{ "date": "2026-10-15", "status": "Working proposals; final method adoption open", "method_id": "irm-x-calibration-proposal-20261014-2", "method_sha256": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "equation": "X = 15 − 13(S + U + J)/15", "criteria": [ { "row": 32, "label": "Proposed rating criteria", "definition": "" }, { "row": 33, "label": "", "definition": "Assess all learned processing that contributes to the result returned by this operation. Use the highest applicable scope rating. Unknown evidence remains unresolved." }, { "row": 34, "label": "S — scope", "definition": "Extent of trained-model determination retained in the returned result." }, { "row": 35, "label": "0", "definition": "No in-scope learned determination remains in the returned result." }, { "row": 36, "label": "1", "definition": "A learned category, scalar or bounded set of global settings; no dense field, structured relations or content construction is learned." }, { "row": 37, "label": "2", "definition": "A learned selection, constraint/dimension set, coordinate or transformation field governing supplied content or geometry, without learned reconstruction of its appearance or new scene content." }, { "row": 38, "label": "3", "definition": "A missing attribute channel or a fixed identified component/property is learned, while the surrounding observed instance defines the task: e.g. chroma, a face-recovery component or a driven mouth region. Evidence must establish this functional restriction." }, { "row": 39, "label": "4", "definition": "Reconstruct an observed instance or its fitted representation, or synthesize a completion under an effective restriction retaining source information outside the generative prediction. Reconstruction includes registered images and scene representations fitted to supplied observations. It does not assert exact pixel copying or unique recovery." }, { "row": 40, "label": "5", "definition": "The operation can generate the entities/parts and arrangement throughout the output domain, with no documented protection/reinstatement restricting retained generative predictions to a component or region. A prompt or reference alone does not establish protection." }, { "row": 41, "label": "", "definition": "" }, { "row": 42, "label": "U — inferential latitude", "definition": "The learned task’s latitude relative to observed material." }, { "row": 43, "label": "0", "definition": "Recognize or locate observed content, or emulate a specified fixed transformation; no restoration, preference assignment or new-content task is included." }, { "row": 44, "label": "1", "definition": "Restore a corrupted observed signal on its existing sampling support under a fidelity target." }, { "row": 45, "label": "2", "definition": "Estimate missing samples of an observed instance under a documented fidelity target, including increased image sampling. Do not infer a fidelity target merely because training examples exist." }, { "row": 46, "label": "3", "definition": "Infer an unobserved attribute, coordinate value, preference or design-intent relation among feasible choices not uniquely specified by the supplied instance and rules; no new semantic scene construction is the task." }, { "row": 47, "label": "4", "definition": "Reconstruct plausible detail or appearance under a documented generative/perceptual/distributional role beyond fidelity-only recovery, or synthesize new content following an explicit supplied structural guide. Such a guide prescribes locations, shapes, poses, trajectories or event ordering for identifiable output parts, in words, coordinates or media; it is more than a high-level subject, style or action request. Evidence must establish that the assessed mechanism uses it. Guidance need not be hard protection. Predicting unobserved samples alone does not establish this anchor." }, { "row": 48, "label": "5", "definition": "Construct a new content/configuration instance without an observed recovery target or an explicit supplied structural guide. Semantic instructions and surrounding context can constrain the task without specifying its internal arrangement." }, { "row": 49, "label": "", "definition": "" }, { "row": 50, "label": "J — learned coordination", "definition": "Relationships the learned process resolves together." }, { "row": 51, "label": "0", "definition": "Separable learned responses; no learned coordination between locations, components or states is established as part of the task." }, { "row": 52, "label": "1", "definition": "Learned local neighboring values/features are resolved together." }, { "row": 53, "label": "2", "definition": "Learned relationships across extended regions or scales of registered supplied content are resolved together." }, { "row": 54, "label": "3", "definition": "Integrate context across the supplied task instance or structure in learned predictions, including a face, a whole subject, a sketch or an observed scene. Establish this functional integration with product-linked task, mechanism or experimental evidence; an undisclosed architecture is not automatically zero. Record the reference domain." }, { "row": 55, "label": "4", "definition": "The learned task jointly constructs semantic and spatial/structural relationships in synthesized content. Mere consistent appearance in a screenshot does not establish the mechanism." }, { "row": 56, "label": "5", "definition": "A learned joint mechanism synthesizes evolving content relationships across time or views. Autoregressive internal tokens, conventional playback or repeated independent calls alone do not establish this." }, { "row": 57, "label": "", "definition": "" }, { "row": 58, "label": "Full method", "definition": "Full rating and evidence rules" } ], "records": [ { "ID": "1", "Software": "Autodesk 3ds Max", "Operation": "Modifier Stack Modeling", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "3ds Max Bend in a specified modifier stack; supplied mesh, angle/direction/local axis, gizmo/center and optional limits. Returned result: Deformed mesh", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Mesh, stack order, deformation parameters and gizmo placement.", "Learned determination": "None in the documented Bend path.", "Conventional processing": "Evaluates the prescribed deformation on mesh coordinates.", "Model contribution in the result": "The transformed geometry retains the designer’s parameterized rule; no learned prediction is added.", "Why this scope score (S)": "S0: Bend transforms supplied vertices from the selected angle, direction, local gizmo axis and limits; no fitted response mapping determines the deformation.", "Why this latitude score (U)": "U0: the operation evaluates the prescribed bend, rather than recognizing, recovering or constructing material through a learned task.", "Why this coordination score (J)": "J0: coordinated changes in vertices follow the same specified deformation rule. Their coupling is conventional geometry, not learned coordination.", "Comparisons and other modes": "Calibration33 Bullet dynamics: Tie at retained-learning endpoint. Both compute substantial coupled geometry from prescribed rules, while their conventional responsibilities differ.\nChallenge: An unexpected deformation might be treated as inference. Checked: Bend’s local-axis/gizmo rules and the January2026 transform report. Result: Surprise and control difficulty do not establish a fitted predictive dependency. Keep the conventional allocation visible.", "Citations": "Autodesk Bend modifier reference — https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Modifiers/files/GUID-C83CBEE2-19FE-4D19-A1F8-52EC04D89A2A.htm", "Source IDs": "N15-BEND", "Supplemental findings": "January2026 OP/replies concern transform state; official help confirms local-gizmo axes. Vendor results were older related reports.", "Evidence limits": "Native Bend only; other modifiers and AI-authored parameter choices are outside this configuration.", "Previous record: scope": "Bend modifier evaluated within a specified 3ds Max modifier stack; assessment ends at the deformed mesh.", "Previous record: rationale": "Prescribed geometric transformation contains no trained predictor. Stack order affects geometry but supplies no inference points.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Angle; Direction; Bend Axis (X/Y/Z); Limit Effect; Upper Limit; Lower Limit; Gizmo; Center. Returns one result; editable structure: mesh with Bend in modifier stack. “The Bend modifier lets you bend the current selection up to 360 degrees about a single axis.” Source • Screenshot", "Collected source links": "https://help.autodesk.com/cloudhelp/2024/ENU/3DSMax-Reference/files/GUID-2C58DCD5-6126-4443-9110-29E47A42E7E7.htm\nhttps://help.autodesk.com/cloudhelp/2022/ENU/3DSMax-Modifiers/files/GUID-C83CBEE2-19FE-4D19-A1F8-52EC04D89A2A.htm\nhttps://help.autodesk.com/cloudhelp/2022/ENU/3DSMax-Modifiers/images/GUID-31B3936C-933D-4DDD-AEDC-2E30154A0747.png", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-1", "Sources": [ { "Source ID": "N15-BEND", "Operation IDs": "1", "Citation": "Autodesk Bend modifier reference", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Modifier Stack and Parameters rollout", "Finding used in this review": "Angle, direction, local axis, gizmo/center and optional limits define the mesh deformation.", "Evidence limits": "Scoped Bend path; no claim about every modifier.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Modifiers/files/GUID-C83CBEE2-19FE-4D19-A1F8-52EC04D89A2A.htm" } ] }, { "ID": "2", "Software": "Autodesk 3ds Max", "Operation": "ReForm Retopology", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "3ds Max2026 ReForm retopology with supplied mesh, target face count and selected Auto Edge constraints; algorithm kept distinct from QuadriFlow/Instant Meshes. Returned result: Retopologized editable quad mesh", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Source mesh, face target, regularity/adaptivity/anisotropy and retained mesh-data constraints.", "Learned determination": "No example-derived predictor is identified in the documented ReForm solver.", "Conventional processing": "Global topology optimization and remeshing under geometric objectives/constraints.", "Model contribution in the result": "The solver reparameterizes a supplied geometric surface under explicit objectives and constraints; the documentation does not establish an example-derived predictive mapping or observation-fitted scene representation.", "Why this scope score (S)": "S0: ReForm solves a quad topology for the supplied surface using explicit geometric objectives, face target and Auto Edge constraints; an example-derived predictive mapping is not established.", "Why this latitude score (U)": "U0: this scoped solver reparameterizes the already supplied surface under those objectives. Its search over topology is not evidence of a retained learned recovery or new-instance task.", "Why this coordination score (J)": "J0: global optimization coordinates mesh topology conventionally. The word global does not establish a learned context mechanism.", "Comparisons and other modes": "Calibration114 specified extrusion: Tie at retained-learning endpoint. Optimization can choose topology more extensively than extrusion while both follow prescribed geometric objectives without established learning.\nChallenge: Topology optimization could be equated with scene-model fitting. Checked: ReForm’s supplied-surface objectives versus v2’s requirement for a fitted predictive mapping/observed-instance representation. Result: Reparameterizing the supplied mesh does not by itself establish such a predictor. Do not import QuadriFlow’s implementation into ReForm.", "Citations": "Autodesk Retopology Tools Interface — https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Retopology/files/GUID-5A960813-FBCC-4A5D-A423-3FCD60825B10.html", "Source IDs": "N15-REFORM", "Supplemental findings": "Recent automated-retopology demonstrations and older dent/failure reports were found; none established a learned ReForm component. Official help warns constraints can cause failure.", "Evidence limits": "Proprietary internals remain undisclosed. X15 is a functional assessment of the documented solver, not source-code proof that every future release lacks learning.", "Previous record: scope": "ReForm reconstructs quad topology through the documented global optimization solver.", "Previous record: rationale": "Search and optimization alone do not establish a trained mechanism. The chosen solver is compared with conventional computation, irrespective of how surprising its topology is.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by search or optimization. Designer controls: Target Size / Face Count; Preprocess Mesh; Quad Tolerance; Subdivision; Regularize; Anisotropy; Adaptivity; Auto Edge (Smoothing Groups, Material IDs, UV Channel, By Angle, etc.); Compute. Returns one result; editable structure: quad mesh. “a state-of-the-art quad topology global optimization technology developed by Autodesk to outperform direct and wave propagation methods significantly, guaranteeing gap and crack-free results” Source • Screenshot", "Collected source links": "https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Retopology/files/GUID-5A960813-FBCC-4A5D-A423-3FCD60825B10.html\nhttps://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-What-s-New/files/GUID-DCF4F5ED-BD3B-4273-88B3-9B2C687752FE.htm\nhttps://igl.ethz.ch/projects/instant-meshes/\nhttps://github.com/hjwdzh/QuadriFlow\nhttps://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Retopology/images/retopo_algorithm_3_reform.png", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-2", "Sources": [ { "Source ID": "N15-REFORM", "Operation IDs": "2", "Citation": "Autodesk Retopology Tools Interface", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "ReForm Settings; Auto Edge Settings", "Finding used in this review": "ReForm is documented as global quad-topology optimization. Face target, regularity, anisotropy, adaptivity and existing mesh-data constraints govern its solution.", "Evidence limits": "Proprietary implementation not inspected. QuadriFlow/Instant Meshes are separate algorithms.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Retopology/files/GUID-5A960813-FBCC-4A5D-A423-3FCD60825B10.html" } ] }, { "ID": "3", "Software": "Ableton Live", "Operation": "Trim and arrange generated video", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "Live12 movie boundary trimming and Arrangement placement at normal playback rate; video remains tempo leader where synchronization is used. Returned result: Edited source-time interval and timeline placement", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "In/out boundaries, retained interval, arrangement position and tempo-leader state.", "Learned determination": "None performed by this editing operation.", "Conventional processing": "Maps retained source frames onto the chosen timeline.", "Model contribution in the result": "The supplied movie may contain earlier generated material; its production precedes this edit and is not newly determined by Live.", "Why this scope score (S)": "S0: the edit selects source-frame boundaries and timeline placement through direct source-time mapping, with no fitted predictor inside the trimming operation.", "Why this latitude score (U)": "U0: the human chooses the interval and placement; this edit does not perform the footage’s earlier generative task.", "Why this coordination score (J)": "J0: frame ordering and playback synchronization follow the specified timeline. The video’s multiple frames do not establish new learned temporal coordination by Live.", "Comparisons and other modes": "ID8 human keyframes/expressions: Tie. Both apply prescribed time/value mappings to supplied material. Neither score measures the entire prior provenance of that material.\nChallenge: Generated video provenance or a follower-warp route might be folded into the edit. Checked: Manual28.2–28.4 and January2026 time-marker report. Result: Keep normal-speed trimming explicit; follower warping is a distinct configuration, still requiring its own dependency account.", "Citations": "Ableton Live12 Working with Video — https://www.ableton.com/en/manual/working-with-video/", "Source IDs": "N15-LIVE", "Supplemental findings": "January13 2026 discussion identifies time-versus-bar marker friction. No matching new learned trimming mechanism found; old audio-trim discussions are mismatched.", "Evidence limits": "", "Previous record: scope": "Live trims movie playback boundaries and places the retained interval in Arrangement View at normal speed.", "Previous record: rationale": "Conventional source-time mapping retains no learned decision from within this editing operation. Earlier movie generation lies outside its scope.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Start Marker; Warp button; Warp Markers; Leader/Follower switch; Video Window. Returns one result; editable structure: clip in Arrangement View timeline. “we drag the Start Marker to the right so the video clip starts at the beginning of the action.” Source • Screenshot", "Collected source links": "https://www.ableton.com/en/manual/working-with-video/\nhttps://help.ableton.com/hc/en-us/articles/209773125-Using-Video", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-3", "Sources": [ { "Source ID": "N15-LIVE", "Operation IDs": "3", "Citation": "Ableton Live12 Working with Video", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "28.2.1; 28.2.3; 28.3 step4; 28.4", "Finding used in this review": "Dragging boundaries chooses source time; Arrangement positions the clip. A video acting as tempo leader plays at its normal rate; followers can warp.", "Evidence limits": "Generation of source footage and audio tempo analysis lie outside this scoped edit.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://www.ableton.com/en/manual/working-with-video/" } ] }, { "ID": "4", "Software": "Adobe After Effects", "Operation": "MCP LLM Composition Construction", "Proposed X": "2.9", "Review status": "Proposed X calculated", "Previous X": "6.0–10.0", "S": "5", "U": "5", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "Pinned Dakkshin MCP bridge; LLM constructs a new still composition from a high-level design request, selecting text/shapes/solids/cameras and their arrangement. No prescribed layout, external generated media or delegated motion sequence. Returned result: Editable native After Effects composition", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "High-level brief, tool access and result acceptance.", "Learned determination": "LLM chooses entities, text, properties and spatial relationships.", "Conventional processing": "The bridge queues commands; native AE construction/rendering executes those choices.", "Model contribution in the result": "Model-selected entities and arrangement remain in editable layers even though native code creates them.", "Why this scope score (S)": "S5: new entities and arrangement are delegated throughout the composition; a fixed repertoire of layer types does not restrict prediction to a supplied component.", "Why this latitude score (U)": "U5: a new composition is constructed without an observed recovery target or prescribed structural arrangement.", "Why this coordination score (J)": "J4: the delegated design integrates semantic choices with spatial layer relationships. Internal code tokens and conventional time evaluation do not establish J5.", "Comparisons and other modes": "Calibration31 Blender MCP: Tie for matched high-level construction. Both retain LLM-selected native entities and arrangement through conventional execution.\nThis still-composition construction is proposed at X 2.9, below human-authored keyframes and expressions (ID8, X 15), because the LLM chooses the composition’s contents and arrangement; ID8 evaluates choices and rules supplied by the designer.\nChallenge: A native-layer repertoire could justify S2, or available keyframes could imply J5. Checked: Pinned tool signatures and v2 functional-target/J rules. Result: Native output media do not erase new-entity construction. Keep animation/delegated-layout configurations separate.", "Citations": "Dakkshin pinned After Effects MCP source — https://raw.githubusercontent.com/Dakkshin/after-effects-mcp/88d5fbf08b7ae9f015ee98e5f8c4904095cf8202/src/index.ts\nDakkshin After Effects MCP README at pinned revision — https://raw.githubusercontent.com/Dakkshin/after-effects-mcp/88d5fbf08b7ae9f015ee98e5f8c4904095cf8202/README.md", "Source IDs": "N15-AEMCP\nN15-MCPREADME", "Supplemental findings": "Read the2026 issue index and August20–21 contributor PR38 live-testing account. It reports effect-name, result-correlation and array-keyframe failures. These affect reliable execution and require checking actual retained output; they do not establish a different allocation for the specified successful still-composition construction.", "Evidence limits": "The allowlist contains20 names, correcting the historical19 count. No local AE execution or specific client-model identity claimed. Contributor PR38 reports execution failures. This is a documentary assessment of a successfully returned composition, not verification that every exposed command works or that proposed fixes are deployed.", "Previous record: scope": "An LLM selects composition objects and their properties; the After Effects MCP bridge executes native construction commands.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The documented model output is a choice of native layer types and their property values, the which-and-where kind of determination REF10 makes, with the appearance of each element prescribed by After Effects; this places the operation nearer REF10 than REF02. The interval stays open below REF10 because the LLM authors the arrangement from a prompt rather than partitioning a supplied image.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. The Dakkshin after-effects-mcp server (31 commits, no release tag, current October 2026) restricts run-script to an allowlist of 19 named ExtendScript operations that create compositions, text, shape and solid layers, cameras and masks, and set properties, keyframes, expressions and effects. No tool calls a generative media model; the LLM supplies layer types and property values, including text strings and expression code, and After Effects renders the pixels. Not established: Whether LLM-written text strings and expressions call for a position closer to REF02 is a comparison-policy question the documentation does not settle. Previous reason: Learned composition decisions survive conventional command execution. Semantic and spatial construction crosses the synthesis and region-determination references; available evidence supports an interval rather than a particular t. raw.githubusercontent.com, main branch, October 2026 • github.com, 31 commits, no release", "Collected source links": "https://github.com/Dakkshin/after-effects-mcp\nhttps://github.com/Dakkshin/after-effects-mcp/blob/88d5fbf08b7ae9f015ee98e5f8c4904095cf8202/src/index.ts\nhttps://raw.githubusercontent.com/Dakkshin/after-effects-mcp/main/src/index.ts", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "43/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-4", "Sources": [ { "Source ID": "N15-AEMCP", "Operation IDs": "4", "Citation": "Dakkshin pinned After Effects MCP source", "Type": "primary code", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "allowedScripts lines163–183; setLayerKeyframe and setLayerExpression; composition/layer tools", "Finding used in this review": "The bridge exposes20 allowlisted command names, including native composition, text, shape, solid, camera, property, mask, keyframe and expression construction. Client-chosen arguments are queued for execution.", "Evidence limits": "Bridge code establishes available actions and transmitted choices, not an LLM architecture or a successful run.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/Dakkshin/after-effects-mcp/88d5fbf08b7ae9f015ee98e5f8c4904095cf8202/src/index.ts" }, { "Source ID": "N15-MCPREADME", "Operation IDs": "4", "Citation": "Dakkshin After Effects MCP README at pinned revision", "Type": "primary code documentation", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Features; Usage Guide; Available MCP Tools", "Finding used in this review": "AI assistants can control composition/layer construction through the bridge. Native objects and property values can be authored and modified.", "Evidence limits": "The assessed configuration explicitly delegates object and arrangement choices to an LLM; literal execution of fully specified commands is different.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/Dakkshin/after-effects-mcp/88d5fbf08b7ae9f015ee98e5f8c4904095cf8202/README.md" } ] }, { "ID": "5", "Software": "Adobe Express", "Operation": "Express AI Assistant", "Proposed X": "2.9", "Review status": "Proposed X calculated", "Previous X": "2.0–10.0", "S": "5", "U": "5", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "Express AI Assistant: construct a new still design from a high-level semantic brief, allowing assistant-chosen text/assets/layout and retained generated imagery. Returned result: New editable still design", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Semantic brief, optional brand/assets, feature selection and acceptance.", "Learned determination": "Assistant plans content and arrangement and invokes generation where the design calls for it.", "Conventional processing": "Native placement, formatting and export apply the assistant’s retained choices.", "Model contribution in the result": "Learned content/layout decisions remain throughout the returned design; conventional editing does not remove them.", "Why this scope score (S)": "S5: this configuration delegates entities and arrangement across a new design, including retained generated media; it is not just a bounded edit to supplied objects.", "Why this latitude score (U)": "U5: the task constructs a new design without a prescribed internal layout.", "Why this coordination score (J)": "J4: design generation jointly resolves content and spatial arrangement. A still-design result does not establish J5.", "Comparisons and other modes": "Calibration31 Blender MCP and ID4 AE MCP: Tie for matched construction. Different toolchains retain the same coarse responsibility for new entities and arrangement.\nExpress repair layout on supplied design: Separate configuration. Structured placement edits to supplied elements fit a different functional scope; this point does not represent every assistant action.\nChallenge: A single assistant label could conceal several modes and exact model routing. Checked: September2026 overview, October2026 FAQ and2025 launch forum. Result: Restrict the point to the stated construction endpoint. Asset search, bounded edits, image-only generation and animation are accounted for as separate follow-up configurations, not silently averaged.", "Citations": "Adobe Express AI Assistant overview — https://helpx.adobe.com/express/web/ai-assistant/adobe-express-ai-assistant-overview.html\nAdobe Express AI Assistant FAQ — https://helpx.adobe.com/express/web/ai-assistant/adobe-express-ai-assistant-faq.html", "Source IDs": "N15-EXPRESS\nN15-EXPFAQ", "Supplemental findings": "Vendor launch discussion and2026 feature announcements describe expanding actions. Current help supersedes old plan/model lists; no firsthand matching result establishes a different profile.", "Evidence limits": "Named partner models are supported by the current overview, not the current FAQ. Exact routing/architecture is unnecessary for this functional construction judgment; other assistant actions remain outside this point.", "Previous record: scope": "Express AI Assistant interprets a design request and orchestrates models, assets and editing tools to construct the design.", "Previous record: rationale": "The retained design includes learned planning and possibly image synthesis. The controller and selected tool path are insufficiently specified for one point within the synthesis-to-region comparison.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). The Adobe Express AI Assistant FAQ (undated, current October 2026) states the assistant combines Adobe Firefly with partner models (Gemini 3.1 with Nano Banana 2, GPT Image 2, FLUX.1 Kontext) and selects the model automatically, with no user choice. Its actions include generative ones (Generate image, Generate template, Insert object, Resize with Expand) and prompted edit actions (change layouts, adjust colors, replace images, choose fonts), so one request can retain synthesized images and placed template elements together, which keeps the 2–10 envelope. Not established: Which model plans the design, and which share of a given design comes from generated images versus templates and edit actions. helpx.adobe.com, undated; page refused WebFetch (403)", "Collected source links": "https://news.adobe.com/news/2025/10/adobe-max-2025-express-ai-assistant\nhttps://helpx.adobe.com/express/web/ai-assistant/adobe-express-ai-assistant-overview.html\nhttps://helpx.adobe.com/express/web/ai-assistant/adobe-express-ai-assistant-faq.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "43/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-5", "Sources": [ { "Source ID": "N15-EXPRESS", "Operation IDs": "5, 8", "Citation": "Adobe Express AI Assistant overview", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "September10 2026: What you can do; Generative AI models used", "Finding used in this review": "Assistant generates designs or finds templates, edits selected elements or entire designs, and restyles images. It automatically chooses Firefly/partner models. Current overview names Gemini3.1/Nano Banana2, GPT Image2 and FLUX.1 Kontext for image generation.", "Evidence limits": "Current model names come from the overview; the FAQ does not identify particular models.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://helpx.adobe.com/express/web/ai-assistant/adobe-express-ai-assistant-overview.html" }, { "Source ID": "N15-EXPFAQ", "Operation IDs": "5", "Citation": "Adobe Express AI Assistant FAQ", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "October5 2026: action table and availability", "Finding used in this review": "Design generation, layout/color/font editing, asset search and image/object generation are distinct assistant actions.", "Evidence limits": "Billing categories do not reveal mechanism or measure model contribution. Current FAQ does not identify particular models.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://helpx.adobe.com/express/web/ai-assistant/adobe-express-ai-assistant-faq.html" } ] }, { "ID": "6", "Software": "Affinity by Canva", "Operation": "Object Selection in the Unified App", "Proposed X": "10.7", "Review status": "Proposed X calculated", "Previous X": "10.0", "S": "2", "U": "0", "J": "3", "Reviewed on": "2026-10-14", "Configuration": "Unified Affinity local Object Selection tool, installed ML model, supplied pixel layer; initial object selection before manual replacement/refinement. Returned result: Object-membership selection", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Image/layer, indicated object or area, add/subtract choices and acceptance.", "Learned determination": "Installed ML processing determines object membership.", "Conventional processing": "Selection application and edge postprocessing carry predicted membership into the result.", "Model contribution in the result": "A semantic selection remains; this is distinct from Canva’s generative-image panels.", "Why this scope score (S)": "S2: learned output is a membership field governing supplied pixels, with no learned appearance reconstruction.", "Why this latitude score (U)": "U0: the task recognizes observed objects; selecting an object does not construct a new content instance.", "Why this coordination score (J)": "J3: the reference domain is the selected object within the supplied image. The learned task determines pixels as belonging to that object, rather than applying independent color-and-tolerance matches. This supports object-level context integration as a functional inference; it does not establish the segmentation architecture or identical weights in the current app and its predecessor.", "Comparisons and other modes": "Calibration388 Photoshop Select Subject Cloud: Tie. Both retain context-based semantic membership of an observed subject.\nAffinity Object Selection is proposed at X 10.7, below Magic Wand (ID390, X 15): the model determines which pixels belong to an object, whereas Magic Wand applies the specified color and tolerance rules.\nChallenge: Predecessor documentation might be borrowed as proof of current weights, or selected area treated as generated content. Checked: Current unified-app ML settings, predecessor task guide and September2026 report distinction. Result: Current official help establishes ML Object Selection; weight identity remains unclaimed and does not prevent this functional profile.", "Citations": "Current Affinity Settings help — https://www.affinity.studio/help/workspace-settings/", "Source IDs": "N15-AFFSET", "Supplemental findings": "Recent indexed report excerpts distinguish local selection from Canva panels. Direct thread retrieval failed and exact-title follow-up returned no full thread. Current official ML-settings text independently establishes the local-model feature; report excerpts are not facet evidence.", "Evidence limits": "AFF2 is related lineage evidence. The actual segmentation architecture, matching weights and accuracy are not established.", "Previous record: scope": "Object Selection in unified Affinity determines object membership in an image; the trained segmentation lineage is documented for its predecessor.", "Previous record: rationale": "A region-determination match follows the operation's task. Current model continuity is inferred and does not establish identical weights or mask quality.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). TidBITS coverage of the unified Affinity app (31 October 2025) lists Object Selection among Canva AI tools, while a user report on the same page states the segmentation model downloads free; Affinity guides describe installing the Segmentation model under Settings > Machine Learning. No generative step is documented for Object Selection. Not established: Whether the unified app uses the Affinity Photo 2.6 segmentation weights, and whether any Object Selection path calls Canva cloud models, is not documented. Serif support pages refused direct fetch. tidbits.com, 2025-10-31 • scrapbookcampus.com, 2025-10", "Collected source links": "https://tidbits.com/2025/10/31/canvas-affinity-combines-photo-designer-and-publisher-into-one-free-app/\nhttps://affinity.help/photo2/English.lproj/pages/Tools/tools_objectSelection.html\nhttps://scrapbookcampus.com/2025/10/object-selection-in-affinity-photo/", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "32/3", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-6", "Sources": [ { "Source ID": "N15-AFFSET", "Operation IDs": "6", "Citation": "Current Affinity Settings help", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Machine Learning Models section", "Finding used in this review": "Current unified-app help states that locally installed ML models enable Object Selection.", "Evidence limits": "Current help confirms ML Object Selection but does not establish identical weights to Affinity Photo 2.6.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://www.affinity.studio/help/workspace-settings/" } ] }, { "ID": "7", "Software": "Adobe After Effects", "Operation": "Content-Aware Fill Object mode on generated video", "Proposed X": "", "Review status": "Numerically unresolved", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "2026-10-14", "Configuration": "After Effects Content-Aware Fill, Object mode on supplied generated video; animated transparency mask; no supplied reference frame; output includes fill layer and composited video. Returned result: Filled region across the clip", "Next step": "Identify whether the current Object mode uses a learned model for background-motion estimates or fill synthesis that remains in the returned video. The available documentation does not establish that dependency, so the S/U/J ratings and X remain unresolved.", "Human input and controls": "Video, mask, work area, Object mode and alpha expansion.", "Learned determination": "Which motion/fill stages retain learned predictions is not established by the available product-linked evidence.", "Conventional processing": "Pixel transfer, fill-layer storage and compositing are documented. Motion estimation is documented, but its learned or prescribed implementation remains unresolved.", "Model contribution in the result": "Reused background and any missing-background synthesis remain, but their learned versus prescribed dependency cannot yet be separated.", "Why this scope score (S)": "Unresolved: The current no-reference Object-mode implementation does not disclose whether retained motion estimates or fill synthesis come from a fitted predictor, and which functional contribution that predictor makes. Help, Cloak, research search and report follow-up do not settle it.", "Why this latitude score (U)": "Unresolved: The current no-reference Object-mode implementation does not disclose whether retained motion estimates or fill synthesis come from a fitted predictor, and which functional contribution that predictor makes. Help, Cloak, research search and report follow-up do not settle it.", "Why this coordination score (J)": "Unresolved: The current no-reference Object-mode implementation does not disclose whether retained motion estimates or fill synthesis come from a fitted predictor, and which functional contribution that predictor makes. Help, Cloak, research search and report follow-up do not settle it.", "Comparisons and other modes": "Calibration63 protected learned inpainting: No numerical order established. A mask and fill endpoint do not establish the same learned dependency as pinned SD3.5 inpainting.\nChallenge: Sensei branding or synthesis wording could be treated as proof of a generative network; historical patch matching could be treated as proof of no learning. Checked: Current Object-mode help, historical Cloak description, targeted Adobe Research mechanism search and recent reference-frame discussion. Result: Neither inference identifies the current contributing components. Withhold X and leave S/U/J blank.", "Citations": "Adobe After Effects Content-Aware Fill — https://helpx.adobe.com/after-effects/desktop/remove-objects-from-your-videos/content-aware-fill.html", "Source IDs": "N15-CAF", "Supplemental findings": "May2026 reference-frame discussion was read; an October2026 failure thread was indexed but direct retrieval failed and exact follow-up returned none. Official help confirms that reference frames change the configuration. Neither route identifies the no-reference learned component.", "Evidence limits": "A private local trial could reveal behavior but would not by itself identify hidden learning. No public product-linked method source located resolves that dependency.", "Previous record: scope": "After Effects Content-Aware Fill Object mode combines motion-based temporal reuse with synthesis of missing background.", "Previous record: rationale": "The sources distinguish it from Firefly but do not resolve the active learned contribution. Determine that contribution before selecting the learned or conventional equation branch; background visibility supplies no score.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Adobe's help page describes Object mode as filling from current and surrounding frames while estimating the motion of the scene behind the object; Adobe Research's Cloak note (11 Dec 2017) describes forward and backward motion estimation with content-aware fill for background hidden in all frames. A third-party 2026 blog claims Generative Fill for Video replaced the legacy engine; no Adobe page located confirms that claim for After Effects 26.x (June 2026), and the January 2026 Adobe blog on After Effects names no change to Content-Aware Fill. Not established: Whether the motion estimator in After Effects 26.x is a trained network or a prescribed optical-flow method; whether background hidden in all frames is synthesized by patch-based search (PatchMatch-type) or by a trained inpainting model; whether any Firefly model participates in Content-Aware Fill. helpx.adobe.com • research.adobe.com, 2017-12-11 • fxguide.com, 2019-04-04", "Collected source links": "https://helpx.adobe.com/after-effects/desktop/remove-objects-from-your-videos/content-aware-fill.html\nhttps://research.adobe.com/news/cloak-remove-unwanted-objects-in-video/\nhttps://helpx.adobe.com/after-effects/using/content-aware-fill.html\nhttps://www.fxguide.com/quicktakes/content-aware-fill-in-ae-and-other-new-gems/\nhttps://web.archive.org/web/20260521182339id_/https://helpx.adobe.com/after-effects/using/projects.html\nhttps://web.archive.org/web/20260928142651id_/https://helpx.adobe.com/after-effects/desktop/remove-objects-from-your-videos/content-aware-fill.html", "Evidence status": "Learned dependency unresolved; X withheld", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-7", "Sources": [ { "Source ID": "N15-CAF", "Operation IDs": "7", "Citation": "Adobe After Effects Content-Aware Fill", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Object mode; reference frames; fill layer and alpha", "Finding used in this review": "Object mode fills a transparent region from current/surrounding frames using estimated background motion. A fill layer is generated; reference frames are optional.", "Evidence limits": "Sensei branding does not establish which current motion or fill components use learned prediction.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://helpx.adobe.com/after-effects/desktop/remove-objects-from-your-videos/content-aware-fill.html" } ] }, { "ID": "8", "Software": "Adobe After Effects", "Operation": "Keyframes and linked expressions on generated artwork", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "After Effects manually authored property keyframes and linked/human-written expressions on supplied artwork; no LLM writing or learned tracking inside this operation. Returned result: Animated property values and rendered artwork", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Keyframe values/times, interpolation/handles, expression code and links.", "Learned determination": "None in the scoped evaluation of supplied rules.", "Conventional processing": "Evaluates time/value interpolation and JavaScript property functions, then renders supplied artwork.", "Model contribution in the result": "Human-specified functional rules govern the animation; artwork generation is inherited input provenance.", "Why this scope score (S)": "S0: AE evaluates the supplied keyframe values/interpolation and human-authored JavaScript property functions; no fitted predictor supplies those rules.", "Why this latitude score (U)": "U0: the scoped operation executes a prescribed time/value function. Auto Bezier interpolation and wiggle randomness do not establish a learned recovery or construction task.", "Why this coordination score (J)": "J0: linked property values can be tightly coordinated, but those links and functions are explicitly authored and conventionally evaluated.", "Comparisons and other modes": "Human-authored keyframes and expressions are proposed at X 15, above LLM composition construction (ID4, X 2.9). After Effects evaluates rules supplied by the designer in this configuration; the LLM chooses content and arrangement in ID4.\nChallenge: Automatic Bezier curves, wiggle randomness or complex linked motion could be treated as learning. Checked: Adobe interpolation and expression definitions. Result: Programmed interpolation/randomness and dependency evaluation do not establish a fitted predictive role. LLM-authored expressions are another configuration.", "Citations": "Adobe keyframe interpolation — https://helpx.adobe.com/after-effects/desktop/animate-in-after-effects/animation-keyframes/keyframe-interpolation.html\nAdobe expression basics — https://helpx.adobe.com/after-effects/desktop/work-with-expressions/expression-basics/expression-basics.html", "Source IDs": "N15-KEYFRAMES\nN15-EXPRESSIONS", "Supplemental findings": "A2024 baking discussion has2026 follow-ups and newer template discussions address value/position behavior; they support checking authoring boundaries, not new learning.", "Evidence limits": "", "Previous record: scope": "After Effects evaluates manually authored keyframes and linked expressions to animate artwork.", "Previous record: rationale": "Interpolation and expression evaluation implement prescribed functions without a trained predictor. Source artwork provenance does not change this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: keyframes; interpolation type (Linear, Bezier, Auto Bezier, Continuous Bezier, Hold); direction handles; expression field; pick whip. Returns one result; editable structure: layer property keyframes and expressions. “Linear interpolation creates a uniform rate of change between keyframes, which can add a mechanical look to animations.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/after-effects/desktop/animate-in-after-effects/animation-keyframes/keyframe-interpolation.html\nhttps://helpx.adobe.com/after-effects/desktop/work-with-expressions/expression-basics/expression-basics.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-8", "Sources": [ { "Source ID": "N15-KEYFRAMES", "Operation IDs": "8", "Citation": "Adobe keyframe interpolation", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Linear, Bezier, Auto Bezier, Continuous Bezier and Hold interpolation", "Finding used in this review": "Specified keyframe values and interpolation rules determine intermediate property values.", "Evidence limits": "Automatic Bezier calculations remain specified interpolation, not evidence of a learned predictor.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://helpx.adobe.com/after-effects/desktop/animate-in-after-effects/animation-keyframes/keyframe-interpolation.html" }, { "Source ID": "N15-EXPRESSIONS", "Operation IDs": "8", "Citation": "Adobe expression basics", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "August10 2026: definition and property linking", "Finding used in this review": "Expressions are JavaScript property-value functions evaluated at a given time; pick-whip links can relate properties.", "Evidence limits": "Human-written expressions are assessed here. LLM-authored expressions require a different determination account.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://helpx.adobe.com/after-effects/desktop/work-with-expressions/expression-basics/expression-basics.html" } ] }, { "ID": "9", "Software": "Adobe After Effects", "Operation": "Roto Brush 3 segmentation of generated video", "Proposed X": "10.7", "Review status": "Proposed X calculated", "Previous X": "10.0", "S": "2", "U": "0", "J": "3", "Reviewed on": "2026-10-14", "Configuration": "After Effects effect Version3.0 Roto Brush; base-frame foreground/background strokes, propagation and optional classic refinement; assess the retained matte on supplied footage. Returned result: Foreground matte across the supplied shot", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Footage, foreground/background strokes, span and corrective/refinement controls.", "Learned determination": "Roto Brush3 predicts object membership and propagates/tracks it across supplied frames.", "Conventional processing": "Matte freezing, classic refinement and compositing retain or refine predicted membership.", "Model contribution in the result": "The matte carries learned recognition across the observed shot; no scene appearance is synthesized by this operation.", "Why this scope score (S)": "S2: learned membership/selection fields govern supplied footage. No learned appearance reconstruction is part of this result.", "Why this latitude score (U)": "U0: selection/tracking recognizes observed footage; motion through frames does not turn recognition into new-instance generation.", "Why this coordination score (J)": "J3: the named reference domain is the supplied shot. Context/identity across its frames is integrated in a retained segmentation task, below J4/J5 new-content synthesis.", "Comparisons and other modes": "Calibration388 Photoshop Select Subject and ID6 Affinity selection: Tie under the frozen anchors. All retain semantic membership of observed content. The shot reference domain and corrective work remain visible even though the scalar does not separately reward temporal recognition.\nChallenge: Temporal propagation might earn J5, or current Object Matte docs might replace the specified Roto Brush version. Checked: Adobe’s Version3.0 selector, new-AI-model account and v2 J5 synthesis requirement. Result: Keep3.0 explicit and J3. This tie exposes the rubric’s coarse treatment of temporal recognition without inventing a new anchor.", "Citations": "Adobe Roto Brush and Refine Matte — https://helpx.adobe.com/after-effects/desktop/roto-brush-and-refine-matte/roto-brush/roto-brush-refine-matte.html", "Source IDs": "N15-ROTO", "Supplemental findings": "Recent discussions include performance and tool/version confusion; older GPU reports are not evidence about a current architecture. Official help confirms3.0 remains selectable.", "Evidence limits": "J3 is a functional inference. Temporal segmentation quality and correction effort are not measured by X.", "Previous record: scope": "Roto Brush 3 predicts and propagates foreground segmentation through video.", "Previous record: rationale": "Temporal mask determination matches the region reference at the present resolution. Repeated frames and corrective strokes do not automatically change the coordinate.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). After Effects 26.2 (15 April 2026) replaced the Roto Brush tool with the Object Matte tool, which selects a subject by click or marquee and propagates a matte through the shot. Adobe staff state that earlier Roto Brush versions remain selectable through a version selector in the Object Matte effect controls. No documented step generates or fills image content; the output remains a matte. Not established: Adobe does not name the Object Matte model or state whether it shares weights with Roto Brush 3. Adobe helpx pages refused direct fetch (HTTP 403). cgchannel.com, 2026-04-15 • community.adobe.com, AE 26.2.1 • community.adobe.com, 2026-05-29", "Collected source links": "https://helpx.adobe.com/after-effects/desktop/roto-brush-and-refine-matte/roto-brush/roto-brush-refine-matte.html\nhttps://research.adobe.com/news/project-fast-mask/\nhttps://www.cgchannel.com/2026/04/adobe-releases-after-effects-26-2/\nhttps://community.adobe.com/questions-529/the-rotobrush-tool-is-missing-from-my-most-recent-update-v26-2-1-it-is-in-the-prior-version-25-6-5-1559301\nhttps://community.adobe.com/announcements-527/new-in-ae-object-matte-ai-powered-rotoscoping-1557065\nhttps://web.archive.org/web/20260521182339id_/https://helpx.adobe.com/after-effects/using/projects.html\nhttps://web.archive.org/web/20260928143017id_/https://helpx.adobe.com/after-effects/desktop/roto-brush-and-refine-matte/roto-brush/roto-brush-refine-matte.html\nhttps://web.archive.org/web/20260928141758id_/https://helpx.adobe.com/after-effects/desktop/apply-effects-and-animation-presets/effects-and-animation-presets/effects-animation-presets-overview.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "32/3", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-9", "Sources": [ { "Source ID": "N15-ROTO", "Operation IDs": "9", "Citation": "Adobe Roto Brush and Refine Matte", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "April15 2026: Next-Gen Roto Brush; Version3.0; propagation/classic controls", "Finding used in this review": "Roto Brush3 uses a new AI model to select and track objects across space/time from representative foreground/background strokes. Version3.0 remains selectable; classic processing can follow it.", "Evidence limits": "Object Matte is a separate newer interface/path. Learned temporal recognition is not new-content video synthesis.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://helpx.adobe.com/after-effects/desktop/roto-brush-and-refine-matte/roto-brush/roto-brush-refine-matte.html" } ] }, { "ID": "10", "Software": "Agisoft Metashape", "Operation": "Photographs through depth-map estimation to editable mesh", "Proposed X": "7.2", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "4", "U": "2", "J": "3", "Reviewed on": "2026-10-14", "Configuration": "Metashape2.3.2 supplied calibrated/overlapping photographs through depth-map estimation and Build Model(Depth Maps source) to the retained editable mesh. Exclude AI masks, learned classification and later texture-generation tools. Returned result: Editable polygon mesh", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Photographs, calibration/reference information, reconstruction volume and quality/filtering/meshing settings.", "Learned determination": "Under v2’s expressly broadened data-fitted representation scope, camera/depth estimates and reconstructed scene geometry fit the supplied observations. This is not a claim of population training or a neural depth model.", "Conventional processing": "Dense stereo correspondences, calibrated geometry, depth-map fusion/filtering, surface construction and interpolation implement the observational fit.", "Model contribution in the result": "The returned scene mesh retains the observation-derived surface estimate, including geometry interpolated between observed samples. It can supply positions/surfaces beyond the directly measured image samples. Fitting and returned representation are within this operation.", "Why this scope score (S)": "S4: the operation estimates and retains a representation of the observed scene from multiple photographs. V2 includes such fitted scene representations; mesh editability or lack of neural pretraining does not remove them. This differs from remeshing a supplied surface in ReForm.", "Why this latitude score (U)": "U2: the task is faithful geometric reconstruction of the observed scene from sampled multiview evidence. The documented matching/filtering target supports recovery, not a generative/perceptual invention role at U4.", "Why this coordination score (J)": "J3: the named domain is the supplied observed scene. Camera/depth relationships across overlapping views jointly support its reconstructed surface. A static fitted scene is not J5 synthesis of evolving new content.", "Comparisons and other modes": "Calibration124 gsplat baseline: Tie at4/2/3 under v2. Both fit and retain an observed-scene representation from multiview samples under a fidelity target. Their numerical algorithms/media differ, but the frozen method includes fitted scene state without requiring a pretrained neural model.\nThe complete Metashape photo-to-mesh route is proposed at X 7.2, below ReForm (ID2, X 15) and specified extrusion (ID114, X 15). Metashape estimates an observed scene’s geometry from photographs; the other operations use supplied geometry and prescribed construction or remeshing rules. Under v2, the distinction is whether the returned result retains a representation fitted to observations, rather than whether the computation uses optimization.\nChallenge: Retain historical X15 because stereo/meshing uses conventional arithmetic and no pretrained network is identified. Checked: Fresh original-first reviewer challenged the zero branch against v2’s explicit fitted-scene inclusion. Manual printed45–53 establishes observation-derived camera/depth state and reconstructed/interpolated scene geometry. Result: Reject the zero branch for the stated complete photos-to-mesh endpoint. Apply the same4/2/3 profile as the fitted-scene comparator. This is a consequence of v2’s broader scope, not discovery of a hidden neural model. A supplied-depth-maps-only conversion is a separate boundary.", "Citations": "Agisoft Metashape2.3 professional manual — https://www.agisoft.com/pdf/metashape-pro_2_3_en.pdf\nAgisoft3D model reconstruction guide — https://agisoft.freshdesk.com/support/solutions/articles/31000152092-3d-model-reconstruction", "Source IDs": "N15-METAMAN\nN15-METAMESH", "Supplemental findings": "Dated2.3 laser-scan matching reports concern alignment and are related only. An October8 2025 prerelease mesh failure is outside the12-month window and not proof of current failure or learning.", "Evidence limits": "Functional expert application of v2’s broadened fitted-representation rule. No neural architecture asserted and no software trial executed. The complete photos/depth/mesh dependency is assessed; an already-supplied-depth-only conversion needs a separate record.", "Previous record: scope": "Metashape depth-map reconstruction produces a mesh using the documented conventional reconstruction route; optional AI masks are excluded.", "Previous record: rationale": "Multiview matching and geometric reconstruction do not establish a retained trained predictor within this defined route.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by reconstruction from captured images. Designer controls: source data (depth maps); surface type; face count / custom face count; interpolation; vertex colors; vertex confidence; build texture; clip to boundary; trimming radius. Returns one result; editable structure: polygonal mesh. “Metashape can reconstruct polygonal mesh model based on the point cloud information (Point Cloud, Tie Points) or based on the depth maps data.” Source • Screenshot", "Collected source links": "https://agisoft.freshdesk.com/support/solutions/articles/31000173952-new-features-in-agisoft-metashape-2-2-x\nhttps://www.agisoft.com/pdf/metashape-pro_1_6_en.pdf\nhttps://www.agisoft.com/pdf/metashape-pro_2_3_en.pdf\nhttps://agisoft.freshdesk.com/support/solutions/articles/31000152092-3d-model-reconstruction\nhttps://s3.amazonaws.com/cdn.freshdesk.com/data/helpdesk/attachments/production/31071870824/original/z0WErJxFLGFSMpUx9fnmy8psKuSbgRNqvg.png", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "36/5", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-10", "Sources": [ { "Source ID": "N15-METAMAN", "Operation IDs": "10", "Citation": "Agisoft Metashape2.3 professional manual", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Printed pp45–53: Building point cloud; Building model; depth filtering and interpolation", "Finding used in this review": "The manual describes pairwise dense stereo matching from calibrated overlapping views, depth-map merging, geometric meshing, connected-component noise filters and interpolation/extrapolation settings.", "Evidence limits": "No neural depth model is established. Broad v2 inclusion of fitted predictive scene state requires an explicit comparison with gsplat; numerical meshing alone is insufficient.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://www.agisoft.com/pdf/metashape-pro_2_3_en.pdf" }, { "Source ID": "N15-METAMESH", "Operation IDs": "10", "Citation": "Agisoft3D model reconstruction guide", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "April8 2026: Align Photos and Build Model", "Finding used in this review": "Depth maps and point clouds are separate mesh source choices. Geometry generation precedes optional texture building; unseen roofs can remain absent.", "Evidence limits": "The assessed route excludes AI masks and later texture generation.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://agisoft.freshdesk.com/support/solutions/articles/31000152092-3d-model-reconstruction" } ] }, { "ID": "11", "Software": "Alpha3D", "Operation": "Single Image to Mesh", "Proposed X": "3.7", "Review status": "Proposed X calculated", "Previous X": "4.0", "S": "5", "U": "4", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "Alpha5 single supplied image to textured mesh; exclude multiview inputs, later retopology/texturing/rigging and Alphred orchestration. Returned result: Generated textured3D asset", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Single reference image, available output settings and acceptance.", "Learned determination": "The image-to3D generator determines unobserved geometry, surface appearance and part relationships.", "Conventional processing": "Mesh/material packaging and export retain those predictions.", "Model contribution in the result": "A new asset is constructed from the supplied2D structural reference; no protected source-mesh region exists.", "Why this scope score (S)": "S5: generation determines geometry/parts and appearance throughout the3D asset; the image does not reinstate a source-mesh complement.", "Why this latitude score (U)": "U4: the supplied reference depicts locations/shapes of identifiable object parts and guides the generated structure. Unseen geometry remains inferred; this is not documented fidelity-only multiview recovery.", "Why this coordination score (J)": "J4: constructing one textured object coordinates semantic parts and spatial geometry. A3D file or a Hunyuan-name inference does not establish learned joint synthesis across views at J5.", "Comparisons and other modes": "Calibration65 active Canny guidance: Tie at5/4/4. Both construct whole-domain content under a supplied structural reference without hard source protection.\nCalibration124 gsplat observations: Different latitude and scope. Alpha5 constructs an asset from one depiction; gsplat’s calibrated multiview task fits an observed scene under a fidelity target.\nChallenge: An image could be treated as copied mesh content, fidelity recovery or automatic evidence for J5. Checked: Current FAQ separation of input routes; archived13250 input/output pair re-inspected; v2 guide and coordination rules. Result: Use functional construction with U4 and J4. The archived pair does not reveal architecture or quantify preserved fractions.", "Citations": "Alpha3D current FAQ — https://www.alpha3d.io/faq\nPreserved Alpha5 input/output observation — archived research evidence; no public link supplied", "Source IDs": "N15-ALPHA\nN15-ALPHAPAIR", "Supplemental findings": "No matching recent Reddit report was returned. Older Rhino discussion concerns multiple tools; it is related context only. Existing published Alpha5 evidence remains available.", "Evidence limits": "Vendor has not confirmed inferred Hunyuan identity; no borrowed Hunyuan architecture claim. One archived pair was re-inspected rather than rerunning the previous five-pair study.", "Previous record: scope": "Alpha3D predicts a three-dimensional asset from an image, ending at the generated mesh.", "Previous record: rationale": "Working v5 documentation coding with 5 Oct 2026 output evidence: 2 kept, 2 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 4.0. Surface appearance and fine detail were coded model from five published Alpha 5 generations, input image and GLB compared by two independent coders (10 of 10 agreement).", "Previous record: calculation": "K=2; M=2; unknown=0; conditional=0; A=4. 2 + 3 × 2 ÷ (4 − 1) = 4.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Alpha3D's own model line, named \"Alpha 5\" in its current release (Alpha3D). Alpha3D describes its AI as proprietary and developed internally. Model choice: Not documented. Not established: Architecture and training of Alpha 5; whether it builds on an open model such as TRELLIS, Hunyuan3D or TripoSG; release date of Alpha 5. alpha3d.io • techcrunch.com, 2023-09-20\nChecked 5 Oct 2026 X Unresolved (lower bound 4.0) → 4.0. Evidence: five Alpha 5 image-to-3D generations from Alpha3D's public gallery (posts 13250, 13297, 13325, 13381, 20–21 Apr 2026; 15445, 13 Jul 2026), each with its input image and GLB. All five GLBs contain base-colour, metallic-roughness and normal maps (4096 px). Two independent coders (10 of 10 agreement): the base colour lacks the input's highlights and directional shading in all five; colours depart in part of each object; fine features are partly retained and partly replaced or added in all five. Surface appearance and fine detail: unknown → model. Model: probably Tencent HY 3D 3.0 (the site code maps the model id \"hunyuan\" to \"Alpha 5\"; request parameters and task-ID format match Tencent's Hunyuan 3D API); not confirmed by Alpha3D or Tencent. Sensitivity: fine detail coded unknown gives 4.0–5.0; the proposed dimension rule for shape (under review) gives 3.0. Limits: five published generations; three inputs are renders or concept art. Records: 06_research_archive/2026-10-05_alpha3d_op11_evidence. Alpha3D public gallery • Tencent SubmitHunyuanTo3DProJob\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (image-to-3D). Designer controls: image upload (single image, or front, back, side and top views). Returns count not documented; editable structure: mesh; GLB, FBX, OBJ, STL, USDZ export. “Users can upload a single product photo, sketch or concept image and receive a textured 3D model automatically.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.alpha3d.io\nhttps://www.alpha3d.io/image-to-3d\nhttps://www.alpha3d.io/news/alpha-5-is-live-a-new-standard-for-ai-3d-generation\nhttps://techcrunch.com/2023/09/20/alpha3d-wants-to-accelerate-digital-asset-and-ai-powered-generation-through-cheaper-hyper-scaling-technology/\nhttps://www.alpha3d.io/explore\nhttps://www.tencentcloud.com/document/product/1284/75540\nhttps://www.alpha3d.io/news/how-to-generate-3d-models-from-text-and-images", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "56/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-11", "Sources": [ { "Source ID": "N15-ALPHA", "Operation IDs": "11", "Citation": "Alpha3D current FAQ", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "About Alpha3D; Image-to-3D; Alpha5 jobs", "Finding used in this review": "A single image can generate a textured3D asset. Multiview input and later retopology/texturing/rigging/agent orchestration are separate routes.", "Evidence limits": "The vendor does not confirm the inferred Hunyuan implementation.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://www.alpha3d.io/faq" }, { "Source ID": "N15-ALPHAPAIR", "Operation IDs": "11", "Citation": "Preserved Alpha5 input/output observation", "Type": "provided artifact", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Archived input13250 and front unlit-basecolour mesh render inspected October14", "Finding used in this review": "The two chairs and connecting unit remain identifiable, while shading, small surface details and supports differ in the generated asset.", "Evidence limits": "One archived pair re-inspected, not a new generation or a representative success-rate test. Image resemblance does not prove protected copying.", "Review status": "read; direct", "Open source": "Archived research evidence", "URL": "" } ] }, { "ID": "12", "Software": "Apple Keynote / Pages / Freeform", "Operation": "Apple Image Playground (Keynote/Pages/Freeform)", "Proposed X": "2.9", "Review status": "Proposed X calculated", "Previous X": "2.0", "S": "5", "U": "5", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "Image Playground new still image from a high-level description and suggested concepts in a selected style; no prescribed layout. Person/photo identity conditioning is discussed separately; gallery edits are outside this point. Returned result: New still image inserted into the chosen app", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "High-level description/concepts, style and aspect ratio, with result choice.", "Learned determination": "Image generation determines new entities/appearance and arrangement.", "Conventional processing": "Gallery storage and insertion into Keynote/Pages/Freeform present the generated image.", "Model contribution in the result": "The image retains whole-domain learned construction; app insertion does not change that responsibility.", "Why this scope score (S)": "S5: the new-image route constructs entities and arrangement across the output; no protected source region is documented.", "Why this latitude score (U)": "U5: high-level concepts specify subject and style without prescribing internal locations, pose or arrangement. A separately assessed route could qualify for U4 only if the input specifies those structural relationships and evidence establishes that the generation mechanism uses them.", "Why this coordination score (J)": "J4: the still-image task jointly constructs semantic/spatial relationships. The Animation style name does not establish a temporal result or J5.", "Comparisons and other modes": "Calibration64 SD3.5 high-level text-to-image: Tie for matched task. Both delegate whole still-image content and arrangement without a prescribed structure.\nSelected-person identity route: Conditional comparison. Identity conditioning alone supplies a subject constraint, not hard protection or automatic structural-layout guidance. It cannot justify the historical kept-content fraction.\nChallenge: All text inputs might be codedU5, or a supplied person/photo might automatically earn protection. Checked: Current macOS27 person/photo note and description/style/gallery controls. Result: Input format does not decide U. A precise layout or pose instruction is a candidate for U4 only with evidence that the mechanism uses the supplied structure; identity-only conditioning does not prove either layout guidance or source copying. Gallery edits remain a separate endpoint.", "Citations": "Apple macOS27 Image Playground guide — https://support.apple.com/en-gb/guide/mac-help/mchld5412d00/27/mac/27\nApple Keynote Add an image — https://support.apple.com/guide/keynote/add-an-image-tan77aea6844/mac", "Source IDs": "N15-APPLE\nN15-KEYNOTE", "Supplemental findings": "June2026 firsthand photo-modification discussion reports intermittent errors; October2026 iOS27 output-failure discussion concerns quality, not a different allocation. Current macOS guide confirms the routes.", "Evidence limits": "The point is for the specified high-level new-image route. Source availability, model makers and identities can change without this functional profile changing.", "Previous record: scope": "Image Playground synthesizes an image through the selected Apple generative image mode.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: From iOS 27 / macOS 27 (announced 2026-06-08): a new Apple generative model running on Private Cloud Compute, part of the next generation of Apple Foundation Models, custom-built in collaboration with Google and its Gemini models; outputs carry Google's SynthID watermark. Earlier: Apple's own image model (Animation, Illustration, Sketch styles, 2024), and from iOS 26 (2025) OpenAI ChatGPT image generation for the ChatGPT styles (Oil Painting, Watercolor, Vector, Anime, Print, Any Style). Model choice: Style choice, with no explicit model picker. Not established: Name, size and maker split of the iOS 27 image model (Apple-trained, Gemini-derived, or a Google image model such as Nano Banana); the SynthID watermark points to Google involvement but Apple does not name an image model. apple.com, 2026-06-08 • macrumors.com, 2026-06-08 • apple.com, 2025-06-09\nChecked 5 Oct 2026 X unchanged (2.0), validated for the description-and-concepts route. Both v5 coders coded all five applicable categories model before adjudication. Under R2 no image input is required, and a text-only input has no instance of layout, shape, surface appearance or fine detail for the output to keep. Coder B's Keynote quotation is confirmed on Apple's page: \"create fun, original images based on a description and suggested concepts.\" The model changes recorded on 4 Oct 2026 do not move X, because all five categories are model-decided under each model. Routes outside the coded scope: Keynote documents an Add to Playground command for text or an image selected on a slide; the macOS 27 Mac User Guide states that when a description does not match a selected person, the image is personalised from the selected photo over the text. In the person route the photograph is required and identity belongs to content; proposed coding for that route: content kept, other four categories model, X = 2.75 (not double-coded). The macOS 27 targeted edits act on an image already in the gallery and form a separate editing operation. Proposed scope sentence (pending decision): \"Image Playground generates a new image from a description and suggested concepts in the selected style.\" Apple Keynote: Add an image • Apple: Image Playground on Mac (macOS 27)\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Apple image model; ChatGPT style routes to ChatGPT). Designer controls: description; suggested concepts; person from Photos (up to three); photo; style (Animation, Illustration, Sketch, ChatGPT); aspect ratio. Returns several to choose from (versions; count not stated); a flat image. “Click the Choose Person button, then select a person. You can add up to three people.” Source • Screenshot", "Collected source links": "https://www.apple.com/newsroom/2024/12/apple-intelligence-now-features-image-playground-genmoji-and-more/\nhttps://support.apple.com/en-sg/guide/mac-help/mchld5412d00/mac\nhttps://www.apple.com/newsroom/2026/06/apple-intelligence-brings-powerful-ai-capabilities-into-everyday-experiences/\nhttps://www.macrumors.com/2026/06/08/apple-revamps-image-playground/\nhttps://www.apple.com/newsroom/2025/06/apple-intelligence-gets-even-more-powerful-with-new-capabilities-across-apple-devices/\nhttps://support.apple.com/guide/keynote/add-an-image-tan77aea6844/mac\nhttps://www.apple.com/newsroom/images/2024/12/apple-intelligence-now-features-genmoji-image-playground-and-more/article/Apple-Intelligence-Keynote-and-Messages_big.jpg.large.jpg", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "43/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-12", "Sources": [ { "Source ID": "N15-APPLE", "Operation IDs": "12", "Citation": "Apple macOS27 Image Playground guide", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Create an image; person/photo note; style and gallery editing", "Finding used in this review": "Descriptions/concepts and optional selected people/photos condition new images. Selected-person identity can override conflicting text; up to3people are allowed. Gallery edits form another route.", "Evidence limits": "Identity conditioning is not pixel protection or proof of layout guidance. No hidden image-model architecture inferred.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://support.apple.com/en-gb/guide/mac-help/mchld5412d00/27/mac/27" }, { "Source ID": "N15-KEYNOTE", "Operation IDs": "12", "Citation": "Apple Keynote Add an image", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Use Image Playground; Add to Playground", "Finding used in this review": "Keynote invokes Image Playground with descriptions/concepts or selected slide text/images.", "Evidence limits": "Keynote insertion does not remove the generation dependency; Pages/Freeform use the same system feature but app-specific controls can differ.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://support.apple.com/guide/keynote/add-an-image-tan77aea6844/mac" } ] }, { "ID": "13", "Software": "Archicad", "Operation": "AI Visualizer", "Proposed X": "3.7", "Review status": "Proposed X calculated", "Previous X": "3.5–4.25", "S": "5", "U": "4", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "Current AI Visualizer3.0 generation using an Archicad model-view snapshot as the composition guide plus high-level material/style text. Assess the generated still, before Edit or Upscale. Returned result: Generated architectural visualization", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Model geometry/view, captured composition, prompt, model option and acceptance.", "Learned determination": "The selected image model interprets the supplied structure and synthesizes appearance/context throughout the visualization.", "Conventional processing": "Archicad captures the view and saves the returned image.", "Model contribution in the result": "The image retains conditioned synthesis; supplied composition constrains structure without documented exact geometry/pixel protection.", "Why this scope score (S)": "S5: the generated image can determine content/appearance across its domain with no protected/reinstated source complement established.", "Why this latitude score (U)": "U4: current Generate-tab help explicitly says image prompts guide composition through detected edges/colors/textures. This supplies a structural guide, unlike a high-level text-only request.", "Why this coordination score (J)": "J4: learned synthesis resolves architectural semantic/spatial relationships in a still visualization, not evolving new content across time/views.", "Comparisons and other modes": "Calibration65 Canny and ID11 image-to-mesh: Tie in coarse profile. All construct new output with an explicit structural reference and no established protected output region.\nThe model-view configuration is proposed at X 3.7. The separately assessed high-level text-only route, without a structural guide, is X 2.9 (5/5/4). Its lower X reflects construction without the supplied view’s structural guide.\nChallenge: Historical2.0 pairs/model names might be treated as current3.0 proof or geometry resemblance as protection. Checked: Transition plan and current Archicad30 Generate/model-selector help. Result: Use current functional documentation. Historical pair observations remain history; model family names do not prove identical weights or copying.\nSeparate text-only configuration:5/5/4, X2.9; full assessment/calculation retained in review history.", "Citations": "Graphisoft AI Visualizer transition plan — https://community.graphisoft.com/t5/Graphisoft-Insights/AI-Visualizer-Service-Update-and-Transition-Plan/ba-p/712925\nGraphisoft Archicad30 Generate tab — https://help.graphisoft.com/AC/30/INT/Topics/100_Visualization/ai_visualizer_generate_tab.html\nGraphisoft current AI Visualizer help — https://help.graphisoft.com/AC/30/INT/Topics/100_Visualization/generating_ai_images.html", "Source IDs": "N15-GSUPDATE\nN15-GSGEN\nN15-GSMAIN", "Supplemental findings": "Recent indexed discussion mostly concerns access/licensing or broad product reactions; older Stable Diffusion posts are version-mismatched. Current help resolves composition use.", "Evidence limits": "The original record’s2.0 scope has been superseded explicitly by3.0 documentation. Text-only, inpainting, tiling and upscaling are distinct modes, with the extra text-only assessment saved separately.", "Previous record: scope": "Archicad 29 AI Visualizer 2.0 generates a visualization from a composition (the 3D window view or an imported image) and a required text prompt through Graphisoft's cloud service.", "Previous record: rationale": "Working v5 coding with 5 Oct 2026 output evidence for AI Visualizer 2.0: 2 kept (content, layout and viewpoint), 2 model-decided (surface appearance, fine detail) and 1 unknown (shape) among 5 applicable. The recorded uncertainty gives 3.5–4.25; no midpoint is used.", "Previous record: calculation": "K=2; M=2; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 2 ÷ (5 − 1) = 3.5. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 5 Oct 2026 X 2.75–5.0 → 3.5–4.25. Version: AI Visualizer 2.0 in Archicad 29 (hotfix 29.2.1, 28 Jul 2026). Graphisoft's transition post (7 Sep 2026) states that AI Visualizer 3.0 arrives with Archicad 30 and that 2.0 updates to 3.0 automatically; 1.0 (Archicad 28) switches off on 9 Oct 2026. Model: the AI Solutions licence lists Stable Diffusion XL, Google Gemini and Qwen Image Edit across AI Solutions without assigning them to functions. Audit: all 13 v5 quotations are present at their sources; they describe three versions (the Archicad 27 add-on with a shape-fidelity slider, 1.0 with a sunset-lighting default, and the 2.0 guide). Evidence: four 2.0 input and output pairs from the 2.0 guide and Graphisoft community posts, coded by two independent coders (92 percent pair-level agreement). Content and layout and viewpoint are kept in all four; surface appearance and fine detail are model in all four; shape is kept in three and changed in one (a louvred panel replaced by stone wall). Creativity 5 and Creativity 25 give the same codes. Previous scope: \"Archicad AI Visualizer uses the documented Stable Diffusion route to synthesize a visualization from a model view.\" Limits: four pairs; inputs shown as panel thumbnails 163–335 px wide. Re-check when AI Visualizer 3.0 ships. Records: 06_research_archive/2026-10-05_op13_op15_evidence. Graphisoft: AI Visualizer transition plan • AI Visualizer 2.0 guide • Graphisoft AI Solutions licence\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Stable Diffusion route). Designer controls: prompt; creativity slider; image size; style; iterations; prompt strength; Lock Result; Upscale Image. Returns one result; a flat image. “AI Visualizer will always generate one image at a time.” Source • Screenshot", "Collected source links": "https://www.graphisoft.com/solutions/innovation/archicad-ai-visualizer/\nhttps://help.graphisoft.com/AC/28/INT/_AC28_Help/100_Visualization/100_Visualization-10.htm\nhttps://community.graphisoft.com/t5/Graphisoft-Insights/AI-Visualizer-Service-Update-and-Transition-Plan/ba-p/712925\nhttps://community.graphisoft.com/gowhw84577/attachments/gowhw84577/hidden-forum/16/2/AI%20Visualizer%202.1.pdf\nhttps://www.graphisoft.com/en-us/company/legal/license-agreements/ai-solutions-eula/", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "56/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-13", "Sources": [ { "Source ID": "N15-GSUPDATE", "Operation IDs": "13", "Citation": "Graphisoft AI Visualizer transition plan", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "September2026: retirement and3.0 transition", "Finding used in this review": "Visualizer1.0 retires October9; Archicad29 Visualizer2.0 automatically changes to3.0 with Archicad30.", "Evidence limits": "Historical 2.0 output pairs do not establish 3.0 behavior.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://community.graphisoft.com/t5/Graphisoft-Insights/AI-Visualizer-Service-Update-and-Transition-Plan/ba-p/712925" }, { "Source ID": "N15-GSGEN", "Operation IDs": "13", "Citation": "Graphisoft Archicad30 Generate tab", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Image Prompts; Prompt; Aspect Ratio; no-image route", "Finding used in this review": "Current view snapshots/imported images guide composition through edges, colors and textures. With no image prompt, generation uses text/settings alone.", "Evidence limits": "Guidance influences composition; exact pixel or geometry protection is not documented.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://help.graphisoft.com/AC/30/INT/Topics/100_Visualization/ai_visualizer_generate_tab.html" }, { "Source ID": "N15-GSMAIN", "Operation IDs": "13", "Citation": "Graphisoft current AI Visualizer help", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "AI-model dropdown; Generate/Edit/Save steps", "Finding used in this review": "Current help documents a model choice including Nano Banana models, image generation and editing, mask-supported editing and upscaling.", "Evidence limits": "An exact backend version cannot be inferred solely from the family name. Original model-view generation remains distinguishable from added text-only/edit/upscale routes.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://help.graphisoft.com/AC/30/INT/Topics/100_Visualization/generating_ai_images.html" } ] }, { "ID": "14", "Software": "Archicad", "Operation": "Parametric Building Element Modeling", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "Native Archicad Wall Tool construction: straight segment from clicked endpoints, specified geometry/material/reference-line/story settings; exclude AI Visualizer and MCP authoring. Returned result: Editable parametric BIM wall", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Element type/settings, endpoints, reference line and story/connection constraints.", "Learned determination": "None in this native modeling route.", "Conventional processing": "Constructs parametric geometry and evaluates prescribed building-element relationships.", "Model contribution in the result": "The result retains explicit settings and conventional dependencies.", "Why this scope score (S)": "S0: clicked wall endpoints and selected native element/reference-line/story settings prescribe the BIM construction; no fitted predictor is established in this route.", "Why this latitude score (U)": "U0: parametric wall construction and updates apply those specified rules, without an in-scope learned prediction task.", "Why this coordination score (J)": "J0: story links and wall junctions are conventional dependencies. Their coordinated updates do not establish learned context integration.", "Comparisons and other modes": "Calibration114 Fusion specified extrusion: Tie. Both build geometry from human-specified profiles/parameters/rules.\nNative wall construction is proposed at X 15, above AI Visualizer’s model-view configuration (ID13, X 3.7). The wall follows the designer’s endpoints and element settings; Visualizer synthesizes a new image guided by a model view.\nChallenge: Automatic joins or the presence of AI elsewhere in Archicad could be imported into this operation. Checked: Native wall instructions and February2026 composite-wall discussion. Result: Conventional junction priorities remain explicit computer computation; they do not establish learned prediction.", "Citations": "Graphisoft Create a Straight Wall — https://help.graphisoft.com/AC/29/INT/_AC29_Help/040_ElementsVB/040_ElementsVB-8.htm", "Source IDs": "N15-WALL", "Supplemental findings": "February22 2026 discussion concerns reference-line and material/junction priorities. It adds control context and does not change the retained-learning profile.", "Evidence limits": "", "Previous record: scope": "Archicad evaluates manually specified building elements and parametric relationships.", "Previous record: rationale": "Native parametric construction supplies no trained predictor in this scope; AI Visualizer is a separate operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: Wall Tool; click endpoints; Info Box geometry method; Tool Settings (Geometry and Positioning, structure, thickness, reference line); Top Link to story. Returns one result; editable structure: BIM building elements. “In both Floor Plan and 3D, you define the length of Wall segments by clicking at their endpoints.” Source • Screenshot", "Collected source links": "https://www.graphisoft.com/solutions/archicad\nhttps://help.graphisoft.com/AC/28/INT/_AC28_Help/040_ElementsVB/040_ElementsVB-1.htm\nhttps://help.graphisoft.com/AC/29/INT/_AC29_Help/040_ElementsVB/040_ElementsVB-8.htm\nhttps://help.graphisoft.com/AC/29/INT/_AC29_Help/150_UserInterfaceToolSettings/150_UserInterfaceToolSettings-3.htm", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-14", "Sources": [ { "Source ID": "N15-WALL", "Operation IDs": "14", "Citation": "Graphisoft Create a Straight Wall", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Straight Wall geometry method and endpoint definition", "Finding used in this review": "The designer selects the wall method/settings and clicks endpoints; native geometry constructs and stretches the building element.", "Evidence limits": "Scope is manual native element construction, not MCP or AI Visualizer.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://help.graphisoft.com/AC/29/INT/_AC29_Help/040_ElementsVB/040_ElementsVB-8.htm" } ] }, { "ID": "15", "Software": "ArkoAI (in Rhino, SketchUp, Autodesk Revit)", "Operation": "Viewport to Image Generation", "Proposed X": "3.7", "Review status": "Proposed X calculated", "Previous X": "Unresolved", "S": "5", "U": "4", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "ArkoAI4.0 viewport-to-image, active supplied model view as the structural guide. Account for Basic Render, Basic Ideate and Pro(BANANA dispatch); no separate masked edit/upscale included. Returned result: Generated architectural image", "Next step": "Review this configuration-specific v2 proposal and its evidence. Final method adoption remains open.", "Human input and controls": "Native geometry/view, prompt/style settings, Basic/Pro/mode and output acceptance.", "Learned determination": "Hosted synthesis interprets viewport structure and determines image appearance/context and possible geometry changes.", "Conventional processing": "Host captures/submits the view; frontend dispatches mode/options and displays the returned image.", "Model contribution in the result": "Viewport-conditioned generation remains. Render/Pro retention claims do not establish exact copying; Ideate can vary geometry while still using the view as a guide.", "Why this scope score (S)": "S5: all three construction paths return an image generated across the output without documented protected source reinstatement.", "Why this latitude score (U)": "U4: the assessed paths use the active model view to guide identifiable structural relationships. Permitting geometry variation does not remove a guide, just as non-hard Canny conditioning can vary from its edges.", "Why this coordination score (J)": "J4: the task constructs semantic/spatial architectural relationships in a still image. Frontend modes/model names do not support J5.", "Comparisons and other modes": "Calibration65 Canny and ID13 Visualizer model view: Tie for guided whole-image construction. Different conditioning strength and geometry fidelity remain in configuration notes; the frozen integer anchors do not quantify those differences.\nArkoAI’s guided viewport configuration is proposed at X 3.7, above the high-level text-only generation comparison at X 2.9. The supplied viewport guides structural relationships even when the generated result departs from them.\nChallenge: Basic Ideate’s geometry changes might implyU5, while Pro’s retention label might implyS4. Checked: Current4.0 public dispatch code and active-view product page; historical Render/Ideate definitions checked only as related context. Result: A guide need not hard-lock geometry; a retention claim is not source reinstatement. Keep5/4/4 for documented guided paths. If an actual backend path ignores the view, it requires a new5/5/4 assessment.", "Citations": "ArkoAI current public plugin code — https://app.arko.ai/\nArkoAI Revit4.0 product page — https://arko.ai/revit", "Source IDs": "N15-ARKO\nN15-ARKOPRO", "Supplemental findings": "No matching recent Reddit results. Vendor forum results describe2023–24 releases; they are historical context. Current frontend confirms three branches; the4.0 announcement currently returns generic content.", "Evidence limits": "Backend models are undisclosed beyond the product’s Pro engine label, and no particular Google model version is assigned. Current modes can tie despite different fidelity; no scalar strength fraction is invented. The claimed active-view configuration does not verify every strength endpoint or hidden backend behavior.", "Previous record: scope": "ArkoAI produces a generated visualization through its hosted viewport-to-image operation.", "Previous record: rationale": "Current placement unresolved. Version 4.0 has three request configurations (Basic + Render, Basic + Ideate, Pro). 5 Oct 2026 output evidence gives Pro 2.75–5.0 from current pairs. The Basic configurations have no current pairs; 2023 Render-mode and Ideate-mode pairs give 2.75–3.5 and 2.0–2.75. With all configurations combined, surface appearance is model and the other four categories are unknown.", "Previous record: calculation": "Operation: lower 2.0; upper undefined while the 4.0 Basic configurations lack current evidence (with the 2023 Basic-mode evidence: 2.0–5.0). Pro: K=1; M=1; unknown=3; A=5; 2.75–5.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented by ArkoAI. One third-party roundup states that ArkoAI release 4.0 (March 2026) added Google's Nano Banana model; no ArkoAI page confirming this was found, and the claim may conflate ArkoAI with Chaos Veras 4.0, which Chaos documents as Nano Banana-powered. Model choice: Not documented. Not established: Which model ArkoAI calls for viewport-to-image renders, before and after any 4.0 release; whether an ArkoAI 4.0 release exists (arko.ai/f/release-40 returned HTTP 404); whether users choose a render engine or model. arko.ai, page date not shown • syntina.ai • aecmag.com\nChecked 5 Oct 2026 X unchanged (Unresolved; lower bound 2.0); configuration values added. Version: ArkoAI 4.0.0 (Rhino package 10 Mar 2026; announcement 11 Mar 2026 at arko.ai/f/arkoai-40-is-here-say-hello-to-nano-banana, which opens on 5 Oct 2026; the 4 Oct note's address /f/release-40 returns 404). Configurations from the plugin's public code (app.arko.ai): Basic + Render, Basic + Ideate (the default) and Pro, which sends request mode BANANA whichever mode button is active. Basic: model not named; the request carries guidance scale, prompt strength, seed and a fixed negative prompt. Pro: named \"Nano Banana engine\" by ArkoAI; the Google model version is not stated. Google's Gemini 3 Pro Image and 3.1 Flash Image model cards list infrequent copying from the input image and state no preservation of unedited content. Audit: all 11 v5 quotations are present; 5 describe 2023 versions. Evidence: 27 pairs coded by two independent coders (87 percent pair-level agreement). Pro, five current pairs: content kept; surface appearance model; layout and viewpoint, shape and fine detail unknown (eye level raised in one pair; framing 1.25–1.45 times tighter in two). 2023 Render mode, seven pairs: content kept; layout and viewpoint model; shape unknown; surface appearance and fine detail model. 2023 Ideate mode, six pairs: content unknown; the other four model. The 2023 outputs are uniformly enlarged by about 4.5 percent, coded as a framing change. Records: 06_research_archive/2026-10-05_op13_op15_evidence. ArkoAI 4.0 announcement • ArkoAI plugin page (public code) • Gemini 3 Pro Image model card\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (hosted image model). Designer controls: Render or Ideate mode; prompt; negative prompt; prompt importance; difference from original; style options (style, environment, lighting); seed; image quality. Returns count not documented; a flat image. “Allows you to maintain the integrity of your 3D geometry while rendering the context and materials.” Source • Screenshot Quote taken from a search result for the vendor page.", "Collected source links": "https://arko.ai/\nhttps://arko.ai/sketchup\nhttps://arko.ai/f/release-20\nhttps://syntina.ai/en/blog/best-ai-architectural-rendering-tools\nhttps://aecmag.com/visualisation/chaos-veras-4-0-powered-by-nano-banana/\nhttps://arko.ai/f/arkoai-40-is-here-say-hello-to-nano-banana\nhttps://www.youtube.com/watch?v=c9JB2Hqv0PU\nhttps://app.arko.ai/\nhttps://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Image-Model-Card.pdf\nhttps://forum.arko.ai/t/arkoai-release-2-0/1473", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "56/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "next15-20261014-15", "Sources": [ { "Source ID": "N15-ARKO", "Operation IDs": "15", "Citation": "ArkoAI current public plugin code", "Type": "primary code", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "currentRenderMode; rendering request branch; getModelSelectValue; BANANA/IDEATE/RENDER", "Finding used in this review": "Public UI code preserves Basic Render, Basic Ideate and Pro branches. Pro sends BANANA regardless of visible mode buttons. Prompt, seed, difference and active view travel through the host-render request.", "Evidence limits": "Frontend dispatch does not disclose backend weights or prove hard geometry protection.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://app.arko.ai/" }, { "Source ID": "N15-ARKOPRO", "Operation IDs": "15", "Citation": "ArkoAI Revit4.0 product page", "Type": "primary help", "Source date": "See locator", "Accessed": "2026-10-14", "Locator": "Basic versus Pro; render directly from Revit", "Finding used in this review": "Current4.0 page describes active-view AI rendering; Pro is presented as geometry-retaining and Basic as better for ideation.", "Evidence limits": "Geometry-retention marketing is evidence of a structural-guide task, not exact copying or a guarantee.", "Review status": "read; direct", "Open source": "Open source", "URL": "https://arko.ai/revit" } ] }, { "ID": "16", "Software": "Autodesk Arnold", "Operation": "Open Image Denoise Pass", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2fe6957a74bf\ncollected-4e11fb7b9e5b\ncollected-caf577624d1d\ncollected-5b86c2d9e423\ncollected-12ff45981523\ncollected-f2d1d51f3abc\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Arnold's Open Image Denoise imager predicts a clean render from noisy render data.", "Previous record: rationale": "Trained noise removal matches fidelity restoration. Albedo and normal guidance describe execution and add no score adjustment.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). Arnold 7.5.2.0 (18 June 2026) keeps the OIDN imager beside the OptiX imager; the latest Open Image Denoise release is v2.5.1 (18 August 2026), a CNN with pre-trained models whose training toolkit pairs noisy renders with noise-free references and defaults to an L1 plus MS-SSIM loss. Intel has announced OIDN 3 for the second half of 2026, moving from direct prediction to kernel prediction; it is unreleased as of 3 October 2026. Not established: Data and loss for the shipped pre-trained weights are not stated apart from the toolkit default. Whether and when Arnold adopts OIDN 3 kernel prediction is not documented. openimagedenoise.org, accessed 2026-10-03 • github.com, v2.5.1 • github.com, master branch, accessed 2026-10-03", "Collected source links": "https://help.autodesk.com/cloudhelp/ENU/AR-Core/files/ac-denoising/arnold_user_guide_ac_denoising_ac_denoiser_oidn_html.html\nhttps://www.openimagedenoise.org/documentation.html\nhttps://github.com/RenderKit/oidn/blob/master/README.md\nhttps://github.com/RenderKit/oidn/blob/master/training/config.py\nhttps://help.autodesk.com/cloudhelp/ENU/AR-Maya/files/am-Arnold_for_Maya_User_Guide/render-settings/arnold/arnold_for_maya_arnold_am_Imagers_html.html\nhttps://help.autodesk.com/cloudhelp/ENU/AR-Maya/files/am-Arnold_for_Maya_User_Guide/rendering/arnold_for_maya_rendering_am_Arnold_RenderView_Window_html.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2fe6957a74bf", "Operation IDs": "16", "Citation": "D1 Autodesk — Open Image Denoise Imager", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/AR-Core/files/ac-denoising/arnold_user_guide_ac_denoising_ac_denoiser_oidn_html.html" }, { "Source ID": "collected-4e11fb7b9e5b", "Operation IDs": "16", "Citation": "D2 Intel — Open Image Denoise Documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.openimagedenoise.org/documentation.html" }, { "Source ID": "collected-caf577624d1d", "Operation IDs": "16, 30", "Citation": "github.com, v2.5.1", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/RenderKit/oidn/blob/master/README.md" }, { "Source ID": "collected-5b86c2d9e423", "Operation IDs": "16", "Citation": "github.com, master branch, accessed 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/RenderKit/oidn/blob/master/training/config.py" }, { "Source ID": "collected-12ff45981523", "Operation IDs": "16", "Citation": "help.autodesk.com — arnold_for_maya_arnold_am_Imagers_html.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/AR-Maya/files/am-Arnold_for_Maya_User_Guide/render-settings/arnold/arnold_for_maya_arnold_am_Imagers_html.html" }, { "Source ID": "collected-f2d1d51f3abc", "Operation IDs": "16", "Citation": "help.autodesk.com — arnold_for_maya_rendering_am_Arnold_RenderView_Window_html.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/AR-Maya/files/am-Arnold_for_Maya_User_Guide/rendering/arnold_for_maya_rendering_am_Arnold_RenderView_Window_html.html" } ] }, { "ID": "17", "Software": "Autodesk Arnold", "Operation": "Path Tracing Final Image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-74bbb0d47ec9\ncollected-2ca1ef651b89\ncollected-f7f32ed37975", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Arnold path tracing evaluates light transport for a defined scene, with learned denoising excluded.", "Previous record: rationale": "Monte Carlo estimation is computational inference in a broad sense but contains no trained predictor in this scope. The current X convention places this conventional operation at 15.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a solver or simulation. Designer controls: Camera (AA) samples; Diffuse samples; Specular samples; Transmission samples; Volume Indirect samples. Returns one result; a flat image. “The higher the number of samples, the better the anti-aliasing quality, and the longer the render times.” Source", "Collected source links": "https://help.autodesk.com/cloudhelp/2024/ENU/AR-Maya/files/am-Arnold_for_Maya_User_Guide/getting-started/arnold_for_maya_getting_started_am_Arnold_html.html\nhttps://help.autodesk.com/cloudhelp/ENU/AR-Core/files/ac-faq/arnold_user_guide_ac_faq_ac_general_faq_html.html\nhttps://help.autodesk.com/cloudhelp/ENU/AR-Core/files/ac-render-settings/ac-sampling/arnold_user_guide_ac_render_settings_ac_samples_html.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-74bbb0d47ec9", "Operation IDs": "17", "Citation": "Autodesk · Arnold Monte Carlo path tracing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2024/ENU/AR-Maya/files/am-Arnold_for_Maya_User_Guide/getting-started/arnold_for_maya_getting_started_am_Arnold_html.html" }, { "Source ID": "collected-2ca1ef651b89", "Operation IDs": "17", "Citation": "B37-A1 Autodesk Arnold general FAQ, path tracing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/AR-Core/files/ac-faq/arnold_user_guide_ac_faq_ac_general_faq_html.html" }, { "Source ID": "collected-f7f32ed37975", "Operation IDs": "17", "Citation": "help.autodesk.com — arnold_user_guide_ac_render_settings_ac_samples_html.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/AR-Core/files/ac-render-settings/ac-sampling/arnold_user_guide_ac_render_settings_ac_samples_html.html" } ] }, { "ID": "18", "Software": "Autodesk AutoCAD", "Operation": "3D Solid Face Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b463849e2c92\ncollected-1ff2babc6e0d\ncollected-acb9a4a38d96", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "AutoCAD SOLIDEDIT offsets selected solid faces according to a specified distance or point.", "Previous record: rationale": "Geometric face modification is prescribed computation; a complex boundary does not imply learned determination.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: SOLIDEDIT Face Offset option; select faces; offset distance or through point. Returns one result; editable structure: 3D solid. “Offsets faces equally by a specified distance or through a specified point.” Source", "Collected source links": "https://www.autodesk.com/eu/products/autocad/features\nhttps://help.autodesk.com/view/ACD/2026/ENU/?caas=caas%2Fdocumentation%2FCIV3D%2F2014%2FENU%2FfilesACD%2FGUID-D54C266B-2B68-4660-ACA5-0579432F149C-htm.html\nhttps://help.autodesk.com/cloudhelp/2026/ENU/AutoCAD-Core/files/GUID-D54C266B-2B68-4660-ACA5-0579432F149C.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b463849e2c92", "Operation IDs": "18", "Citation": "Autodesk · AutoCAD overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/eu/products/autocad/features" }, { "Source ID": "collected-1ff2babc6e0d", "Operation IDs": "18", "Citation": "Autodesk · SOLIDEDIT 3D face editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/view/ACD/2026/ENU/?caas=caas%2Fdocumentation%2FCIV3D%2F2014%2FENU%2FfilesACD%2FGUID-D54C266B-2B68-4660-ACA5-0579432F149C-htm.html" }, { "Source ID": "collected-acb9a4a38d96", "Operation IDs": "18", "Citation": "help.autodesk.com — GUID-D54C266B-2B68-4660-ACA5-0579432F149C.htm", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/AutoCAD-Core/files/GUID-D54C266B-2B68-4660-ACA5-0579432F149C.htm" } ] }, { "ID": "19", "Software": "Autodesk AutoCAD", "Operation": "Markup Assist to Native Drawing Objects", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-83318bb94cb8\ncollected-6c980987d555\ncollected-6381329fbf0a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Markup Assist recognizes written annotations and converts them into native drawing text or leaders.", "Previous record: rationale": "Recognition estimates structured content, between region determination and fidelity reconstruction. Public task evidence does not support a more precise position within 8–10.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). AutoCAD help states that Markup Assist uses machine learning to convert markups into AutoCAD geometry; the 2026 update lets several detected text markups go into one mtext or mleader insertion and sizes revision clouds from the detected markup border. The documented learned output is detection and text recognition of existing markups, and inserted objects carry the recognized text or detected outline; no generated content is documented. Not established: Model type and training data are not published, and changes in AutoCAD 2027 are unconfirmed: the 2027 What's New index returned 404 and Autodesk blog pages refused fetch. Documentation does not place recognition of markups within 8–10. help.autodesk.com, AutoCAD 2025 • help.autodesk.com, AutoCAD 2026", "Collected source links": "https://help.autodesk.com/view/ACD/2023/ENU/?guid=GUID-64CFD65E-ABD0-49A8-9218-D4E2D22BC070\nhttps://help.autodesk.com/cloudhelp/2025/ENU/AutoCAD-Subscription/files/GUID-75DC6F68-0CCA-4968-AB81-74B38B335B1E.htm\nhttps://help.autodesk.com/cloudhelp/2026/ENU/AutoCAD-WhatsNew/files/GUID-4811E96E-4CF9-4A02-B3A0-2C9080BBB57C.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-83318bb94cb8", "Operation IDs": "19", "Citation": "Autodesk · Markup Assist", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/view/ACD/2023/ENU/?guid=GUID-64CFD65E-ABD0-49A8-9218-D4E2D22BC070" }, { "Source ID": "collected-6c980987d555", "Operation IDs": "19", "Citation": "Autodesk · Markup Assist recognition and CAD conversion", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2025/ENU/AutoCAD-Subscription/files/GUID-75DC6F68-0CCA-4968-AB81-74B38B335B1E.htm" }, { "Source ID": "collected-6381329fbf0a", "Operation IDs": "19", "Citation": "help.autodesk.com, AutoCAD 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/AutoCAD-WhatsNew/files/GUID-4811E96E-4CF9-4A02-B3A0-2C9080BBB57C.htm" } ] }, { "ID": "20", "Software": "Autodesk AutoCAD", "Operation": "2027 Smart Blocks Detect and Convert", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec9b990b2e74\ncollected-277237ebcb16\ncollected-2855dd88f3ad\ncollected-a5cd4a4e6a58", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Smart Blocks Detect and Convert identifies similar object groups before creating native block definitions.", "Previous record: rationale": "Learned grouping is the retained determination; native conversion transmits it. Match region determination without awarding points for editable blocks.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). The AutoCAD 2027 help states that Detect and Convert uses Autodesk AI to identify and group similar objects for block conversion; the 2025 tech-preview help states that drawing geometry goes to an AutoCAD machine learning service. The task remains detection and grouping; no generative step is documented. Not established: Model type, training data, and whether 2027 uses the same service as the 2025 preview are undisclosed. help.autodesk.com, AutoCAD 2027 • help.autodesk.com, AutoCAD 2025", "Collected source links": "https://help.autodesk.com/view/ACD/2025/ENU/?guid=GUID-8D87AF53-BD0C-4E01-A015-42E92E6AA715\nhttps://help.autodesk.com/cloudhelp/2024/ENU/AutoCAD-LT/files/GUID-8D87AF53-BD0C-4E01-A015-42E92E6AA715.htm\nhttps://help.autodesk.com/cloudhelp/2027/ENU/AutoCAD-MAC-Core/files/GUID-B3E5C654-612C-4395-B3A8-7316DB678465.htm\nhttps://help.autodesk.com/cloudhelp/2025/ENU/AutoCAD-WhatsNew/files/GUID-25BD5FB7-119A-42A8-B1C1-62BB812A3F4F.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec9b990b2e74", "Operation IDs": "20", "Citation": "Autodesk · Smart Blocks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/view/ACD/2025/ENU/?guid=GUID-8D87AF53-BD0C-4E01-A015-42E92E6AA715" }, { "Source ID": "collected-277237ebcb16", "Operation IDs": "20", "Citation": "Autodesk · Smart Blocks placement and replacement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2024/ENU/AutoCAD-LT/files/GUID-8D87AF53-BD0C-4E01-A015-42E92E6AA715.htm" }, { "Source ID": "collected-2855dd88f3ad", "Operation IDs": "20", "Citation": "help.autodesk.com, AutoCAD 2027", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/AutoCAD-MAC-Core/files/GUID-B3E5C654-612C-4395-B3A8-7316DB678465.htm" }, { "Source ID": "collected-a5cd4a4e6a58", "Operation IDs": "20", "Citation": "help.autodesk.com, AutoCAD 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2025/ENU/AutoCAD-WhatsNew/files/GUID-25BD5FB7-119A-42A8-B1C1-62BB812A3F4F.htm" } ] }, { "ID": "21", "Software": "Autodesk Forma", "Operation": "Detailed Noise Analysis Visual Map", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-dd6f2f9ad8d5\ncollected-bd0336649984\ncollected-1e13566c5143\ncollected-502cf9a0e764", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Forma Detailed Noise Analysis computes a visual noise map through the documented detailed propagation simulation.", "Previous record: rationale": "The selected detailed solver is conventional. A predicted environmental quantity and a complex visualization do not by themselves establish trained inference.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a solver or simulation. Designer controls: roads and traffic data (database or own data); traffic speed; traffic volume. Returns one result; form not documented. “The noise calculations in Forma analysis are in accordance with the Common NOise aSSessment methOdS in Europe (CNOSSOS-EU) developed by the European Commission.” Source • Screenshot", "Collected source links": "https://www.autodesk.com/products/forma-site-design/environmental-impact-analysis\nhttps://help.autodesk.com/cloudhelp/ENU/Forma-SD-Get-Started/files/Getting_Started_SD.html\nhttps://blogs.autodesk.com/forma/2023/09/18/detailed-noise-analysis-helps-architects-do-groundwork-for-noise-assessments/\nhttps://blogs.autodesk.com/forma/wp-content/uploads/sites/163/2023-09-18-09_43_51-Noise-analysis-detailed-1-1024x576.jpg", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-dd6f2f9ad8d5", "Operation IDs": "21", "Citation": "Autodesk · Forma Detailed Analyses", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/products/forma-site-design/environmental-impact-analysis" }, { "Source ID": "collected-bd0336649984", "Operation IDs": "21", "Citation": "Autodesk · Forma Site Design massing workspace", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/Forma-SD-Get-Started/files/Getting_Started_SD.html" }, { "Source ID": "collected-1e13566c5143", "Operation IDs": "21", "Citation": "blogs.autodesk.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blogs.autodesk.com/forma/2023/09/18/detailed-noise-analysis-helps-architects-do-groundwork-for-noise-assessments/" }, { "Source ID": "collected-502cf9a0e764", "Operation IDs": "21", "Citation": "blogs.autodesk.com — 2023-09-18-09_43_51-Noise-analysis-detailed-1-1024x576.jpg", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blogs.autodesk.com/forma/wp-content/uploads/sites/163/2023-09-18-09_43_51-Noise-analysis-detailed-1-1024x576.jpg" } ] }, { "ID": "22", "Software": "Autodesk Forma", "Operation": "Rapid Noise Analysis Visual Map", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-566ac9755bc1\ncollected-446a62e041fd\ncollected-926ee8ed1e39", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Forma Rapid Noise Analysis uses a trained predictor to approximate detailed noise analysis as a visual site map.", "Previous record: rationale": "A learned surrogate for a defined analysis function matches fixed-operator emulation. Its scientific subject matter does not make it generative image synthesis.", "Previous record: calculation": "X = 12 + 0 × (12 − 12) = 12.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (12.0). Autodesk describes Rapid Noise Analysis as a neural network surrogate trained on tens of thousands of detailed noise simulations; 2026 Forma Site Design materials still list rapid and detailed noise options. The help article cited in the record now redirects to the Forma login page. Not established: Model architecture and the current training set are undisclosed. blogs.autodesk.com, 2023-07-03 • autodesk.com, 2026", "Collected source links": "https://blogs.autodesk.com/forma/2023/06/06/formas-rapid-wind-analysis-offers-instant-insights-into-wind-conditions/\nhttps://blogs.autodesk.com/forma/2023/07/03/game-changing-rapid-noise-analysis-comes-to-forma-2/\nhttps://www.autodesk.com/learn/ondemand/curated/forma-quick-start-guide/2NXymFoWSdr487mpHTw4In", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-566ac9755bc1", "Operation IDs": "22", "Citation": "Autodesk · Forma Rapid Analyses", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blogs.autodesk.com/forma/2023/06/06/formas-rapid-wind-analysis-offers-instant-insights-into-wind-conditions/" }, { "Source ID": "collected-446a62e041fd", "Operation IDs": "22", "Citation": "Autodesk · Forma neural rapid noise analysis", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blogs.autodesk.com/forma/2023/07/03/game-changing-rapid-noise-analysis-comes-to-forma-2/" }, { "Source ID": "collected-926ee8ed1e39", "Operation IDs": "22", "Citation": "autodesk.com, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/learn/ondemand/curated/forma-quick-start-guide/2NXymFoWSdr487mpHTw4In" } ] }, { "ID": "23", "Software": "Autodesk Forma", "Operation": "Site Massing Object Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-901199ef907e\ncollected-f2b47b7fce5d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Forma direct massing edits create and modify cuboids, faces and edges with specified dimensions.", "Previous record: rationale": "Conventional geometric modeling contains no trained predictor within this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by the designer’s marks. Designer controls: Cuboid tool; Line tool; Arc by three points; Move tool; face extrude; edge selection; typed dimensions. Returns one result; editable structure: 3D massing volumes. “Select a face and click again to extrude it by 100'.” Source", "Collected source links": "https://illustrarch.com/articles/design-softwares/73363-autodesk-forma-review.html\nhttps://blogs.autodesk.com/forma/2026/04/21/how-to-create-a-3d-sketch-building-in-forma-site-design/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-901199ef907e", "Operation IDs": "23", "Citation": "illustrarch · Forma Core", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://illustrarch.com/articles/design-softwares/73363-autodesk-forma-review.html" }, { "Source ID": "collected-f2b47b7fce5d", "Operation IDs": "23", "Citation": "B41-M1 Autodesk Forma 3D Sketch building tutorial April 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blogs.autodesk.com/forma/2026/04/21/how-to-create-a-3d-sketch-building-in-forma-site-design/" } ] }, { "ID": "24", "Software": "Autodesk Maya", "Operation": "MayaMCP LLM Scene Object Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c8c9781ea545\ncollected-a744a220692e\ncollected-5fd3cc2c73c6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "A MayaMCP LLM selects native objects and transform attributes; Maya executes the construction.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The model determines which native objects or prescribed procedural models appear and their parameter and transform values, while geometry comes from Maya primitives and the server's prescribed generators; this is the which-and-where kind of determination REF10 makes and places the operation nearer REF10 than REF02. The interval stays open below REF10 because the LLM authors the arrangement from a prompt.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. MayaMCP v0.2.0 (6 commits) exposes predefined tools: create_object (cube, cone, sphere, cylinder, camera, three light types), attribute setting, mesh operations, and parameterized procedural generators (create_advanced_model, generate_scene) that run prescribed Python through Maya's command port. No tool calls a generative model or runs LLM-written code. Not established: No dated release; the LLM in use is chosen by the client. Previous reason: Learned scene decisions remain in native geometry. The source supports the controller-to-tool dependency but not a precise cross-domain position within 2–10. raw.githubusercontent.com, v0.2.0", "Collected source links": "https://github.com/PatrickPalmer/MayaMCP\nhttps://github.com/PatrickPalmer/MayaMCP/blob/5aa3d7b575facf699f88d5334aafa785516bb7e5/src/mayatools/object/create_object.py\nhttps://raw.githubusercontent.com/PatrickPalmer/MayaMCP/main/README.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c8c9781ea545", "Operation IDs": "24", "Citation": "GitHub · MayaMCP", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/PatrickPalmer/MayaMCP" }, { "Source ID": "collected-a744a220692e", "Operation IDs": "24", "Citation": "B42-M1 MayaMCP primitive creation pinned source", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/PatrickPalmer/MayaMCP/blob/5aa3d7b575facf699f88d5334aafa785516bb7e5/src/mayatools/object/create_object.py" }, { "Source ID": "collected-5fd3cc2c73c6", "Operation IDs": "24", "Citation": "raw.githubusercontent.com, v0.2.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/PatrickPalmer/MayaMCP/main/README.md" } ] }, { "ID": "25", "Software": "Beautiful.ai", "Operation": "Slide AI image regeneration", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-eddfa67b7de3\ncollected-a7c6b9d939fb\ncollected-9c30dd290bb2\ncollected-ca614cba419d\ncollected-c50e2ccf4e29", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Beautiful.ai Slide AI regenerates one slide image while the surrounding slide remains outside the assessed endpoint.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented for current Slide AI image regeneration. Beautiful.ai's January 2023 DesignerBot launch named OpenAI's DALL-E as the image generator; no later source names the image model. Model choice: No model choice documented. Not established: Whether Slide AI image regeneration still uses DALL-E or moved to GPT Image or another provider. prnewswire.com, 2023-01-11 • beautiful.ai, 2025-12-19\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model. Designer controls: prompt (re-prompt per image); deck AI image style. Returns count not documented; a flat image. “create an image unique to your story” Source • Screenshot Quote taken from a search result for the vendor page.", "Collected source links": "https://support.beautiful.ai/hc/en-us/articles/43350069148557-Create-and-Edit-your-Slides-with-Slide-AI\nhttps://www.beautiful.ai/product-updates\nhttps://www.prnewswire.com/news-releases/beautifulai-launches-designerbot-to-automatically-create-rich-presentations-about-anything-and-everything-using-generative-ai-301718769.html\nhttps://www.beautiful.ai/blog/represent-2025-a-year-in-review\nhttps://cdn.prod.website-files.com/59e16042ec229e00016d3a66/69cfa81f79bbf894975314d6_improved-ai-images.avif", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-eddfa67b7de3", "Operation IDs": "25", "Citation": "support.beautiful.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.beautiful.ai/hc/en-us/articles/43350069148557-Create-and-Edit-your-Slides-with-Slide-AI" }, { "Source ID": "collected-a7c6b9d939fb", "Operation IDs": "25", "Citation": "BA2 Beautiful.ai Product Updates, releases 03172026 and 12152025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.beautiful.ai/product-updates" }, { "Source ID": "collected-9c30dd290bb2", "Operation IDs": "25", "Citation": "prnewswire.com, 2023-01-11", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.prnewswire.com/news-releases/beautifulai-launches-designerbot-to-automatically-create-rich-presentations-about-anything-and-everything-using-generative-ai-301718769.html" }, { "Source ID": "collected-ca614cba419d", "Operation IDs": "25", "Citation": "beautiful.ai, 2025-12-19", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.beautiful.ai/blog/represent-2025-a-year-in-review" }, { "Source ID": "collected-c50e2ccf4e29", "Operation IDs": "25", "Citation": "cdn.prod.website-files.com — 69cfa81f79bbf894975314d6_improved-ai-images.avif", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.prod.website-files.com/59e16042ec229e00016d3a66/69cfa81f79bbf894975314d6_improved-ai-images.avif" } ] }, { "ID": "26", "Software": "Beautiful.ai", "Operation": "Smart Slide photo layout using generated images", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-06403e049baa\ncollected-db32c38699c9", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Beautiful.ai Photo Grid automatically fits and crops supplied photos into a selected layout.", "Previous record: rationale": "The recorded behavior establishes automatic layout but does not identify whether learned saliency or a prescribed fit rule determines the crop. Resolve that mechanism before selecting an equation branch.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Beautiful.ai's Photo Grid help article (search-result text; page refused retrieval) lists per-cell controls to size the photo up or down, set it to fill the cell, edit, add alt text or replace it. Beautiful.ai describes Smart Slides as a patented layout engine applying design rules; no Beautiful.ai page or patent located states how the default crop position inside a cell is chosen. Not established: Whether the default crop is placed by a fixed rule (center or fit/fill) or by face, saliency or other learned detection. support.beautiful.ai", "Collected source links": "https://support.beautiful.ai/hc/en-us/articles/40184565610893-Photo-Grid\nhttps://support.beautiful.ai/hc/en-us/articles/360018624371-How-do-I-add-an-image-to-a-slide", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-06403e049baa", "Operation IDs": "26", "Citation": "PG1 Beautiful.ai Photo Grid, updated 04072026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.beautiful.ai/hc/en-us/articles/40184565610893-Photo-Grid" }, { "Source ID": "collected-db32c38699c9", "Operation IDs": "26", "Citation": "PG2 Beautiful.ai How do I add an image to a slide, consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.beautiful.ai/hc/en-us/articles/360018624371-How-do-I-add-an-image-to-a-slide" } ] }, { "ID": "27", "Software": "Blender", "Operation": "Conventional Geometry Nodes Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-41d75d2f874b\ncollected-55bf4030d2bd\ncollected-682cb211fe49\ncollected-a7e9833e9288", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Blender evaluates a conventional Geometry Nodes graph to construct geometry.", "Previous record: rationale": "The declared primitive, field and instancing nodes implement prescribed geometry functions; no neural add-on is included.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: Geometry Nodes modifier; node graph in Geometry Node editor; node inputs (e.g. Mesh Line, Instance on Points, Object Info). Returns one result; editable structure: node graph and generated geometry. “Geometry Nodes is a system for modifying the geometry of an object with node-based operations.” Source • Screenshot", "Collected source links": "https://docs.blender.org/manual/en/4.5/modeling/geometry_nodes/introduction.html\nhttps://docs.blender.org/manual/en/4.5/modeling/geometry_nodes/mesh/primitives/mesh_line.html\nhttps://docs.blender.org/manual/en/latest/modeling/geometry_nodes/introduction.html\nhttps://docs.blender.org/manual/en/latest/_images/modeling_geometry-nodes_introduction_properties.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-41d75d2f874b", "Operation IDs": "27", "Citation": "docs.blender.org", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/4.5/modeling/geometry_nodes/introduction.html" }, { "Source ID": "collected-55bf4030d2bd", "Operation IDs": "27", "Citation": "docs.blender.org", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/4.5/modeling/geometry_nodes/mesh/primitives/mesh_line.html" }, { "Source ID": "collected-682cb211fe49", "Operation IDs": "27", "Citation": "docs.blender.org — introduction.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/latest/modeling/geometry_nodes/introduction.html" }, { "Source ID": "collected-a7e9833e9288", "Operation IDs": "27", "Citation": "docs.blender.org — modeling_geometry-nodes_introduction_properties.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/latest/_images/modeling_geometry-nodes_introduction_properties.png" } ] }, { "ID": "28", "Software": "Blender", "Operation": "Conventional Image Compositing Graph", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6bec999f5510\ncollected-39b0ec42ab12\ncollected-67a7c3afcc53", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Blender Alpha Over composites two images using the supplied alpha and blend factor.", "Previous record: rationale": "The inspected implementation performs compositing arithmetic. Image provenance does not change this isolated compositor's coordinate.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Factor; background Image input; foreground Image input; Convert Premultiplied. Returns one result; a flat image. “The Alpha Over node is used to layer an image on top of another with alpha blending.” Source • Screenshot", "Collected source links": "https://raw.githubusercontent.com/blender/blender/v4.5.0/source/blender/nodes/composite/nodes/node_composite_alpha_over.cc\nhttps://docs.blender.org/UATEST/manual/en/4.5/compositing/types/color/mix/alpha_over.html\nhttps://docs.blender.org/UATEST/manual/en/4.5/_images/node-types_CompositorNodeAlphaOver.webp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6bec999f5510", "Operation IDs": "28", "Citation": "BL1 Blender v4.5.0 Alpha Over node source, inspected 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/blender/blender/v4.5.0/source/blender/nodes/composite/nodes/node_composite_alpha_over.cc" }, { "Source ID": "collected-39b0ec42ab12", "Operation IDs": "28", "Citation": "BL2 Blender 4.5 LTS Alpha Over node manual, consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/UATEST/manual/en/4.5/compositing/types/color/mix/alpha_over.html" }, { "Source ID": "collected-67a7c3afcc53", "Operation IDs": "28", "Citation": "docs.blender.org — node-types_CompositorNodeAlphaOver.webp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/UATEST/manual/en/4.5/_images/node-types_CompositorNodeAlphaOver.webp" } ] }, { "ID": "29", "Software": "Blender", "Operation": "Correlation Tracking with Editable Markers", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-01bd288a1c7e\ncollected-25d1b92226ad\ncollected-ea9b9910e0b6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Blender tracks a seeded marker through the selected correlation-based libmv tracking route.", "Previous record: rationale": "Conventional pattern matching and numerical optimization do not establish a trained predictor. Automated marker motion remains conventional in this scope.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by search or optimization. Designer controls: placed marker; Pattern Size; Search Size; Motion Model; Match (Keyframe, Previous Frame); Prepass; Normalize; Correlation; Margin; Use Mask; Frames Limit; R/G/B channels; Weight. Returns one result; editable structure: marker track positions per frame. “This value defines the minimal correlation between a matched pattern and a reference to be considered a successful tracking.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.blender.org/releases/4-5/\nhttps://docs.blender.org/manual/en/4.4/movie_clip/tracking/clip/toolbar/track.html\nhttps://docs.blender.org/manual/en/latest/movie_clip/tracking/clip/toolbar/track.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-01bd288a1c7e", "Operation IDs": "29, 31, 32, 33, 34", "Citation": "B45-R Blender 4.5 LTS release history; 4.5.14 dated 09152026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.blender.org/releases/4-5/" }, { "Source ID": "collected-25d1b92226ad", "Operation IDs": "29", "Citation": "B45-T1 Blender 4.4 manual: Track settings and pattern matching", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/4.4/movie_clip/tracking/clip/toolbar/track.html" }, { "Source ID": "collected-ea9b9910e0b6", "Operation IDs": "29", "Citation": "docs.blender.org — track.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/latest/movie_clip/tracking/clip/toolbar/track.html" } ] }, { "ID": "30", "Software": "Blender", "Operation": "Cycles Open Image Denoise Pass", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7f31a6de0b21\ncollected-28a8bfa42b97\ncollected-b01a0f5f919a\ncollected-e7af858be84f\ncollected-caf577624d1d\ncollected-593450b4512c\ncollected-310eaea79912\ncollected-335ee64abea0\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Cycles Open Image Denoise predicts a clean image from a noisy render, optionally using auxiliary passes.", "Previous record: rationale": "The retained learned denoising responsibility matches fidelity restoration, independently of the host renderer.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). Blender 5.2 LTS (14 July 2026) is the current release; its Cycles notes mention OIDN only to say GPU acceleration was dropped on AMD RDNA2, and the Blender 5.1 notes report improved denoising with textured transparency and roughness. The OIDN mechanism (CNN trained on noisy and noise-free reference pairs; library v2.5.1 of 18 August 2026) is unchanged, while OIDN 3, announced for the second half of 2026, switches to kernel prediction and is unreleased as of 3 October 2026. Not established: The OIDN version bundled with Blender 5.2 is not stated in the release notes. Adoption of OIDN 3 kernel prediction in Cycles is not documented. developer.blender.org, Blender 5.2 LTS, 2026-07-14 • developer.blender.org, Blender 5.1, 2026-03-17 • github.com, v2.5.1", "Collected source links": "https://docs.blender.org/manual/en/4.5/render/cycles/render_settings/sampling.html\nhttps://www.openimagedenoise.org/\nhttps://developer.blender.org/docs/release_notes/5.2/cycles/\nhttps://developer.blender.org/docs/release_notes/5.1/cycles/\nhttps://github.com/RenderKit/oidn/blob/master/README.md\nhttps://docs.blender.org/manual/en/latest/scene_layout/scene/introduction.html\nhttps://docs.blender.org/manual/en/latest/render/cycles/render_settings/sampling.html\nhttps://docs.blender.org/manual/en/latest/render/output/properties/post_processing.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7f31a6de0b21", "Operation IDs": "30", "Citation": "docs.blender.org", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/4.5/render/cycles/render_settings/sampling.html" }, { "Source ID": "collected-28a8bfa42b97", "Operation IDs": "30", "Citation": "openimagedenoise.org", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.openimagedenoise.org/" }, { "Source ID": "collected-b01a0f5f919a", "Operation IDs": "30", "Citation": "developer.blender.org, Blender 5.2 LTS, 2026-07-14", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.blender.org/docs/release_notes/5.2/cycles/" }, { "Source ID": "collected-e7af858be84f", "Operation IDs": "30", "Citation": "developer.blender.org, Blender 5.1, 2026-03-17", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.blender.org/docs/release_notes/5.1/cycles/" }, { "Source ID": "collected-caf577624d1d", "Operation IDs": "16, 30", "Citation": "github.com, v2.5.1", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/RenderKit/oidn/blob/master/README.md" }, { "Source ID": "collected-593450b4512c", "Operation IDs": "30", "Citation": "docs.blender.org — introduction.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/latest/scene_layout/scene/introduction.html" }, { "Source ID": "collected-310eaea79912", "Operation IDs": "30", "Citation": "docs.blender.org — sampling.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/latest/render/cycles/render_settings/sampling.html" }, { "Source ID": "collected-335ee64abea0", "Operation IDs": "30", "Citation": "docs.blender.org — post_processing.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/latest/render/output/properties/post_processing.html" } ] }, { "ID": "31", "Software": "Blender", "Operation": "MCP LLM Scene Object Construction", "Proposed X": "2.9", "Review status": "Proposed X calculated", "Previous X": "2.0–10.0", "S": "5", "U": "5", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "LLM chooses new objects and arrangement from a high-level semantic request and emits native Blender code through MCP; exclude prescribed layouts/scene graphs, external image/mesh generators and fully specified literal code translation.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "High-level scene request and acceptance of the result, excluding prescribed layouts and external asset-generation paths.", "Learned determination": "The LLM chooses new objects, relationships and arrangement and emits Blender code.", "Conventional processing": "Native code execution creates the chosen entities and placements.", "Model contribution in the result": "The scene retains model-chosen objects and spatial relationships. Internal code tokens do not make this a temporally synthesized result.", "Why this scope score (S)": "S5 whole-scene entities/arrangement; U5 new scene; J4 semantic and spatial relationships. Code tokens are an internal representation and do not establish J5.", "Why this latitude score (U)": "S5 whole-scene entities/arrangement; U5 new scene; J4 semantic and spatial relationships. Code tokens are an internal representation and do not establish J5.", "Why this coordination score (J)": "S5 whole-scene entities/arrangement; U5 new scene; J4 semantic and spatial relationships. Code tokens are an internal representation and do not establish J5.", "Comparisons and other modes": "X2.9 ties SD3.5 whole-image synthesis at this coarse profile despite different output media. A fully specified literal translation or an external asset-generation path is a different configuration.", "Citations": "Blender MCP execution implementation — https://raw.githubusercontent.com/ahujasid/mcp-for-blender/main/src/blender_mcp/server.py\nBlender MCP user scene demonstration — https://www.reddit.com/r/OpenAI/comments/1wem6s7/chatgpt_blender_mcp_built_this_from_scratch/", "Source IDs": "CAL2-MCP\nCAL2-REPORTMCP", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies scene request and accepts result. LLM resolves object/relationship choices; conventional code execution creates those chosen entities and placements.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X2.9 ties SD3.5 whole-image synthesis at this coarse profile despite different output media. A fully specified literal translation or an external asset-generation path is a different configuration.", "Previous record: scope": "The Blender MCP integration executes an LLM's native scene-object construction commands.", "Previous record: rationale": "Learned object and arrangement choices survive command execution. External asset generators are excluded; the semantic construction comparison supports 2–10 without a selected point.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). The project is now mcp-for-blender (formerly blender-mcp), version 2.1.3 on PyPI, 30 September 2026 (2.0.0 on 16 September 2026). It has no dedicated object-creation tool: native construction runs through execute_blender_code, which executes LLM-written Python, and Tripo joins Hyper3D Rodin and Hunyuan3D as external generators. Because LLM-written Python can author mesh data directly, the REF02 end of the envelope stays supported with external generators excluded. Not established: Which share of typical constructions uses Blender primitive operators versus code-authored geometry. raw.githubusercontent.com, main branch, October 2026 • pypi.org, 2026-09-30 • github.com, 233 commits", "Collected source links": "https://www.blender.org/releases/4-5/\nhttps://github.com/ahujasid/mcp-for-blender/blob/41a184322db3ccdcb2fdbc1f6994afe71bf9163c/README.md\nhttps://raw.githubusercontent.com/ahujasid/mcp-for-blender/main/src/blender_mcp/server.py\nhttps://pypi.org/project/mcp-for-blender/\nhttps://github.com/ahujasid/mcp-for-blender", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "43/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-31", "Sources": [ { "Source ID": "CAL2-MCP", "Operation IDs": "31", "Citation": "Blender MCP execution implementation", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Blender MCP execution implementation", "Finding used in this review": "execute_blender_code runs generated bpy code; external asset generation tools are separate routes.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/ahujasid/mcp-for-blender/main/src/blender_mcp/server.py" }, { "Source ID": "CAL2-REPORTMCP", "Operation IDs": "31", "Citation": "Blender MCP user scene demonstration", "Type": "user_report_or_forum", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Blender MCP user scene demonstration", "Finding used in this review": "Recent user demonstration lead; displayed success cannot establish architecture or exclude hidden external assets.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://www.reddit.com/r/OpenAI/comments/1wem6s7/chatgpt_blender_mcp_built_this_from_scratch/" } ] }, { "ID": "32", "Software": "Blender", "Operation": "Mesh and Unapplied Modifier Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-01bd288a1c7e\ncollected-70e2e9c8faf2\ncollected-7bf16a7123d0\ncollected-99163a935288", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Blender evaluates direct mesh edits and conventional Mirror or Subdivision modifiers.", "Previous record: rationale": "Mesh editing and subdivision implement geometric rules without a retained trained predictor.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: mesh edit tools; Mirror (Axis, Bisect, Flip, Mirror Object, Clipping, Merge, Merge Distance); Subdivision Surface levels; Apply. Returns one result; editable structure: mesh with unapplied modifier stack. “Modifiers are automatic operations that affect an object's geometry in a non-destructive way.” Source • Screenshot", "Collected source links": "https://www.blender.org/releases/4-5/\nhttps://docs.blender.org/manual/id/4.5/modeling/modifiers/generate/mirror.html\nhttps://docs.blender.org/manual/en/4.5/modeling/modifiers/introduction.html\nhttps://docs.blender.org/manual/en/4.5/_images/modeling_modifiers_generate_mirror_panel.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-01bd288a1c7e", "Operation IDs": "29, 31, 32, 33, 34", "Citation": "B45-R Blender 4.5 LTS release history; 4.5.14 dated 09152026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.blender.org/releases/4-5/" }, { "Source ID": "collected-70e2e9c8faf2", "Operation IDs": "32", "Citation": "docs.blender.org", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/id/4.5/modeling/modifiers/generate/mirror.html" }, { "Source ID": "collected-7bf16a7123d0", "Operation IDs": "32", "Citation": "docs.blender.org — introduction.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/4.5/modeling/modifiers/introduction.html" }, { "Source ID": "collected-99163a935288", "Operation IDs": "32", "Citation": "docs.blender.org — modeling_modifiers_generate_mirror_panel.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/4.5/_images/modeling_modifiers_generate_mirror_panel.png" } ] }, { "ID": "33", "Software": "Blender", "Operation": "Rigid Body Constraint Simulation", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "Supplied bodies, masses, gravity, hinge/other constraints and time step; Bullet returns simulated body transforms.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Supplied bodies, masses, gravity, constraints, time step and duration.", "Learned determination": "None in the scoped Bullet dynamics path.", "Conventional processing": "The constraint solver integrates specified dynamics and returns body transforms.", "Model contribution in the result": "No learned prediction contributes.", "Why this scope score (S)": "S0/U0/J0 because no data-fitted predictive state contributes; complex coupled numerical behavior does not establish learning.", "Why this latitude score (U)": "S0/U0/J0 because no data-fitted predictive state contributes; complex coupled numerical behavior does not establish learning.", "Why this coordination score (J)": "S0/U0/J0 because no data-fitted predictive state contributes; complex coupled numerical behavior does not establish learning.", "Comparisons and other modes": "Can surprise the designer and automate many updates while tying a manual conventional operation at X15. This tie is appropriate for retained learning, not for total automation.", "Citations": "Blender 4.5 Bullet interface — https://raw.githubusercontent.com/blender/blender/blender-v4.5-release/intern/rigidbody/rb_bullet_api.cpp", "Source IDs": "CAL2-BULLET", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human specifies the physical setup. A conventional constraint solver integrates dynamics. No learned response is identified in this scoped path.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. Can surprise the designer and automate many updates while tying a manual conventional operation at X15. This tie is appropriate for retained learning, not for total automation.", "Previous record: scope": "Blender's rigid-body solver evaluates constraints, collisions and motion for the declared simulation.", "Previous record: rationale": "Numerical physics simulation uses prescribed dynamics. Emergent or unexpected motion is not evidence of a trained mechanism.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a solver or simulation. Designer controls: constraint type (e.g. Hinge); Object 1; Object 2; Enabled; Disable Collisions; Breakable; Override Iterations; Speed; Split Impulse; Substeps Per Frame; Solver Iterations; cache start/end. Returns one result; form not documented. “Number of simulation steps taken per frame (higher values are more accurate but slower).” Source", "Collected source links": "https://www.blender.org/releases/4-5/\nhttps://docs.blender.org/UATEST/manual/en/4.5/physics/rigid_body/constraints/types/hinge.html\nhttps://docs.blender.org/manual/en/latest/physics/rigid_body/world.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-33", "Sources": [ { "Source ID": "CAL2-BULLET", "Operation IDs": "33", "Citation": "Blender 4.5 Bullet interface", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Blender 4.5 Bullet interface", "Finding used in this review": "Constraint solver, gravity, iteration count and simulation step; conventional numerical dynamics.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/blender/blender/blender-v4.5-release/intern/rigidbody/rb_bullet_api.cpp" } ] }, { "ID": "34", "Software": "Blender", "Operation": "Sculpting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-01bd288a1c7e\ncollected-d9fc6d0d5609\ncollected-b7bb330c7484\ncollected-10198ec2f967\ncollected-0983d78ee886\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Blender sculpt brushes directly change mesh geometry through their selected conventional brush algorithms.", "Previous record: rationale": "Brush geometry and remeshing are conventional within the defined operation; previous generated content does not change the brush mechanism.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: brush (e.g. Grab); Radius; Strength; Normal Radius; Auto-Smooth; Grab Active Vertex; Grab Silhouette; Falloff Shape. Returns continuous live output; editable structure: mesh. “Grab only moves the vertices that are under the brush radius at the start of the stroke.” Source", "Collected source links": "https://www.blender.org/releases/4-5/\nhttps://docs.blender.org/manual/id/4.5/sculpt_paint/sculpting/brushes/grab.html\nhttps://docs.blender.org/manual/en/4.5/sculpt_paint/sculpting/brushes/grab.html\nhttps://docs.blender.org/manual/en/latest/interface/undo_redo.html\nhttps://web.archive.org/web/20260706143814id_/https://docs.blender.org/manual/en/latest/sculpt_paint/sculpting/introduction/general.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-01bd288a1c7e", "Operation IDs": "29, 31, 32, 33, 34", "Citation": "B45-R Blender 4.5 LTS release history; 4.5.14 dated 09152026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.blender.org/releases/4-5/" }, { "Source ID": "collected-d9fc6d0d5609", "Operation IDs": "34", "Citation": "B46-S1 Blender 4.5 manual: Grab brush", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/id/4.5/sculpt_paint/sculpting/brushes/grab.html" }, { "Source ID": "collected-b7bb330c7484", "Operation IDs": "34", "Citation": "docs.blender.org — grab.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/4.5/sculpt_paint/sculpting/brushes/grab.html" }, { "Source ID": "collected-10198ec2f967", "Operation IDs": "34", "Citation": "docs.blender.org — undo_redo.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.blender.org/manual/en/latest/interface/undo_redo.html" }, { "Source ID": "collected-0983d78ee886", "Operation IDs": "34", "Citation": "web.archive.org — general.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260706143814id_/https://docs.blender.org/manual/en/latest/sculpt_paint/sculpting/introduction/general.html" } ] }, { "ID": "35", "Software": "Blender (addon)", "Operation": "ComfyUI-BlenderAI-node Material Generation and Baking", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.0–4.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6dfadad04293\ncollected-26228c143cce\ncollected-b35dcafcff3d\ncollected-4fadfcbedae1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ComfyUI-BlenderAI-node generates texture imagery and Blender bakes the generated appearance onto a model.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 2 model-decided, 1 unknown and 0 setting-dependent categories among 4 applicable. The recorded uncertainty gives 3.0–4.0; no midpoint is used.", "Previous record: calculation": "K=1; M=2; unknown=1; conditional=0; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. Upper: 2 + 3 × 2 ÷ (4 − 1) = 4.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: User-supplied. The addon runs ComfyUI workflows inside Blender; Model choice: Yes. Not established: Which preset the one-click material generation uses by default and which checkpoint the developers tested it with. github.com • github.com, page date not shown\nDesigner side, 3 Oct 2026 Other. Decided by a trained model (ComfyUI workflow), then Blender bake). Designer controls: nodes/parameters (Shift+A node tree, presets); prompt; seed; model choice; Mat Image input from object; bake pass; bake resolution; UV layer. Returns count not documented; editable structure: baked image texture in the object material. “【New】AI material creation and texture baking” Source • Screenshot", "Collected source links": "https://github.com/AIGODLIKE/ComfyUI-BlenderAI-node/blob/b6729c96797e340c7688de00e547f1c41372ec76/README.md\nhttps://github.com/AIGODLIKE/ComfyUI-BlenderAI-node/releases/tag/v2.0.0\nhttps://github.com/AIGODLIKE/ComfyUI-BlenderAI-node\nhttps://github.com/AIGODLIKE/ComfyUI-BlenderAI-node/tree/main/presets/08BlenderAI", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6dfadad04293", "Operation IDs": "35", "Citation": "B46-M1 ComfyUI-BlenderAI-node README at inspected commit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/AIGODLIKE/ComfyUI-BlenderAI-node/blob/b6729c96797e340c7688de00e547f1c41372ec76/README.md" }, { "Source ID": "collected-26228c143cce", "Operation IDs": "35", "Citation": "B46-M2 ComfyUI-BlenderAI-node v2.0.0 release dated 07252025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/AIGODLIKE/ComfyUI-BlenderAI-node/releases/tag/v2.0.0" }, { "Source ID": "collected-b35dcafcff3d", "Operation IDs": "35", "Citation": "github.com — ComfyUI-BlenderAI-node", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/AIGODLIKE/ComfyUI-BlenderAI-node" }, { "Source ID": "collected-4fadfcbedae1", "Operation IDs": "35", "Citation": "github.com, page date not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/AIGODLIKE/ComfyUI-BlenderAI-node/tree/main/presets/08BlenderAI" } ] }, { "ID": "36", "Software": "Blender through StableGen", "Operation": "FLUX.2 Klein Texturing and Projection", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.0–4.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1b14c4fc478b\ncollected-e57afba8b07e\ncollected-fd03a88722de\ncollected-8ad92eff40f4", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "StableGen uses FLUX.2 Klein to synthesize texture views and projects those results onto existing geometry.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 2 model-decided, 1 unknown and 0 setting-dependent categories among 4 applicable. The recorded uncertainty gives 3.0–4.0; no midpoint is used.", "Previous record: calculation": "K=1; M=2; unknown=1; conditional=0; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. Upper: 2 + 3 × 2 ÷ (4 − 1) = 4.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (FLUX.2 Klein), then rule-based projection). Designer controls: prompt; camera viewpoints (count, placement); generation mode (separate, sequential, grid); reference image; bake to UV textures. Returns one result; editable structure: material node tree blending per-camera projected images; optional UV bake. “Projects an image onto all mesh objects using UV Project Modifier.” Source • Screenshot", "Collected source links": "https://github.com/sakalond/StableGen/blob/1d6e83fc966afc8b2c196c011b75ce7cdaa07c83/stablegen/util/workflow_templates.py\nhttps://github.com/sakalond/StableGen/releases/tag/v0.3.1\nhttps://github.com/sakalond/StableGen/blob/1d6e83fc966afc8b2c196c011b75ce7cdaa07c83/stablegen/texturing/projection.py\nhttps://github.com/sakalond/StableGen", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1b14c4fc478b", "Operation IDs": "36", "Citation": "B47-S1 StableGen Klein workflow template at inspected commit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/sakalond/StableGen/blob/1d6e83fc966afc8b2c196c011b75ce7cdaa07c83/stablegen/util/workflow_templates.py" }, { "Source ID": "collected-e57afba8b07e", "Operation IDs": "36", "Citation": "B47-S2 StableGen v0.3.1 release dated 06122026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/sakalond/StableGen/releases/tag/v0.3.1" }, { "Source ID": "collected-fd03a88722de", "Operation IDs": "36", "Citation": "github.com — projection.py", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/sakalond/StableGen/blob/1d6e83fc966afc8b2c196c011b75ce7cdaa07c83/stablegen/texturing/projection.py" }, { "Source ID": "collected-8ad92eff40f4", "Operation IDs": "36, 37", "Citation": "github.com — StableGen", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/sakalond/StableGen" } ] }, { "ID": "37", "Software": "Blender (addon)", "Operation": "StableGen TRELLIS.2 Image-to-Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8ad92eff40f4\ncollected-9b25fb7cead4\ncollected-0b2dd780f37c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "StableGen's TRELLIS.2 mesh-only route predicts a structured three-dimensional representation and decodes a mesh.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=2; conditional=0; A=3. Lower: 2 + 3 × 1 ÷ (3 − 1) = 3.5. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.5; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (TRELLIS.2 4B). Designer controls: reference image or text prompt; texture mode (None); resolution mode (512 to 1536 Cascade); decimation / remesh; import scale. Returns one result; editable structure: mesh imported into Blender. “('none', 'None', 'Shape only, no texture generation')” Source • Screenshot", "Collected source links": "https://github.com/sakalond/StableGen\nhttps://openaccess.thecvf.com/content/CVPR2026/papers/Xiang_Native_and_Compact_Structured_Latents_for_3D_Generation_CVPR_2026_paper.pdf\nhttps://github.com/sakalond/StableGen/blob/fae5474098c40ce149a93dbfffe02b33876b3cd4/stablegen/core/properties.py", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8ad92eff40f4", "Operation IDs": "36, 37", "Citation": "github.com — StableGen", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/sakalond/StableGen" }, { "Source ID": "collected-9b25fb7cead4", "Operation IDs": "37", "Citation": "openaccess.thecvf.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://openaccess.thecvf.com/content/CVPR2026/papers/Xiang_Native_and_Compact_Structured_Latents_for_3D_Generation_CVPR_2026_paper.pdf" }, { "Source ID": "collected-0b2dd780f37c", "Operation IDs": "37", "Citation": "github.com — properties.py", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/sakalond/StableGen/blob/fae5474098c40ce149a93dbfffe02b33876b3cd4/stablegen/core/properties.py" } ] }, { "ID": "38", "Software": "Blender (addon, Video Sequence Editor)", "Operation": "Pallaidium FLUX.2 Klein Image Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d8bee1c392fa\ncollected-1887d40c9fff\ncollected-28a5826b3b79", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Pallaidium invokes FLUX.2 Klein 4B image generation from Blender's video sequencer.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (FLUX.2 Klein 4B). Designer controls: prompt; seed; steps; guidance; width / height; input strip (img2img, inpaint); LoRA; reference images. Returns one result; a flat image: image added as a strip on a new VSE channel. “automatically added to your VSE timeline on a new channel” Source", "Collected source links": "https://github.com/tin2tin/Pallaidium/blob/80083591cfe90ce7039d475265213467fc3ae4d5/README.md\nhttps://github.com/tin2tin/Pallaidium/blob/80083591cfe90ce7039d475265213467fc3ae4d5/models_plugins/image/flux2_klein_4b.py\nhttps://github.com/tin2tin/Pallaidium", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d8bee1c392fa", "Operation IDs": "38", "Citation": "B47-P1 Pallaidium README at inspected commit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/tin2tin/Pallaidium/blob/80083591cfe90ce7039d475265213467fc3ae4d5/README.md" }, { "Source ID": "collected-1887d40c9fff", "Operation IDs": "38", "Citation": "B47-P2 Pallaidium Klein 4B plugin input handling and generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/tin2tin/Pallaidium/blob/80083591cfe90ce7039d475265213467fc3ae4d5/models_plugins/image/flux2_klein_4b.py" }, { "Source ID": "collected-28a5826b3b79", "Operation IDs": "38", "Citation": "github.com — Pallaidium", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/tin2tin/Pallaidium" } ] }, { "ID": "39", "Software": "Blender (plugin + hosted service)", "Operation": "Meshy for Blender Bridge", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c0aa2dbb644e\ncollected-0a3f911f7907\ncollected-1fbf4c2050da", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Meshy Bridge downloads and imports an existing model through glTF import and conventional mesh/material setup.", "Previous record: rationale": "The import command contains no trained generator. Earlier asset generation is separately scoped and does not give the import an AI score.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Run Bridge button; Send to Blender (DCC Bridge menu); local port 5324. Returns one result; editable structure: Blender mesh objects with materials, prefixed Meshy_. “The Bridge creates a seamless connection between MeshyAI and Blender, allowing you to transfer 3D models with a single click.” Source • Screenshot", "Collected source links": "https://docs.meshy.ai/en/webapp/plugins/blender/bridge-to-blender\nhttps://cdn.meshy.ai/meshy-for-blender/meshy-blender-plugin-v0.6.1.zip\nhttps://cdn.meshy.ai/docs-assets/blender-plugin/blender-plugin-v055-newpanel/doc_sendtoblender.webp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c0aa2dbb644e", "Operation IDs": "39", "Citation": "B47-M1 Meshy Bridge documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/webapp/plugins/blender/bridge-to-blender" }, { "Source ID": "collected-0a3f911f7907", "Operation IDs": "39", "Citation": "B47-M2 Official Meshy v0.6.1 archive: BridgePanel.py", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.meshy.ai/meshy-for-blender/meshy-blender-plugin-v0.6.1.zip" }, { "Source ID": "collected-1fbf4c2050da", "Operation IDs": "39", "Citation": "cdn.meshy.ai — doc_sendtoblender.webp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.meshy.ai/docs-assets/blender-plugin/blender-plugin-v055-newpanel/doc_sendtoblender.webp" } ] }, { "ID": "40", "Software": "Browzwear platform", "Operation": "AI Avatar Manual Pose Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-fdcb4271d280\ncollected-fa1a202208d2\ncollected-ec005e376215\ncollected-01206861a59b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Browzwear's pose editor manually changes an existing avatar pose.", "Previous record: rationale": "The specified pose-editing operation uses conventional controls. AI-branded avatar provenance does not establish learning in the edit itself.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: skeleton joints selection; Edit Pose pane; Save. Returns one result; editable structure: avatar pose skeleton. “The avatar pose is displayed in the form of a skeleton with bones and selectable joints.” Source • Screenshot", "Collected source links": "https://help.browzwear.com/en/articles/16992808-vstitcher-2026-2-1-release-notes\nhttps://help.browzwear.com/en/articles/13065644-designing-avatar-poses-overview\nhttps://help.browzwear.com/en/articles/13065645-editing-a-pose\nhttps://browzwear.intercom-attachments-7.com/i/o/eyv2k68l/1874399976/ad29eff69223f1a67322c8be80c0/40488782577689", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-fdcb4271d280", "Operation IDs": "40, 41", "Citation": "B48-R VStitcher 2026.2.1 release dated 09222026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.browzwear.com/en/articles/16992808-vstitcher-2026-2-1-release-notes" }, { "Source ID": "collected-fa1a202208d2", "Operation IDs": "40", "Citation": "B48-P1 Designing Avatar Poses Overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.browzwear.com/en/articles/13065644-designing-avatar-poses-overview" }, { "Source ID": "collected-ec005e376215", "Operation IDs": "40", "Citation": "help.browzwear.com — 13065645-editing-a-pose", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.browzwear.com/en/articles/13065645-editing-a-pose" }, { "Source ID": "collected-01206861a59b", "Operation IDs": "40", "Citation": "browzwear.intercom-attachments-7.com — 40488782577689", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://browzwear.intercom-attachments-7.com/i/o/eyv2k68l/1874399976/ad29eff69223f1a67322c8be80c0/40488782577689" } ] }, { "ID": "41", "Software": "Browzwear VStitcher", "Operation": "Pattern and Sewing Garment Simulation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-fdcb4271d280\ncollected-ce8f882ec53f\ncollected-902154c8c007", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "VStitcher simulates sewn garment patterns using the selected fabric and avatar configuration.", "Previous record: rationale": "The scoped cloth simulation is prescribed numerical modeling; separately offered learned avatar or image generation is excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by a solver or simulation. Designer controls: stitches (piece edges, internal edges, points); fabric physics (mass, friction, thickness, bend, stretch, shear, shrink); Dress; Finish; Undress; Update 3D; One Step Simulation. Returns one result; editable structure: draped 3D garment tied to 2D patterns. “Virtual stitching in VStitcher involves stitching together a combination of piece edges, internal edges, or points, to simulate physical stitching on garments.” Source • Screenshot", "Collected source links": "https://help.browzwear.com/en/articles/16992808-vstitcher-2026-2-1-release-notes\nhttps://help.browzwear.com/en/articles/16111588-vstitcher-2026-2-release-notes\nhttps://help.browzwear.com/en/articles/13065399-stitching-overview", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-fdcb4271d280", "Operation IDs": "40, 41", "Citation": "B48-R VStitcher 2026.2.1 release dated 09222026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.browzwear.com/en/articles/16992808-vstitcher-2026-2-1-release-notes" }, { "Source ID": "collected-ce8f882ec53f", "Operation IDs": "41", "Citation": "B48-G1 VStitcher 2026.2 release dated 07302026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.browzwear.com/en/articles/16111588-vstitcher-2026-2-release-notes" }, { "Source ID": "collected-902154c8c007", "Operation IDs": "41", "Citation": "help.browzwear.com — 13065399-stitching-overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.browzwear.com/en/articles/13065399-stitching-overview" } ] }, { "ID": "42", "Software": "Canva", "Operation": "Conversational generation of an editable social graphic", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-183fc5a562ff\ncollected-74779dbc3cf2\ncollected-59145a5fc55e\ncollected-d1773598d596", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Canva's Design action generates a layered social graphic through its documented design model.", "Previous record: rationale": "Learned layout and visual composition survive in editable elements. The present image references bound this structured design comparison at 2–10 without establishing a precise point.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). TechCrunch (30 October 2025) and CMSWire (16 April 2026, Canva AI 2.0) describe the Canva Design Model as a foundation model trained on Canva design elements that generates designs with editable layers and objects. Neither source states whether images inside a generated design are synthesized or drawn from Canva's library, so the split between generated content and placed assets stays open. Not established: Source of image elements in a generated design; share of layout generation versus asset selection. techcrunch.com, 2025-10-30 • cmswire.com, 2026-04-16", "Collected source links": "https://www.canva.com/newsroom/news/canva-create-2026-ai/\nhttps://www.canva.com/ai-assistant/\nhttps://techcrunch.com/2025/10/30/canva-launches-its-own-design-model-adds-new-ai-features-to-the-platform/\nhttps://www.cmswire.com/digital-experience/canva-ai-20-adds-agentic-design-tools/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-183fc5a562ff", "Operation IDs": "42", "Citation": "B48-C1 Canva AI 2.0 announcement and Design Model", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/newsroom/news/canva-create-2026-ai/" }, { "Source ID": "collected-74779dbc3cf2", "Operation IDs": "42", "Citation": "B48-C2 Canva AI assistant: editable components and brand context", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/ai-assistant/" }, { "Source ID": "collected-59145a5fc55e", "Operation IDs": "42", "Citation": "techcrunch.com, 2025-10-30", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://techcrunch.com/2025/10/30/canva-launches-its-own-design-model-adds-new-ai-features-to-the-platform/" }, { "Source ID": "collected-d1773598d596", "Operation IDs": "42", "Citation": "cmswire.com, 2026-04-16", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cmswire.com/digital-experience/canva-ai-20-adds-agentic-design-tools/" } ] }, { "ID": "43", "Software": "Canva", "Operation": "Manual layout using generated imagery", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-dce788f2cef5\ncollected-5c59c34bc1e4\ncollected-41f09b2e2bb7", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Canva's manual editor places, sizes, groups and orders existing image and text elements.", "Previous record: rationale": "The layout commands implement the specified edits conventionally. Generated asset provenance does not change their mechanism.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Position (toolbar); Layers tab; Forward / Backward / To front / Back; Group (Ctrl/Command+G); Align to page; Align elements; drag and resize. Returns one result; editable structure: design elements in layers. “When elements are added to your design, they are in layers, with front layers on top of the layers at the back.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.canva.com/en_au/help/layer-group-align-variantb/\nhttps://www.canva.com/newsroom/news/canva-ai-launches/\nhttps://www.canva.com/help/layer-group-align/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-dce788f2cef5", "Operation IDs": "43", "Citation": "B49-C1 Canva Group layer and align elements — consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/en_au/help/layer-group-align-variantb/" }, { "Source ID": "collected-5c59c34bc1e4", "Operation IDs": "43", "Citation": "Canva · Canva overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/newsroom/news/canva-ai-launches/" }, { "Source ID": "collected-41f09b2e2bb7", "Operation IDs": "43", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/layer-group-align/" } ] }, { "ID": "44", "Software": "Canva", "Operation": "Magic Design template arrangement of generated imagery", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-80649fb439f5\ncollected-baa4c61d2e53\ncollected-9c5d983004c2\ncollected-847e5a4d1b94\ncollected-d87331c601c3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magic Design recommends templates and arranges supplied imagery into designs.", "Previous record: rationale": "The sources do not separate learned recommendation or composition from prescribed template fitting. Template availability and AI branding cannot establish the equation branch.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Canva introduced a design model in October 2025 that generates layouts with separately editable layers and objects (Wikipedia). A November 2025 review states Magic Design generates designs that incorporate uploaded assets and match a brand palette. No Canva source retrieved states whether Magic Design template suggestion and image arrangement run on that design model, on a learned recommender, or on prescribed template fitting; canva.com and openai.com pages refused retrieval. Not established: Which mechanism selects templates and places supplied images in Magic Design as of 2026, and whether the October 2025 Canva design model participates. en.wikipedia.org, 2025-10 • vantagelabsai.com, 2025-11-28 • canva.dev, 2025-01-28", "Collected source links": "https://www.canva.com/help/use-magic-design/\nhttps://openai.com/index/canva/\nhttps://en.wikipedia.org/wiki/Canva\nhttps://www.vantagelabsai.com/blog/canva-ai-magic-studio-review\nhttps://www.canva.dev/blog/engineering/image-replacement-in-canva-designs-using-reverse-image-search/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-80649fb439f5", "Operation IDs": "44", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/use-magic-design/" }, { "Source ID": "collected-baa4c61d2e53", "Operation IDs": "44", "Citation": "OpenAI and Canva · Magic Design model and template combination", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://openai.com/index/canva/" }, { "Source ID": "collected-9c5d983004c2", "Operation IDs": "44", "Citation": "en.wikipedia.org, 2025-10", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://en.wikipedia.org/wiki/Canva" }, { "Source ID": "collected-847e5a4d1b94", "Operation IDs": "44", "Citation": "vantagelabsai.com, 2025-11-28", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.vantagelabsai.com/blog/canva-ai-magic-studio-review" }, { "Source ID": "collected-d87331c601c3", "Operation IDs": "44", "Citation": "canva.dev, 2025-01-28", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.dev/blog/engineering/image-replacement-in-canva-designs-using-reverse-image-search/" } ] }, { "ID": "45", "Software": "Canva", "Operation": "Magic Edit regional replacement", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–3.5", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f4a6160d343b\ncollected-6157a0e442fe\ncollected-1ce52cb38bb3\ncollected-4f314d8df294\ncollected-90a7d84164ab\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magic Edit synthesizes replacement content for a brushed image region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 3 model-decided, 2 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–3.5; no midpoint is used.", "Previous record: calculation": "K=0; M=3; unknown=2; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 2 ÷ (5 − 1) = 3.5. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Google Nano Banana (Gemini image model; version not stated), per a Canva feature page as shown in search-result excerpts. Model choice: No model choice documented. Not established: Which Nano Banana version (Gemini 2.5 Flash Image, Gemini 3 Pro Image or Gemini 3.1 Flash Image) and since when; what model Magic Edit used between its 2023 launch and the switch (search summaries attribute the 2023 version to Stable Diffusion; no primary source for that was found). canva.com pages returned HTTP 403 to direct fetch, so the quote was not checked against the full page. canva.com\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model. Designer controls: brush; prompt; Regenerate. Returns one result; a flat image. “Describe what you want to add or replace in the selected area and select Generate.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.canva.com/help/using-magic-edit/\nhttps://www.canva.com/features/ai-photo-editing/\nhttps://www.canva.com/help/undo-and-redo-changes/\nhttps://web.archive.org/web/20260930082825id_/https://www.canva.com/features/ai-photo-editing/\nhttps://www.canva.com/features/ai-replace/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f4a6160d343b", "Operation IDs": "45", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/using-magic-edit/" }, { "Source ID": "collected-6157a0e442fe", "Operation IDs": "45, 47", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/features/ai-photo-editing/" }, { "Source ID": "collected-1ce52cb38bb3", "Operation IDs": "45, 46", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/undo-and-redo-changes/" }, { "Source ID": "collected-4f314d8df294", "Operation IDs": "45", "Citation": "web.archive.org — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260930082825id_/https://www.canva.com/features/ai-photo-editing/" }, { "Source ID": "collected-90a7d84164ab", "Operation IDs": "45", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/features/ai-replace/" } ] }, { "ID": "46", "Software": "Canva", "Operation": "Magic Expand image outpainting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-882dcf486932\ncollected-f4401f129dbb\ncollected-2e8561204024\ncollected-1ce52cb38bb3\ncollected-54678ac8cd2d\ncollected-9b257454ce62\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magic Expand synthesizes image content beyond the existing canvas boundary.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Model choice: No model choice documented. Not established: Which model performs Magic Expand outpainting (in-house, Stable Diffusion-based, Leonardo, or Google). voicebot.ai, 2023-10-05\nDesigner side, 3 Oct 2026 Other. Decided by a trained model. Designer controls: size or aspect ratio choice; crop handles; Expand; Generate new results. Returns several to choose from (4); a flat image. “gives you four options to choose from” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.canva.com/help/using-magic-expand/\nhttps://www.canva.com/features/ai-image-expander/\nhttps://voicebot.ai/2023/10/05/canvas-new-magic-studio-pulls-enormous-generative-ai-toolkit-and-200m-creator-fund-out-of-a-hat/\nhttps://www.canva.com/help/undo-and-redo-changes/\nhttps://web.archive.org/web/20260904170511id_/https://www.canva.com/features/ai-image-expander/\nhttps://www.bettercontentcreators.com/blog/canva-magic-features", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-882dcf486932", "Operation IDs": "46", "Citation": "Canva Help — Expand images; workflow and limitations, consulted 09212026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/using-magic-expand/" }, { "Source ID": "collected-f4401f129dbb", "Operation IDs": "46", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/features/ai-image-expander/" }, { "Source ID": "collected-2e8561204024", "Operation IDs": "46, 47, 48", "Citation": "voicebot.ai, 2023-10-05", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://voicebot.ai/2023/10/05/canvas-new-magic-studio-pulls-enormous-generative-ai-toolkit-and-200m-creator-fund-out-of-a-hat/" }, { "Source ID": "collected-1ce52cb38bb3", "Operation IDs": "45, 46", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/undo-and-redo-changes/" }, { "Source ID": "collected-54678ac8cd2d", "Operation IDs": "46", "Citation": "web.archive.org — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260904170511id_/https://www.canva.com/features/ai-image-expander/" }, { "Source ID": "collected-9b257454ce62", "Operation IDs": "46", "Citation": "www.bettercontentcreators.com — canva-magic-features", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.bettercontentcreators.com/blog/canva-magic-features" } ] }, { "ID": "47", "Software": "Canva", "Operation": "Magic Grab object extraction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4966946e038c\ncollected-6157a0e442fe\ncollected-2e8561204024", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magic Grab extracts an object and completes the background exposed by extraction.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=4; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Described only as AI object segmentation that separates the subject and fills the area behind it. Model choice: No model choice documented. Not established: Which segmentation model isolates the object and which generative model fills the hole left behind. voicebot.ai, 2023-10-05\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (subject separation and background completion). Designer controls: select subject on photo. Returns one result; editable structure: subject and background as separate design elements. “select and separate the subject of your photo, making it an editable element that you can edit, reposition, or resize” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.canva.com/help/using-magic-grab/\nhttps://www.canva.com/features/ai-photo-editing/\nhttps://voicebot.ai/2023/10/05/canvas-new-magic-studio-pulls-enormous-generative-ai-toolkit-and-200m-creator-fund-out-of-a-hat/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4966946e038c", "Operation IDs": "47", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/using-magic-grab/" }, { "Source ID": "collected-6157a0e442fe", "Operation IDs": "45, 47", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/features/ai-photo-editing/" }, { "Source ID": "collected-2e8561204024", "Operation IDs": "46, 47, 48", "Citation": "voicebot.ai, 2023-10-05", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://voicebot.ai/2023/10/05/canvas-new-magic-studio-pulls-enormous-generative-ai-toolkit-and-200m-creator-fund-out-of-a-hat/" } ] }, { "ID": "48", "Software": "Canva", "Operation": "Magic Media text-to-image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7c8d02217a24\ncollected-fe923a28a5d5\ncollected-2e8561204024\ncollected-2176fd514f09\ncollected-cc29f193d975", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magic Media generates an image from the selected text-to-image action.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Stable Diffusion (Stability AI) at the 2022 Text to Image launch and the 2023 Magic Studio launch; version not stated. In July 2024 Canva stated plans to integrate Leonardo.Ai's Phoenix foundation model into Magic Studio and Magic Media; Model choice: No model choice inside Magic Media documented. Not established: Whether Magic Media text-to-image has switched from fine-tuned Stable Diffusion to Leonardo Phoenix, Lucid Origin or another model, and which Stable Diffusion version Canva fine-tuned. voicebot.ai, 2023-10-05 • maginative.com, 2024-07-30 • capitalbrief.com, 2024-10-23\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model. Designer controls: prompt; Inspire me; style (photography, digital art, fine art); reference image; layout (square, landscape, portrait); Generate again; Generate more like this. Returns several to choose from (count not stated); a flat image. “To generate a new set of images, select Generate again.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.canva.com/help/using-magic-media/\nhttps://www.canva.com/id_id/generator-model-3d-ai/\nhttps://voicebot.ai/2023/10/05/canvas-new-magic-studio-pulls-enormous-generative-ai-toolkit-and-200m-creator-fund-out-of-a-hat/\nhttps://www.maginative.com/article/canva-acquires-leonardo-ai/\nhttps://www.capitalbrief.com/briefing/canva-unveils-leonardoai-powered-dream-lab-a011b92e-3236-418b-b630-113c071439f5/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7c8d02217a24", "Operation IDs": "48", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/help/using-magic-media/" }, { "Source ID": "collected-fe923a28a5d5", "Operation IDs": "48", "Citation": "www.canva.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.canva.com/id_id/generator-model-3d-ai/" }, { "Source ID": "collected-2e8561204024", "Operation IDs": "46, 47, 48", "Citation": "voicebot.ai, 2023-10-05", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://voicebot.ai/2023/10/05/canvas-new-magic-studio-pulls-enormous-generative-ai-toolkit-and-200m-creator-fund-out-of-a-hat/" }, { "Source ID": "collected-2176fd514f09", "Operation IDs": "48", "Citation": "maginative.com, 2024-07-30", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.maginative.com/article/canva-acquires-leonardo-ai/" }, { "Source ID": "collected-cc29f193d975", "Operation IDs": "48", "Citation": "capitalbrief.com, 2024-10-23", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.capitalbrief.com/briefing/canva-unveils-leonardoai-powered-dream-lab-a011b92e-3236-418b-b630-113c071439f5/" } ] }, { "ID": "49", "Software": "CapCut (ByteDance)", "Operation": "Auto Captions editable timed text", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-47f57d16e604\ncollected-42eada6dc39a\ncollected-1024ac54c26e\ncollected-a8b34e7bea73\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "CapCut Auto Captions recognizes speech and places editable text on a timeline.", "Previous record: rationale": "Learned transcription and alignment estimate structured content. The 8–10 comparison spans fidelity reconstruction and relational determination; caption editability is not an X adjustment.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). CapCut's help page states that, as of January 2026, Auto Captions uses AI speech-to-text to generate subtitles from spoken audio with editable text and timing, and that speaker differentiation and punctuation refinement vary by platform. The documented learned output is a time-aligned transcript; no generated text beyond the transcript is documented for this route. Not established: The speech model and its training, and whether punctuation refinement uses a separate language model, are not published. Documentation does not place the route within 8–10. capcut.com, as of January 2026", "Collected source links": "https://www.capcut.com/help/how-to-recognise-subtitles\nhttps://www.capcut.com/help/modify-history-task\nhttps://www.capcut.com/help/auto-captions-in-capcut\nhttps://www.capcut.com/help/fix-apply-to-all-captions-error", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-47f57d16e604", "Operation IDs": "49", "Citation": "capcut.com, as of January 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.capcut.com/help/how-to-recognise-subtitles" }, { "Source ID": "collected-42eada6dc39a", "Operation IDs": "49", "Citation": "www.capcut.com — modify-history-task", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.capcut.com/help/modify-history-task" }, { "Source ID": "collected-1024ac54c26e", "Operation IDs": "49", "Citation": "www.capcut.com — auto-captions-in-capcut", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.capcut.com/help/auto-captions-in-capcut" }, { "Source ID": "collected-a8b34e7bea73", "Operation IDs": "49", "Citation": "www.capcut.com — fix-apply-to-all-captions-error", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.capcut.com/help/fix-apply-to-all-captions-error" } ] }, { "ID": "50", "Software": "Capture One Pro", "Operation": "Subject mask on generated imagery", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f6c4494188e1\ncollected-1d60b1ffad9b\ncollected-801d695c8c8b\ncollected-e7e45907d8c3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Capture One's Subject command predicts a foreground mask before image adjustments.", "Previous record: rationale": "Learned subject membership matches region determination. Other masking modes and subsequent adjustments are excluded.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Capture One 16.8 (May 2026) and service releases through 16.8.4 (July 2026) add Enhanced Denoise, tethering changes and AppleScript access to Subject, Background and People masking; no change to the Subject mask model and no generative step is documented. Not established: Capture One does not disclose the Subject mask model or its processing location. Capture One support pages refused direct fetch (HTTP 403); search-result text was used. support.captureone.com, 16.8.2, 2026 • captureone.com, May 2026 (16.8)", "Collected source links": "https://support.captureone.com/hc/en-us/articles/35279291925277-Capture-One-16-7-7-release-notes\nhttps://support.captureone.com/hc/en-us/articles/14055231933853-AI-Masking\nhttps://support.captureone.com/hc/en-us/articles/37047448168093-Capture-One-16-8-2-release-notes\nhttps://www.captureone.com/en/explore-features/whats-new/16-8", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f6c4494188e1", "Operation IDs": "50, 51, 52", "Citation": "B50-R Capture One 16.7.7 release notes — released 04162026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/35279291925277-Capture-One-16-7-7-release-notes" }, { "Source ID": "collected-1d60b1ffad9b", "Operation IDs": "50", "Citation": "B50-M Capture One AI Masking — consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/14055231933853-AI-Masking" }, { "Source ID": "collected-801d695c8c8b", "Operation IDs": "50", "Citation": "support.captureone.com, 16.8.2, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/37047448168093-Capture-One-16-8-2-release-notes" }, { "Source ID": "collected-e7e45907d8c3", "Operation IDs": "50", "Citation": "captureone.com, May 2026 (16.8)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.captureone.com/en/explore-features/whats-new/16-8" } ] }, { "ID": "51", "Software": "Capture One Pro", "Operation": "Manual masked curves on generated imagery", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f6c4494188e1\ncollected-2fe022ece9a6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Capture One applies manually painted masks and specified curves to a compatible image.", "Previous record: rationale": "The scoped mask painting and tonal functions are conventional. AI Masking and Match Look are separate operations.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Draw Mask brush; brush Size; Hardness; Opacity; Flow; Auto Mask; Curve tool points; adjustment layer. Returns one result; editable structure: adjustment layer with mask. “Once a mask has been created, you apply the image adjustments in the same way as you would when making global adjustments.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://support.captureone.com/hc/en-us/articles/35279291925277-Capture-One-16-7-7-release-notes\nhttps://support.captureone.com/hc/en-us/articles/360002601658-Overview-of-Layers-and-Masks", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f6c4494188e1", "Operation IDs": "50, 51, 52", "Citation": "B50-R Capture One 16.7.7 release notes — released 04162026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/35279291925277-Capture-One-16-7-7-release-notes" }, { "Source ID": "collected-2fe022ece9a6", "Operation IDs": "51", "Citation": "B50-L Capture One Overview of Layers and Masks — consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/360002601658-Overview-of-Layers-and-Masks" } ] }, { "ID": "52", "Software": "Capture One Pro", "Operation": "Match Look reference-based adjustments", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f6c4494188e1\ncollected-0ca95e10a1ab\ncollected-ba217f5224ee\ncollected-3ef9a40deb2e\ncollected-351946b33c50", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Capture One Match Look estimates adjustment settings from a reference look and applies existing image controls.", "Previous record: rationale": "The retained learned responsibility is choosing parameter values that prescribed Capture One operators then apply, which matches learned emulation of a fixed image operator (REF12); the perceptual-reconstruction end of the former range has no documented counterpart in this operation.", "Previous record: calculation": "X = 10 + 1 × (12 − 10) = 12.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–12.0 → 12.0. Capture One's feature page and 2024 reviews of 16.5 describe Match Look as setting regular Capture One adjustments (White Balance, Exposure, Contrast, Color Editor and related tools) that stay editable, matching tone and colour and leaving clarity, vignetting and dehaze untouched. The trained step outputs settings for prescribed tools; no documented step synthesizes pixels. Not established: Model architecture, training data and how multiple-reference weighting is computed. Capture One support pages refused fetch; the support-article quote comes from search-result text. Previous reason: A trained appearance-mapping role lies between fixed-operator emulation and perceptual reconstruction. Platform correspondence is qualified; the record supports 5–12 rather than a precise t. alexonraw.com, 2024-11-05 • blog.thomasfitzgeraldphotography.com, 2024-10-28 • support.captureone.com, not shown", "Collected source links": "https://support.captureone.com/hc/en-us/articles/35279291925277-Capture-One-16-7-7-release-notes\nhttps://support.captureone.com/hc/en-us/articles/22162074930589-Capture-One-16-5-0-Release-Notes\nhttps://alexonraw.com/capture-one-match-look-and-new-features-of-capture-one-16-5/\nhttps://blog.thomasfitzgeraldphotography.com/blog/2024/10/capture-one-165-match-look-a-remarkable-new-feature\nhttps://support.captureone.com/hc/en-us/articles/22188770298269-Match-Look-Tool", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f6c4494188e1", "Operation IDs": "50, 51, 52", "Citation": "B50-R Capture One 16.7.7 release notes — released 04162026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/35279291925277-Capture-One-16-7-7-release-notes" }, { "Source ID": "collected-0ca95e10a1ab", "Operation IDs": "52", "Citation": "B50-R2 Capture One 16.5.0 release notes — released 10232024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/22162074930589-Capture-One-16-5-0-Release-Notes" }, { "Source ID": "collected-ba217f5224ee", "Operation IDs": "52", "Citation": "alexonraw.com, 2024-11-05", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://alexonraw.com/capture-one-match-look-and-new-features-of-capture-one-16-5/" }, { "Source ID": "collected-3ef9a40deb2e", "Operation IDs": "52", "Citation": "blog.thomasfitzgeraldphotography.com, 2024-10-28", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.thomasfitzgeraldphotography.com/blog/2024/10/capture-one-165-match-look-a-remarkable-new-feature" }, { "Source ID": "collected-351946b33c50", "Operation IDs": "52", "Citation": "support.captureone.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.captureone.com/hc/en-us/articles/22188770298269-Match-Look-Tool" } ] }, { "ID": "53", "Software": "Cavalry", "Operation": "Procedural animation of generated SVG artwork", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-72d6f2a19114\ncollected-13533435c1b8\ncollected-52df20ffdf6d\ncollected-0ee5ee4f102d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Cavalry imports SVG artwork and evaluates Duplicator rules and authored transform animation.", "Previous record: rationale": "Prescribed procedural distribution and animation contain no trained predictor within this scope.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by procedural generation. Designer controls: SVG Shape (File, SVG Materials, SVG Strokes); Duplicator (Input Shapes, Distribution, Shape Position, Rotation, Scale, Opacity); Linear Distribution; Time Editor keyframes. Returns one result; editable structure: shape layers, Duplicator node, keyframes. “The Duplicator can be used to copy and distribute Shapes to create grids, circles and other patterns.” Source • Screenshot", "Collected source links": "https://cavalry.studio/docs/tech-info/release-notes/2.6/2-6-0-release-notes/\nhttps://cavalry.studio/docs/nodes/shapes/svg-shape/\nhttps://cavalry.studio/docs/nodes/shapes/duplicator/\nhttps://cavalry.studio/docs/assets/images/duplicator-intro-1ffe8ef7216bf0cd7eb4635c5dd9110b.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-72d6f2a19114", "Operation IDs": "53", "Citation": "B51-C1 Cavalry 2.6 release notes — released 02042026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cavalry.studio/docs/tech-info/release-notes/2.6/2-6-0-release-notes/" }, { "Source ID": "collected-13533435c1b8", "Operation IDs": "53", "Citation": "B51-C2 Cavalry SVG Shape — consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cavalry.studio/docs/nodes/shapes/svg-shape/" }, { "Source ID": "collected-52df20ffdf6d", "Operation IDs": "53", "Citation": "cavalry.studio — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cavalry.studio/docs/nodes/shapes/duplicator/" }, { "Source ID": "collected-0ee5ee4f102d", "Operation IDs": "53", "Citation": "cavalry.studio — duplicator-intro-1ffe8ef7216bf0cd7eb4635c5dd9110b.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cavalry.studio/docs/assets/images/duplicator-intro-1ffe8ef7216bf0cd7eb4635c5dd9110b.png" } ] }, { "ID": "54", "Software": "chaiNNer", "Operation": "ESRGAN image upscaling chain", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-92a09aee44ea\ncollected-e8f10232fecf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "chaiNNer loads an ESRGAN RRDB upscaler and applies it to an image through its node graph.", "Previous record: rationale": "The documented GAN-trained ESRGAN reconstruction matches the perceptual reconstruction reference. A different loaded checkpoint requires a new recorded configuration.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0). chaiNNer's latest release remains v0.25.1 (23 October 2025), the version the record cites; it contains compatibility and bug fixes with no change to ESRGAN model loading or upscaling. The repository has later commits (pushed 1 October 2026) without a new release. github.com, 2025-10-23", "Collected source links": "https://github.com/chaiNNer-org/chaiNNer/releases/tag/v0.25.1\nhttps://github.com/chaiNNer-org/chaiNNer/blob/v0.25.1/backend/src/packages/chaiNNer_pytorch/pytorch/io/load_model.py", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-92a09aee44ea", "Operation IDs": "54", "Citation": "B51-E1 chaiNNer v0.25.1 release — released 10232025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/chaiNNer-org/chaiNNer/releases/tag/v0.25.1" }, { "Source ID": "collected-e8f10232fecf", "Operation IDs": "54", "Citation": "B51-E2 chaiNNer v0.25.1 Load Model source — consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/chaiNNer-org/chaiNNer/blob/v0.25.1/backend/src/packages/chaiNNer_pytorch/pytorch/io/load_model.py" } ] }, { "ID": "55", "Software": "Chaos V-Ray", "Operation": "NVIDIA AI Denoiser Image Pass", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c4e310c9828f\ncollected-429602ddd37f\ncollected-19baaaf7ea30\ncollected-294ddfaa2e77", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "V-Ray uses the explicitly selected NVIDIA AI denoiser to reconstruct the noisy beauty pass.", "Previous record: rationale": "The trained noise-removal responsibility matches fidelity restoration. A fallback to the default conventional engine would be a different scoped operation.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). V-Ray 7 documentation (2026) offers NVIDIA AI (OptiX) as a denoiser engine beside the default V-Ray denoiser and Intel OIDN. NVIDIA's OptiX denoiser page states the network was trained on tens of thousands of images rendered from one thousand 3D scenes, using an autoencoder after NVIDIA's recurrent denoising autoencoder paper. Not established: The loss for the shipped OptiX weights and the OptiX version bundled with the current V-Ray build are not published. Chaos documentation pages returned no engine text for fetch; the engine list comes from a search result. developer.nvidia.com, accessed 2026-10-03 (OptiX 8 page) • documentation.chaos.com, V-Ray 7 documentation, 2026", "Collected source links": "https://docs-chaos.atlassian.net/wiki/spaces/VMAX/pages/113580676\nhttps://docs.chaos.com/vray_app_sdk/doc/csharp/class_v_ray_1_1_plugins_1_1_render_channel_denoiser.html\nhttps://developer.nvidia.com/optix-denoiser\nhttps://documentation.chaos.com/space/VMAYA/111739233/Denoiser", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c4e310c9828f", "Operation IDs": "55", "Citation": "docs-chaos.atlassian.net", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs-chaos.atlassian.net/wiki/spaces/VMAX/pages/113580676" }, { "Source ID": "collected-429602ddd37f", "Operation IDs": "55", "Citation": "B51-V2 Chaos RenderChannelDenoiser SDK reference — consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.chaos.com/vray_app_sdk/doc/csharp/class_v_ray_1_1_plugins_1_1_render_channel_denoiser.html" }, { "Source ID": "collected-19baaaf7ea30", "Operation IDs": "55", "Citation": "developer.nvidia.com, accessed 2026-10-03 (OptiX 8 page)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.nvidia.com/optix-denoiser" }, { "Source ID": "collected-294ddfaa2e77", "Operation IDs": "55", "Citation": "documentation.chaos.com, V-Ray 7 documentation, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documentation.chaos.com/space/VMAYA/111739233/Denoiser" } ] }, { "ID": "56", "Software": "ChatGPT", "Operation": "Current conversational image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7f4bf97ad90a\ncollected-121b461632ab\ncollected-3ef79fa4c44c\ncollected-39c9752fb8f0\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ChatGPT's image-generation service synthesizes the requested image through its managed image backend.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (managed image backend). Designer controls: prompt; reference image upload; area selection with description; canvas multi-select with comments; $imagegen invocation. Returns count not documented; a flat image: API guide: png default, jpeg or webp. “You can also select a specific area of an image and describe the change for that area.” Source", "Collected source links": "https://learn.chatgpt.com/docs/image-generation\nhttps://developers.openai.com/api/docs/guides/image-generation\nhttps://web.archive.org/web/20260920044555id_/https://help.openai.com/en/articles/11084440-chatgpt-image-library\nhttps://web.archive.org/web/20260929213057id_/https://help.openai.com/en/articles/6825453-chatgpt-release-notes", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7f4bf97ad90a", "Operation IDs": "56", "Citation": "learn.chatgpt.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.chatgpt.com/docs/image-generation" }, { "Source ID": "collected-121b461632ab", "Operation IDs": "56, 248", "Citation": "developers.openai.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developers.openai.com/api/docs/guides/image-generation" }, { "Source ID": "collected-3ef79fa4c44c", "Operation IDs": "56", "Citation": "web.archive.org — 11084440-chatgpt-image-library", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260920044555id_/https://help.openai.com/en/articles/11084440-chatgpt-image-library" }, { "Source ID": "collected-39c9752fb8f0", "Operation IDs": "56", "Citation": "web.archive.org — 6825453-chatgpt-release-notes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260929213057id_/https://help.openai.com/en/articles/6825453-chatgpt-release-notes" } ] }, { "ID": "57", "Software": "Cinema 4D", "Operation": "MCP LLM Scene Object Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a52522da9567\ncollected-58c22fe588dd\ncollected-9a34795d74fe", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "The Cinema 4D MCP bridge executes an LLM's native primitive and transform commands.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. Within the recorded scope the model determines which native primitives appear and their transform values, with geometry prescribed by Cinema 4D; this is the which-and-where kind of determination REF10 makes and places the operation nearer REF10 than REF02. The interval stays open below REF10 because the LLM authors the arrangement from a prompt.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. The cinema4d-mcp README (43 commits, no release tag, current October 2026) lists add_primitive, modify_object, create_camera, create_light, MoGraph and material tools, plus execute_python for custom Python; no tool calls a generative media model. The record's scope covers the primitive and transform commands. Not established: If execute_python were brought into scope, LLM-written code could author geometry and the REF02 end would apply again, as for the Blender bridge. Previous reason: Learned scene planning survives the conventional host commands. Current reference coverage supports 2–10; host compatibility is not a measure of inference. github.com, 43 commits", "Collected source links": "https://github.com/ttiimmaacc/cinema4d-mcp/blob/6bab376f0adbbfe20698c79689a14547e6bd9bb9/README.md\nhttps://github.com/ttiimmaacc/cinema4d-mcp/blob/6bab376f0adbbfe20698c79689a14547e6bd9bb9/src/cinema4d_mcp/server.py\nhttps://github.com/ttiimmaacc/cinema4d-mcp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a52522da9567", "Operation IDs": "57", "Citation": "B52-M1 Cinema 4D MCP README at inspected commit — 03072026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/ttiimmaacc/cinema4d-mcp/blob/6bab376f0adbbfe20698c79689a14547e6bd9bb9/README.md" }, { "Source ID": "collected-58c22fe588dd", "Operation IDs": "57", "Citation": "B52-M2 Cinema 4D MCP server tool definitions at inspected commit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/ttiimmaacc/cinema4d-mcp/blob/6bab376f0adbbfe20698c79689a14547e6bd9bb9/src/cinema4d_mcp/server.py" }, { "Source ID": "collected-9a34795d74fe", "Operation IDs": "57", "Citation": "github.com, 43 commits", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/ttiimmaacc/cinema4d-mcp" } ] }, { "ID": "58", "Software": "Clip Studio Paint", "Operation": "Colorize on generated line art", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-22368421d16d\ncollected-2b9810c9d57f\ncollected-3395ab223947\ncollected-907c805ab5de", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Clip Studio Paint Colorize uses its machine-learning service to estimate colorization for line art.", "Previous record: rationale": "Learned appearance allocation crosses the fixed-mapping and perceptual-reconstruction references. Optional color hints do not determine a numerical discount; retain 5–12.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). The Clip Studio Paint manual now documents Colorize on its AI Tools page: line art merged into one layer is uploaded to a CELSYS server, where the AI colors the image, with optional hint layers. A February 2024 CELSYS statement says its AI tools come from consented illustration data and that the company declined generative image AI; neither states the model's architecture or objective. Not established: Model architecture and training objective; whether shading in the colored layer is synthesized or region fills dominate. Training-data provenance does not move the coordinate under the comparison policy. help.clip-studio.com, © 2023 CELSYS • x.com, 2024-02", "Collected source links": "https://www.clipstudio.net/en/dl/release_note/latest/\nhttps://www.celsys.com/en/topic/2018112901\nhttps://help.clip-studio.com/en-us/manual_en/390_filters/AI_Tools.htm\nhttps://x.com/clipstudiopaint/status/1760284266272031138", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-22368421d16d", "Operation IDs": "58, 59, 60", "Citation": "B52-C1 Clip Studio Paint 5.1.4 release notes — released 09082026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.clipstudio.net/en/dl/release_note/latest/" }, { "Source ID": "collected-2b9810c9d57f", "Operation IDs": "58", "Citation": "B52-C2 Celsys 1.8.4 announcement — 11292018", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.celsys.com/en/topic/2018112901" }, { "Source ID": "collected-3395ab223947", "Operation IDs": "58", "Citation": "help.clip-studio.com, © 2023 CELSYS", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.clip-studio.com/en-us/manual_en/390_filters/AI_Tools.htm" }, { "Source ID": "collected-907c805ab5de", "Operation IDs": "58", "Citation": "x.com, 2024-02", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://x.com/clipstudiopaint/status/1760284266272031138" } ] }, { "ID": "59", "Software": "Clip Studio Paint", "Operation": "Hand Scanner to Editable 3D Hand Pose", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-22368421d16d\ncollected-fd11045bc108\ncollected-06f1535d3dc9", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Hand Scanner estimates a three-dimensional hand pose from camera imagery and applies it to the selected hand.", "Previous record: rationale": "Learned pose reconstruction crosses relational determination and fidelity estimation. Native pose controls transmit the inference; retain 8–10.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). The Clip Studio Paint manual (current with Ver. 5.1.4, 8 September 2026) describes the Hand Scanner as capturing camera footage and applying the hand pose to a 3D drawing figure, and release notes from Ver. 5.1.0 to 5.1.4 list no Hand Scanner change. The learned output is a pose estimate applied to the model's joints; no generated content is documented, and the model is not named. Not established: Model identity, on-device or server processing for the Hand Scanner, and training are not published. Documentation does not place pose estimation within 8–10. help.clip-studio.com, accessed 2026-10-03 • clipstudio.net, 2026-09-08", "Collected source links": "https://www.clipstudio.net/en/dl/release_note/latest/\nhttps://www.clipstudio.net/en/dl/release_note/v2/\nhttps://help.clip-studio.com/en-us/manual_en/660_3d/Posing_3D_drawing_figures.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-22368421d16d", "Operation IDs": "58, 59, 60", "Citation": "B52-C1 Clip Studio Paint 5.1.4 release notes — released 09082026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.clipstudio.net/en/dl/release_note/latest/" }, { "Source ID": "collected-fd11045bc108", "Operation IDs": "59", "Citation": "H2 Celsys — Version 2.0.0 release and Hand Scanner introduction 03142023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.clipstudio.net/en/dl/release_note/v2/" }, { "Source ID": "collected-06f1535d3dc9", "Operation IDs": "59", "Citation": "help.clip-studio.com, accessed 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.clip-studio.com/en-us/manual_en/660_3d/Posing_3D_drawing_figures.htm" } ] }, { "ID": "60", "Software": "Clip Studio Paint", "Operation": "Manual vector stroke revision of generated artwork", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-22368421d16d\ncollected-54d11c9d904c\ncollected-f2b70c5e6c2c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Clip Studio Paint moves a selected control point in an existing vector stroke.", "Previous record: rationale": "Conventional vector geometry evaluates the specified change. The provenance of the drawing reference is outside this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Object tool; Control point tool (Move, Add, Delete, Switch corner, Adjust line width, Adjust opacity, Split line); Correct line tools. Returns one result; editable structure: vector layer strokes. “Lines drawn in the vector format are made from a 'path' and 'control points'.” Source • Screenshot", "Collected source links": "https://www.clipstudio.net/en/dl/release_note/latest/\nhttps://help.clip-studio.com/en-us/manual_en/180_layers/Vector_layers.htm\nhttps://help.clip-studio.com/en-us/manual_en/images/180_layers_0084.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-22368421d16d", "Operation IDs": "58, 59, 60", "Citation": "B52-C1 Clip Studio Paint 5.1.4 release notes — released 09082026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.clipstudio.net/en/dl/release_note/latest/" }, { "Source ID": "collected-54d11c9d904c", "Operation IDs": "60", "Citation": "V2 Celsys — Vector layers and Control point editing; consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.clip-studio.com/en-us/manual_en/180_layers/Vector_layers.htm" }, { "Source ID": "collected-f2b70c5e6c2c", "Operation IDs": "60", "Citation": "help.clip-studio.com — 180_layers_0084.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.clip-studio.com/en-us/manual_en/images/180_layers_0084.png" } ] }, { "ID": "61", "Software": "CLO 3D", "Operation": "Sewn Pattern Garment Simulation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2583df80979b\ncollected-3fd813b38eae\ncollected-fdd479c289a5\ncollected-c6bf8911e7e9\ncollected-b6f49922172c\ncollected-f851194341f9\ncollected-33ba6e8ac066\ncollected-28eff88fac18\ncollected-77703c62df06\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "CLO simulates a garment from sewing relationships, fabric parameters and an avatar.", "Previous record: rationale": "The scoped cloth dynamics is conventional numerical simulation. Unanticipated drape does not establish a trained predictor.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by a solver or simulation. Designer controls: segment sewing; fabric physical properties (Stretch Weft/Warp/Shear, Bending); Particle Distance; simulation mode (Normal, Fitting/Accurate Fabric, Animation/Stable); collision offset and thickness. Returns one result; editable structure: draped 3D garment tied to 2D patterns. “Garment simulation is more efficient with faster speed in Normal mode.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://developer.clo3d.com/download.html\nhttps://developer.clo3d.com/changelog.html\nhttps://support.clo3d.com/hc/en-us/articles/30201166554649-Edit-Simulation-Properties\nhttps://support.clo3d.com/hc/en-us/articles/49988591822873-AI-Studio\nhttps://support.clo3d.com/hc/en-us/articles/49343947711769-CLOFAB\nhttps://support.clo3d.com/hc/en-us/articles/53856382869657-Simulation\nhttps://support.clo3d.com/hc/en-us/articles/115002686988-Project-File-ZPRJ-Open-Add-Save\nhttps://support.clo3d.com/hc/en-us/articles/115012227227-Particle-Distance\nhttps://support.clo3d.com/hc/en-us/articles/115012381188-Reset-3D-Arrangement-All", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2583df80979b", "Operation IDs": "61", "Citation": "C1 CLO — API/SDK Download, CLO 2026.1.224; consulted 09282026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.clo3d.com/download.html" }, { "Source ID": "collected-3fd813b38eae", "Operation IDs": "61", "Citation": "C2 CLO — SDK 11.0.0 changelog, August 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.clo3d.com/changelog.html" }, { "Source ID": "collected-fdd479c289a5", "Operation IDs": "61", "Citation": "CLO simulation properties", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.clo3d.com/hc/en-us/articles/30201166554649-Edit-Simulation-Properties" }, { "Source ID": "collected-c6bf8911e7e9", "Operation IDs": "61", "Citation": "CLO AI Studio", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.clo3d.com/hc/en-us/articles/49988591822873-AI-Studio" }, { "Source ID": "collected-b6f49922172c", "Operation IDs": "61", "Citation": "support.clo3d.com — 49343947711769-CLOFAB", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.clo3d.com/hc/en-us/articles/49343947711769-CLOFAB" }, { "Source ID": "collected-f851194341f9", "Operation IDs": "61", "Citation": "support.clo3d.com — 53856382869657-Simulation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.clo3d.com/hc/en-us/articles/53856382869657-Simulation" }, { "Source ID": "collected-33ba6e8ac066", "Operation IDs": "61", "Citation": "support.clo3d.com — 115002686988-Project-File-ZPRJ-Open-Add-Save", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.clo3d.com/hc/en-us/articles/115002686988-Project-File-ZPRJ-Open-Add-Save" }, { "Source ID": "collected-28eff88fac18", "Operation IDs": "61", "Citation": "support.clo3d.com — 115012227227-Particle-Distance", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.clo3d.com/hc/en-us/articles/115012227227-Particle-Distance" }, { "Source ID": "collected-77703c62df06", "Operation IDs": "61", "Citation": "support.clo3d.com — 115012381188-Reset-3D-Arrangement-All", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.clo3d.com/hc/en-us/articles/115012381188-Reset-3D-Arrangement-All" } ] }, { "ID": "62", "Software": "ComfyUI", "Operation": "SD3.5 Large Image to Image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6601fbc003b6\ncollected-1ed4abb05f36\ncollected-4c1055dfe15c\ncollected-bc0eb3f49d29", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ComfyUI encodes an image, runs SD3.5 Large image-to-image sampling and decodes the generated result.", "Previous record: rationale": "Current placement unresolved. K=3; M=1; unknown=0; conditional=1; A=5. Lower: 2 + 3 × 3 ÷ (5 − 1) = 4.25. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. This arithmetic reproduces the v5 proposal; it is not a verified current range for SD3.5.", "Previous record: calculation": "No current numeric placement; the latest record states the remaining requirement. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (SD3.5 Large). Designer controls: input image; prompt (positive, negative); denoise; seed; steps; cfg; sampler; scheduler; nodes/parameters. Returns one result; a flat image: Save Image to ComfyUI/output. “The smaller the denoise value, the smaller the difference between the generated image and the reference image” Source • Screenshot", "Collected source links": "https://baeseokjae.github.io/posts/comfyui-workflow-ai-image-generation-guide-2026/\nhttps://comfyanonymous.github.io/ComfyUI_examples/img2img/\nhttps://docs.comfy.org/tutorials/basic/image-to-image\nhttps://raw.githubusercontent.com/Comfy-Org/example_workflows/refs/heads/main/image_to_image/workflow.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6601fbc003b6", "Operation IDs": "62", "Citation": "Baeseok Jae · img2img", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://baeseokjae.github.io/posts/comfyui-workflow-ai-image-generation-guide-2026/" }, { "Source ID": "collected-1ed4abb05f36", "Operation IDs": "62", "Citation": "ComfyUI · Image to image generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://comfyanonymous.github.io/ComfyUI_examples/img2img/" }, { "Source ID": "collected-4c1055dfe15c", "Operation IDs": "62", "Citation": "docs.comfy.org — image-to-image", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.comfy.org/tutorials/basic/image-to-image" }, { "Source ID": "collected-bc0eb3f49d29", "Operation IDs": "62", "Citation": "raw.githubusercontent.com — workflow.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/Comfy-Org/example_workflows/refs/heads/main/image_to_image/workflow.png" } ] }, { "ID": "63", "Software": "ComfyUI", "Operation": "SD3.5 Large Masked Latent Inpainting", "Proposed X": "3.7", "Review status": "Proposed X calculated", "Previous X": "Unresolved", "S": "4", "U": "5", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "Pinned ComfyUI sampler; SD3.5 Large; effective latent mask contains active support and a nonempty exactly zero complement with source-latent reinstatement; new content generated inside.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Source image, prompt, sampler settings and effective latent mask with both active support and an exactly zero protected complement.", "Learned determination": "SD3.5 denoising synthesizes regional content; learned encoding/decoding also reconstructs the image.", "Conventional processing": "The sampler reinstates source latents in the protected complement after each prediction.", "Model contribution in the result": "Generative predictions remain in the active region and learned reconstruction remains elsewhere. Latent reinstatement does not establish exact source-pixel copying.", "Why this scope score (S)": "S4 protected completion/registered reconstruction; U5 new regional content without a supplied internal structural guide; J4 joint semantic/spatial synthesis.", "Why this latitude score (U)": "S4 protected completion/registered reconstruction; U5 new regional content without a supplied internal structural guide; J4 joint semantic/spatial synthesis.", "Why this coordination score (J)": "S4 protected completion/registered reconstruction; U5 new regional content without a supplied internal structural guide; J4 joint semantic/spatial synthesis.", "Comparisons and other modes": "X3.7 versus whole generation X2.9. An all-one mask with pure-noise initialization removes this protection and becomes 5/5/4. A soft mask of 0.5 everywhere does not prove protection. Small and nearly full protected regions tie under this coarse classification.", "Citations": "ComfyUI pinned sampler code — https://raw.githubusercontent.com/Comfy-Org/ComfyUI/5c460d8172fe30761ff67c0df3d5643bb74e0d70/comfy/samplers.py\nSD3.5 Large model card — https://huggingface.co/stabilityai/stable-diffusion-3.5-large\nScaling Rectified Flow Transformers — https://arxiv.org/html/2403.03206v1\nComfyUI user report and replies — https://www.reddit.com/r/comfyui/comments/1tua87l/why_do_inpaint_and_outpaint_workflows_stop_working/", "Source IDs": "CAL2-MASK\nCAL2-SDMODEL\nCAL2-SDPAPER\nCAL2-REPORTMASK", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human provides source, mask and prompt. Learned encoding/denoising/decoding contributes. Sampler reinstatement restricts new generative predictions; the VAE can still reconstruct the whole image.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X3.7 versus whole generation X2.9. An all-one mask with pure-noise initialization removes this protection and becomes 5/5/4. A soft mask of 0.5 everywhere does not prove protection. Small and nearly full protected regions tie under this coarse classification.", "Previous record: scope": "ComfyUI applies a latent noise mask to ordinary SD3.5 Large sampling and decodes the inpainted image.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=2; conditional=3; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (SD3.5 Large, ordinary checkpoint). Designer controls: mask (MaskEditor brush or alpha channel); prompt (positive, negative); denoise; seed; steps; cfg; sampler; scheduler; nodes/parameters. Returns one result; a flat image: Save Image to ComfyUI/output. “This node is designed to apply a noise mask to a set of latent samples.” Source • Screenshot\nEarlier X comparison finding, 4 October 2026. For a nontrivial mask in this workflow, the sampler restores source latent outside the mask and the VAE decodes the full latent. This establishes latent retention, not exact source-pixel preservation or a measured reduction in decisions. A full-frame mask removes the spatial restriction. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://comfyanonymous.github.io/ComfyUI_examples/inpaint/\nhttps://comfyanonymous.github.io/ComfyUI_examples/sd3/\nhttps://docs.comfy.org/built-in-nodes/SetLatentNoiseMask\nhttps://mintcdn.com/dripart/OltlUSVBSNcJsDMs/images/tutorial/basic/inpaint/inpaint-maskeditor.gif\nhttps://github.com/Comfy-Org/ComfyUI/blob/5c460d8172fe30761ff67c0df3d5643bb74e0d70/comfy/model_base.py\nhttps://huggingface.co/stabilityai/stable-diffusion-3.5-large\nhttps://github.com/Comfy-Org/ComfyUI/blob/5c460d8172fe30761ff67c0df3d5643bb74e0d70/comfy/model_sampling.py\nhttps://forum.comfy.org/t/comfyui-crashes-on-generate/4200\nhttps://github.com/Comfy-Org/ComfyUI/blob/5c460d8172fe30761ff67c0df3d5643bb74e0d70/nodes.py\nhttps://arxiv.org/html/2403.03206v1\nhttps://github.com/Comfy-Org/ComfyUI/blob/5c460d8172fe30761ff67c0df3d5643bb74e0d70/comfy/samplers.py", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "56/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-63", "Sources": [ { "Source ID": "CAL2-MASK", "Operation IDs": "63", "Citation": "ComfyUI pinned sampler code", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: ComfyUI pinned sampler code", "Finding used in this review": "KSamplerX0Inpaint blends source latent outside effective mask; empty sigma schedule differs from full denoising.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/Comfy-Org/ComfyUI/5c460d8172fe30761ff67c0df3d5643bb74e0d70/comfy/samplers.py" }, { "Source ID": "CAL2-SDMODEL", "Operation IDs": "63, 64, 65", "Citation": "SD3.5 Large model card", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: SD3.5 Large model card", "Finding used in this review": "Text-to-image model and MMDiT family; product-linked synthesis evidence.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://huggingface.co/stabilityai/stable-diffusion-3.5-large" }, { "Source ID": "CAL2-SDPAPER", "Operation IDs": "63, 64, 65", "Citation": "Scaling Rectified Flow Transformers", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Scaling Rectified Flow Transformers", "Finding used in this review": "Joint text/image attention supplies the product-linked basis for semantic and spatial coordination, rather than inferring it from screenshots.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://arxiv.org/html/2403.03206v1" }, { "Source ID": "CAL2-REPORTMASK", "Operation IDs": "63", "Citation": "ComfyUI user report and replies", "Type": "user_report_or_forum", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: ComfyUI user report and replies", "Finding used in this review": "Replies identify denoise changes and package drift as possible causes; exact SD3.5 configuration is not established. Leads checked against code, not treated as facet evidence.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://www.reddit.com/r/comfyui/comments/1tua87l/why_do_inpaint_and_outpaint_workflows_stop_working/" } ] }, { "ID": "64", "Software": "ComfyUI", "Operation": "SD3.5 Large Text to Image", "Proposed X": "2.9", "Review status": "Proposed X calculated", "Previous X": "2.0", "S": "5", "U": "5", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "SD3.5 Large, high-level semantic request and empty/noise latent; no prescribed layout/pose/trajectory or protected source region; return decoded image.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "High-level semantic prompt, model/sampler settings and noise latent, with no prescribed internal layout.", "Learned determination": "SD3.5 determines entities, appearance and arrangement throughout the image.", "Conventional processing": "Sampling and decoding carry the learned choices into the image.", "Model contribution in the result": "Whole-domain learned synthesis remains; choosing a seed makes the run repeatable without specifying its content.", "Why this scope score (S)": "S5 whole-domain generation; U5 new instance; J4 learned semantic/spatial coordination supported by the product-linked MMDiT account.", "Why this latitude score (U)": "S5 whole-domain generation; U5 new instance; J4 learned semantic/spatial coordination supported by the product-linked MMDiT account.", "Why this coordination score (J)": "S5 whole-domain generation; U5 new instance; J4 learned semantic/spatial coordination supported by the product-linked MMDiT account.", "Comparisons and other modes": "X2.9, not a universal X2 for all generative operations. A fixed seed makes a run repeatable without replacing learned content with human specification.", "Citations": "SD3.5 Large model card — https://huggingface.co/stabilityai/stable-diffusion-3.5-large\nScaling Rectified Flow Transformers — https://arxiv.org/html/2403.03206v1", "Source IDs": "CAL2-SDMODEL\nCAL2-SDPAPER", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies semantic instructions and controls. Learned synthesis resolves entities, appearance and arrangement across the image; sampling and decoding retain it.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X2.9, not a universal X2 for all generative operations. A fixed seed makes a run repeatable without replacing learned content with human specification.", "Previous record: scope": "ComfyUI runs SD3.5 Large text encoding, learned latent prediction and learned image decoding.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (SD3.5 Large). Designer controls: prompt (positive, negative); width; height; batch size; seed; steps; cfg; sampler; scheduler; checkpoint and text encoders; nodes/parameters. Returns one result (batch_size default 1, max 4096); a flat image: Save Image to ComfyUI/output. “io.Int.Input(\"batch_size\", default=1, min=1, max=4096)” Source • Screenshot\nEarlier X comparison finding, 4 October 2026. Compared with masked inpainting (63), this mode retains newly sampled latent content across the image. Canny (65) and Depth (66) add structural conditions; those conditions do not establish a percentage reduction in model decisions. Entry 68 describes the same generation-and-decode route and remains tied in all tested coding cases. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://docs.comfy.org/tutorials/basic/text-to-image\nhttps://comfyanonymous.github.io/ComfyUI_examples/sd3/\nhttps://github.com/Comfy-Org/ComfyUI/blob/95539f56344958339e39b7582a476267d489b0ee/comfy_extras/nodes_sd3.py\nhttps://comfyanonymous.github.io/ComfyUI_examples/sd3/sd3.5_simple_example.png\nhttps://docs.comfy.org/development/api-development/workflow-metadata\nhttps://docs.comfy.org/basic-concepts/nodes\nhttps://huggingface.co/stabilityai/stable-diffusion-3.5-large\nhttps://forum.comfy.org/t/comfyui-crashes-on-generate/4200\nhttps://arxiv.org/html/2403.03206v1", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "43/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-64", "Sources": [ { "Source ID": "CAL2-SDMODEL", "Operation IDs": "63, 64, 65", "Citation": "SD3.5 Large model card", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: SD3.5 Large model card", "Finding used in this review": "Text-to-image model and MMDiT family; product-linked synthesis evidence.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://huggingface.co/stabilityai/stable-diffusion-3.5-large" }, { "Source ID": "CAL2-SDPAPER", "Operation IDs": "63, 64, 65", "Citation": "Scaling Rectified Flow Transformers", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Scaling Rectified Flow Transformers", "Finding used in this review": "Joint text/image attention supplies the product-linked basis for semantic and spatial coordination, rather than inferring it from screenshots.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://arxiv.org/html/2403.03206v1" } ] }, { "ID": "65", "Software": "ComfyUI", "Operation": "SD3.5 Large ControlNet Canny Generation", "Proposed X": "3.7", "Review status": "Proposed X calculated", "Previous X": "3.5–5.0", "S": "5", "U": "4", "J": "4", "Reviewed on": "2026-10-14", "Configuration": "SD3.5 Large plus active Canny ControlNet, supplied edge map, prompt and noise latent; no protected source-pixel complement.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "An explicit edge map, prompt and active Canny ControlNet; no protected source region.", "Learned determination": "Conditioned synthesis interprets the edge structure and constructs the whole image.", "Conventional processing": "Canny edge extraction can prepare the guide; sampling and decoding retain learned synthesis.", "Model contribution in the result": "The guide constrains locations and shape relationships without copying or protecting source pixels.", "Why this scope score (S)": "S5 remains whole-domain generation. Proposed U4 recognizes explicit supplied structural guidance. J4 remains joint semantic/spatial synthesis. v1 U5 produced the same profile as unconditioned text to image.", "Why this latitude score (U)": "S5 remains whole-domain generation. Proposed U4 recognizes explicit supplied structural guidance. J4 remains joint semantic/spatial synthesis. v1 U5 produced the same profile as unconditioned text to image.", "Why this coordination score (J)": "S5 remains whole-domain generation. Proposed U4 recognizes explicit supplied structural guidance. J4 remains joint semantic/spatial synthesis. v1 U5 produced the same profile as unconditioned text to image.", "Comparisons and other modes": "Proposed X3.7 versus v1 X2.9. This categorical step represents constrained structure, not a measured percentage. Zero control returns to the text-to-image reference. Canny and protected inpainting tie through different profiles; they do not impose identical constraints.", "Citations": "Stability SD3.5 Large ControlNets — https://stability.ai/news-updates/sd3-5-large-controlnets\nSD3.5 ControlNets model card — https://huggingface.co/stabilityai/stable-diffusion-3.5-controlnets/blob/main/README.md\nSD3.5 Large model card — https://huggingface.co/stabilityai/stable-diffusion-3.5-large\nScaling Rectified Flow Transformers — https://arxiv.org/html/2403.03206v1", "Source IDs": "CAL2-CONTROL\nCAL2-CONTROLCARD\nCAL2-SDMODEL\nCAL2-SDPAPER", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies spatial edge relationships. Conventional Canny extraction can produce the map. Learned conditioned synthesis interprets it and constructs the image; guidance does not copy source pixels.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. Proposed X3.7 versus v1 X2.9. This categorical step represents constrained structure, not a measured percentage. Zero control returns to the text-to-image reference. Canny and protected inpainting tie through different profiles; they do not impose identical constraints.", "Previous record: scope": "SD3.5 Large Canny ControlNet processes a supplied edge map during learned generation and decoding.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 1 model-decided, 1 unknown and 1 setting-dependent categories among 5 applicable. The recorded uncertainty gives 3.5–5.0; no midpoint is used.", "Previous record: calculation": "K=2; M=1; unknown=1; conditional=1; A=5. Lower: 2 + 3 × 2 ÷ (5 − 1) = 3.5. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (SD3.5 Large with Canny ControlNet). Designer controls: nodes/parameters; prompt (positive, negative); edge map image; ControlNet strength; start / end percent; seed; steps; cfg; sampler; scheduler; batch size. Returns one result (batch_size default 1, adjustable); a flat image: Save Image to ComfyUI/output. “Leverage Canny edge maps to structure your generated images.” Source • Screenshot\nEarlier X comparison finding, 4 October 2026. The edge image guides a trained generator; supplied guidance is not copied output geometry and does not remove model resolution. The source review does not establish a greater/lesser order between Canny (65) and Depth (66). The original label-count ordering depends on the disputed shape label. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://stability.ai/news-updates/sd3-5-large-controlnets\nhttps://stability.ai/news-updates?category=Research\nhttps://comfyanonymous.github.io/ComfyUI_examples/sd3/sd3.5_large_canny_controlnet_example.png", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "56/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-65", "Sources": [ { "Source ID": "CAL2-CONTROL", "Operation IDs": "65", "Citation": "Stability SD3.5 Large ControlNets", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Stability SD3.5 Large ControlNets", "Finding used in this review": "Canny structural guidance and depth spatial guidance; no exact-copy guarantee.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://stability.ai/news-updates/sd3-5-large-controlnets" }, { "Source ID": "CAL2-CONTROLCARD", "Operation IDs": "65", "Citation": "SD3.5 ControlNets model card", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: SD3.5 ControlNets model card", "Finding used in this review": "Canny preprocessing and active conditioning; strength is not a percentage of determination.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://huggingface.co/stabilityai/stable-diffusion-3.5-controlnets/blob/main/README.md" }, { "Source ID": "CAL2-SDMODEL", "Operation IDs": "63, 64, 65", "Citation": "SD3.5 Large model card", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: SD3.5 Large model card", "Finding used in this review": "Text-to-image model and MMDiT family; product-linked synthesis evidence.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://huggingface.co/stabilityai/stable-diffusion-3.5-large" }, { "Source ID": "CAL2-SDPAPER", "Operation IDs": "63, 64, 65", "Citation": "Scaling Rectified Flow Transformers", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Scaling Rectified Flow Transformers", "Finding used in this review": "Joint text/image attention supplies the product-linked basis for semantic and spatial coordination, rather than inferring it from screenshots.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://arxiv.org/html/2403.03206v1" } ] }, { "ID": "66", "Software": "ComfyUI", "Operation": "SD3.5 Large ControlNet Depth Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.5–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-843fb18cc85c\ncollected-2974b63ec542\ncollected-087b1b5618e1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SD3.5 Large Depth ControlNet processes a supplied depth map during learned generation and decoding.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 1 model-decided, 1 unknown and 1 setting-dependent categories among 5 applicable. The recorded uncertainty gives 3.5–5.0; no midpoint is used.", "Previous record: calculation": "K=2; M=1; unknown=1; conditional=1; A=5. Lower: 2 + 3 × 2 ÷ (5 − 1) = 3.5. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (SD3.5 Large with Depth ControlNet). Designer controls: nodes/parameters; prompt (positive, negative); depth map image; ControlNet strength; start / end percent; seed; steps; cfg; sampler; scheduler; batch size. Returns one result (batch_size default 1, adjustable); a flat image: Save Image to ComfyUI/output. “Use depth maps, generated by DepthFM, to guide image generation.” Source • Screenshot\nEarlier X comparison finding, 4 October 2026. The depth map guides a trained generator; it is not copied output geometry. No greater/lesser order relative to Canny (65) is established by the reviewed sources. Keep structural guidance distinct from source retention and assess any learned depth estimator separately. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://stability.ai/news-updates/sd3-5-large-controlnets\nhttps://stability.ai/news-updates?category=Research\nhttps://docs.comfy.org/tutorials/controlnet/depth-controlnet", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-843fb18cc85c", "Operation IDs": "65, 66, 67", "Citation": "R1 Stability AI — SD3.5 Large ControlNets announcement 11262024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news-updates/sd3-5-large-controlnets" }, { "Source ID": "collected-2974b63ec542", "Operation IDs": "65, 66, 67", "Citation": "R2 Stability AI — dated release index 11262024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news-updates?category=Research" }, { "Source ID": "collected-087b1b5618e1", "Operation IDs": "66", "Citation": "docs.comfy.org — depth-controlnet", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.comfy.org/tutorials/controlnet/depth-controlnet" } ] }, { "ID": "67", "Software": "ComfyUI", "Operation": "SD3.5 Large ControlNet Blur Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-843fb18cc85c\ncollected-2974b63ec542\ncollected-6d4ac28b21e3\ncollected-c9a4a99bb1f0\ncollected-e0a4fcb4f414", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SD3.5 Large Blur ControlNet reconstructs image detail through the documented blur-conditioned generation route.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Stability AI's model card describes the Blur ControlNet (released 26 Nov 2024) as one of the ControlNet models trained for Stable Diffusion 3.5 Large and recommends tiled upscaling at ControlNet strength 0.7-0.8. The ControlNet paper defines ControlNet as spatial conditioning added to a pretrained text-to-image diffusion model, so each tile is sampled by the SD3.5 Large diffusion generator under blur conditioning; Stability's phrase 'extremely high fidelity upscaling' describes output quality and names no fidelity training objective. Not established: How far blur conditioning constrains content at a given strength; Stability publishes no measured fidelity to the source. Previous reason: Its restoration task and generative detail prior cross fidelity restoration and perceptual reconstruction. Retain 5–8; the blur amount itself does not supply t. huggingface.co, not shown • arxiv.org, v3, 26 Nov 2023 • blog.comfy.org, 26 Nov 2024", "Collected source links": "https://stability.ai/news-updates/sd3-5-large-controlnets\nhttps://stability.ai/news-updates?category=Research\nhttps://huggingface.co/stabilityai/stable-diffusion-3.5-controlnets\nhttps://arxiv.org/abs/2302.05543\nhttps://blog.comfy.org/p/sd3-5-large-controlnet", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-843fb18cc85c", "Operation IDs": "65, 66, 67", "Citation": "R1 Stability AI — SD3.5 Large ControlNets announcement 11262024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news-updates/sd3-5-large-controlnets" }, { "Source ID": "collected-2974b63ec542", "Operation IDs": "65, 66, 67", "Citation": "R2 Stability AI — dated release index 11262024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news-updates?category=Research" }, { "Source ID": "collected-6d4ac28b21e3", "Operation IDs": "67", "Citation": "huggingface.co, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://huggingface.co/stabilityai/stable-diffusion-3.5-controlnets" }, { "Source ID": "collected-c9a4a99bb1f0", "Operation IDs": "67", "Citation": "arxiv.org, v3, 26 Nov 2023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://arxiv.org/abs/2302.05543" }, { "Source ID": "collected-e0a4fcb4f414", "Operation IDs": "67", "Citation": "blog.comfy.org, 26 Nov 2024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.comfy.org/p/sd3-5-large-controlnet" } ] }, { "ID": "68", "Software": "ComfyUI", "Operation": "SD3.5 Large text-to-image generation through VAE decode", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-06d2114e2823\nsd-card\ncollected-0eacb14fff6e\ncollected-224a92876fcd\ncollected-e2e1192b6668\ncollected-85a8f1042708\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SD3.5 Large text-to-image generation includes learned prompt encoding, latent prediction and VAE decoding.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (MMDiT with text encoders and VAE). Designer controls: prompt (positive, negative); width; height; batch size; seed; steps; cfg; sampler; scheduler; checkpoint and text encoders; nodes/parameters. Returns one result (batch_size default 1); a flat image: VAE-decoded image. “Model type: MMDiT text-to-image generative model” Source • Screenshot\nEarlier X comparison finding, 4 October 2026. Entries 64 and 68 remain tied in every tested weight-and-label case; the descriptions identify the same generation-and-decode route. This analysis provides no basis for separating their X placements. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://stability.ai/news-updates/introducing-stable-diffusion-3-5\nhttps://huggingface.co/stabilityai/stable-diffusion-3.5-large\nhttps://comfyanonymous.github.io/ComfyUI_examples/sd3/sd3.5_simple_example.png\nhttps://docs.comfy.org/development/api-development/workflow-metadata\nhttps://docs.comfy.org/tutorials/basic/text-to-image\nhttps://docs.comfy.org/basic-concepts/nodes", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-06d2114e2823", "Operation IDs": "68", "Citation": "stability.ai — introducing-stable-diffusion-3-5", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news-updates/introducing-stable-diffusion-3-5" }, { "Source ID": "sd-card", "Operation IDs": "63, 64, 68", "Citation": "Stability AI. (n.d.). Stable Diffusion 3.5 Large model card. Model description and model details. Accessed 2026-10-14.", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Model description and model details", "Finding used in this review": "SD3.5 Large is text-to-image MMDiT with three fixed pretrained text encoders and QK normalization.", "Evidence limits": "An architecture name alone does not establish the operation’s scope, latitude or coordination ratings.", "Review status": "Read for this review", "Open source": "Open source", "URL": "https://huggingface.co/stabilityai/stable-diffusion-3.5-large" }, { "Source ID": "collected-0eacb14fff6e", "Operation IDs": "64, 68", "Citation": "comfyanonymous.github.io — sd3.5_simple_example.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://comfyanonymous.github.io/ComfyUI_examples/sd3/sd3.5_simple_example.png" }, { "Source ID": "collected-224a92876fcd", "Operation IDs": "64, 68", "Citation": "docs.comfy.org — workflow-metadata", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.comfy.org/development/api-development/workflow-metadata" }, { "Source ID": "collected-e2e1192b6668", "Operation IDs": "64, 68", "Citation": "ComfyUI · Text to image generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.comfy.org/tutorials/basic/text-to-image" }, { "Source ID": "collected-85a8f1042708", "Operation IDs": "64, 68", "Citation": "docs.comfy.org — nodes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.comfy.org/basic-concepts/nodes" } ] }, { "ID": "69", "Software": "Corel Painter 2023", "Operation": "Flat Color brushstroke using Enhanced Cover", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "Specified Flat Color brush subcategory, stroke path, color and brush settings; return painted marks.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Stroke path, color, Flat Color brush subcategory and settings.", "Learned determination": "None in this brush path.", "Conventional processing": "Programmed dabs, antialiasing and blending form the painted marks.", "Model contribution in the result": "No learned prediction contributes; AI Style is a separate operation.", "Why this scope score (S)": "S0/U0/J0: the documented scoped function is conventional raster processing.", "Why this latitude score (U)": "S0/U0/J0: the documented scoped function is conventional raster processing.", "Why this coordination score (J)": "S0/U0/J0: the documented scoped function is conventional raster processing.", "Comparisons and other modes": "Changing GPU speed or antialiasing quality does not create learned determination. Compare with Blender simulation: different conventional work, same retained-learning endpoint.", "Citations": "Corel Painter methods and subcategories — https://product.corel.com/help/Painter/540111162/Corel-Painter-en/Corel-Painter-Methods-and-subcategories.html", "Source IDs": "CAL2-PAINTER", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human sets the stroke and controls. Programmed dabs, antialiasing and blending produce the mark; this is not the AI Style route.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. Changing GPU speed or antialiasing quality does not create learned determination. Compare with Blender simulation: different conventional work, same retained-learning endpoint.", "Previous record: scope": "Painter's Flat Color brush applies the documented Enhanced Cover raster painting method.", "Previous record: rationale": "Antialiasing and artifact-reduction noise are conventional brush processing. Randomness does not establish a trained mechanism.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: brush variant (Flat Color); method (Cover); subcategory (Enhanced). Returns one result; a flat image. “The new Enhanced Cover methods available only in Painter 2023 take advantage of the latest technology advancements for an improved brushing experience.” Source", "Collected source links": "https://product.corel.com/help/Painter/540111155/Corel-Painter-en/Corel-Painter-Whats-new-in-Painter.html\nhttps://product.corel.com/help/Painter/540111162/Corel-Painter-en/Corel-Painter-Methods-and-subcategories.html\nhttps://product.corel.com/help/Painter/540111162/Corel-Painter-en/Corel-Painter-2023.pdf", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-69", "Sources": [ { "Source ID": "CAL2-PAINTER", "Operation IDs": "69", "Citation": "Corel Painter methods and subcategories", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Corel Painter methods and subcategories", "Finding used in this review": "Enhanced Cover dab antialiasing, color precision and blending; assessed Flat Color, not AI Style.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://product.corel.com/help/Painter/540111162/Corel-Painter-en/Corel-Painter-Methods-and-subcategories.html" } ] }, { "ID": "70", "Software": "CSM Cube", "Operation": "Single Image to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a718eb889d8a\ncollected-8e5df897e83f\ncollected-0579fea10158\ncollected-77ce4fef7996\ncollected-0565ec720d46\ncollected-ac262b1772d2\ncollected-52f1799df37c\ncollected-d6780e92f2e1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "CSM Cube Base Single Image to Mesh predicts an untextured three-dimensional asset.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=2; conditional=0; A=3. Lower: 2 + 3 × 1 ÷ (3 − 1) = 3.5. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.5; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Common Sense Machines' own 3D foundation model (Cube 2.0 generation and later), described as a blend of diffusion models, transformers and neural radiance fields, trained on NVIDIA H100 GPUs. No public model name beyond the Cube product versions. Model choice: Yes, at the geometry stage in the CSM API: Base (most consistent) and Turbo (fastest, sometimes lower quality). Not established: Service status: csm.ai and docs.csm.ai did not resolve on 2026-10-04, and one third-party source reports Cube shut down on 2026-01-05; no vendor shutdown notice was found. docs.csm.ai • csm.ai • the-decoder.com, 2026-01-24\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Base geometry model). Designer controls: image upload; geometry model (Turbo, Base); texture (None); manual segmentation; number of variations (1x to 3x); mesh topology (tris, quads); target poly count; symmetry; bounding box. Returns several to choose from (1 to 3, set by designer); editable structure: mesh; untextured with Texture None. “Select how many mesh variations you want: 1x: Generates one mesh. 2x: Generates two meshes. 3x: Generates three meshes.” Source • Screenshot", "Collected source links": "https://docs.csm.ai/?trk=public_post_main-feed-card-text\nhttps://www.metavert.io/videos/artificial-intelligence-livestream-tejas-kulkarni-common-sense-machines\nhttps://docs.csm.ai/image-to-3d/quick-primer\nhttps://www.csm.ai/blog/image-to-3d-in-seconds-is-now-better-than-ever\nhttps://the-decoder.com/google-deepmind-goes-on-acquisition-spree-with-three-ai-deals-in-one-week/\nhttps://web.archive.org/web/20251211002547/https://www.csm.ai/academy/master-the-settings-for-image-to-3d\nhttps://web.archive.org/web/20251014161443/https://docs.csm.ai/sessions/image-to-3d\nhttps://framerusercontent.com/images/EpuRVeQr8rUIM8FXOc7jZHKK2OQ.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a718eb889d8a", "Operation IDs": "70", "Citation": "CSM API v3 introduction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.csm.ai/?trk=public_post_main-feed-card-text" }, { "Source ID": "collected-8e5df897e83f", "Operation IDs": "70", "Citation": "Founder interview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.metavert.io/videos/artificial-intelligence-livestream-tejas-kulkarni-common-sense-machines" }, { "Source ID": "collected-0579fea10158", "Operation IDs": "70", "Citation": "docs.csm.ai — quick-primer", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.csm.ai/image-to-3d/quick-primer" }, { "Source ID": "collected-77ce4fef7996", "Operation IDs": "70", "Citation": "www.csm.ai — image-to-3d-in-seconds-is-now-better-than-ever", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.csm.ai/blog/image-to-3d-in-seconds-is-now-better-than-ever" }, { "Source ID": "collected-0565ec720d46", "Operation IDs": "70", "Citation": "the-decoder.com, 2026-01-24", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://the-decoder.com/google-deepmind-goes-on-acquisition-spree-with-three-ai-deals-in-one-week/" }, { "Source ID": "collected-ac262b1772d2", "Operation IDs": "70", "Citation": "Archived settings guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20251211002547/https://www.csm.ai/academy/master-the-settings-for-image-to-3d" }, { "Source ID": "collected-52f1799df37c", "Operation IDs": "70", "Citation": "Archived image-to-3D docs", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20251014161443/https://docs.csm.ai/sessions/image-to-3d" }, { "Source ID": "collected-d6780e92f2e1", "Operation IDs": "70", "Citation": "framerusercontent.com — EpuRVeQr8rUIM8FXOc7jZHKK2OQ.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://framerusercontent.com/images/EpuRVeQr8rUIM8FXOc7jZHKK2OQ.png" } ] }, { "ID": "71", "Software": "Blackmagic Design DaVinci Resolve", "Operation": "Manual roll edit of one video edit point", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-138132678af4\ncollected-7040295dacbd\ncollected-bcb124c28458", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolve's manual roll edit moves the shared boundary of adjacent clips while preserving their combined duration.", "Previous record: rationale": "The edit applies a specified time offset without learned shot selection or frame synthesis.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Trim Edit Mode (T); drag center of cut with rolling trim pointer; viewer; edit selection toggle (U). Returns one result; editable structure: timeline clips. “rolling trims remove an equal amount from one side of a cut as it takes off from the other side” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf\nhttps://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-20-Editors-Guide.pdf\nhttps://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-17-Beginners-Guide.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-138132678af4", "Operation IDs": "71, 72, 74, 75, 76, 77, 401", "Citation": "Resolve21 manual, p3204", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf" }, { "Source ID": "collected-7040295dacbd", "Operation IDs": "71", "Citation": "Resolve20 Editor guide, pp86–89", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-20-Editors-Guide.pdf" }, { "Source ID": "collected-bcb124c28458", "Operation IDs": "71", "Citation": "documents.blackmagicdesign.com — DaVinci-Resolve-17-Beginners-Guide.pdf", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-17-Beginners-Guide.pdf" } ] }, { "ID": "72", "Software": "DaVinci Resolve Studio", "Operation": "DaVinci Resolve MCP LLM Timeline Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-27abc84f9980\ncollected-5bdd68806ad1\ncollected-245f48d8629b\ncollected-d49b57a341e2\ncollected-72499ec56c00\ncollected-138132678af4\ncollected-eb1d82c3635a\ncollected-a54724cbb437\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "An external LLM selects and orders complete clips, and the Resolve MCP bridge appends them to a timeline.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. In the append operation the model determines which supplied clips or ranges appear and their order, with frame content taken unchanged from the media pool; this which-and-when determination is the REF10 kind and places the operation nearer REF10 than REF02. The interval stays open below REF10 because the LLM authors the order.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. davinci-resolve-mcp is at v4.8.27 (3 October 2026). Appending to a timeline takes whole clips through clip_ids or frame ranges through clip_infos (PR #256, 19 September 2026); Resolve 21 AI Extras and optional local Whisper and CLIP analysis exist as separate tools outside the append operation. Not established: Whether analysis tools (transcription, visual similarity) inform the clip choice in a given session is not recorded by the bridge. Previous reason: Learned temporal composition remains in the conventional timeline. The choice among whole clips is substantive, with a 2–10 synthesis-to-relational comparison envelope. github.com, 2026-10-03 • github.com, 2026-09-19 • github.com, v4.8.27", "Collected source links": "https://github.com/samuelgursky/davinci-resolve-mcp/blob/89da04b19b0dbcbf6e5a53a54fe80542a8d2ef23/README.md\nhttps://github.com/samuelgursky/davinci-resolve-mcp/blob/89da04b19b0dbcbf6e5a53a54fe80542a8d2ef23/src/granular/media_pool.py\nhttps://github.com/samuelgursky/davinci-resolve-mcp/releases\nhttps://github.com/samuelgursky/davinci-resolve-mcp/pull/256\nhttps://github.com/samuelgursky/davinci-resolve-mcp\nhttps://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf\nhttps://raw.githubusercontent.com/samuelgursky/davinci-resolve-mcp/main/docs/reference/resolve_scripting_api.txt\nhttps://raw.githubusercontent.com/samuelgursky/davinci-resolve-mcp/main/docs/SKILL.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-27abc84f9980", "Operation IDs": "72", "Citation": "github.com — README.md", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/samuelgursky/davinci-resolve-mcp/blob/89da04b19b0dbcbf6e5a53a54fe80542a8d2ef23/README.md" }, { "Source ID": "collected-5bdd68806ad1", "Operation IDs": "72", "Citation": "github.com — media_pool.py", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/samuelgursky/davinci-resolve-mcp/blob/89da04b19b0dbcbf6e5a53a54fe80542a8d2ef23/src/granular/media_pool.py" }, { "Source ID": "collected-245f48d8629b", "Operation IDs": "72", "Citation": "github.com, 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/samuelgursky/davinci-resolve-mcp/releases" }, { "Source ID": "collected-d49b57a341e2", "Operation IDs": "72", "Citation": "github.com, 2026-09-19", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/samuelgursky/davinci-resolve-mcp/pull/256" }, { "Source ID": "collected-72499ec56c00", "Operation IDs": "72", "Citation": "github.com, v4.8.27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/samuelgursky/davinci-resolve-mcp" }, { "Source ID": "collected-138132678af4", "Operation IDs": "71, 72, 74, 75, 76, 77, 401", "Citation": "Resolve21 manual, p3204", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf" }, { "Source ID": "collected-eb1d82c3635a", "Operation IDs": "72", "Citation": "raw.githubusercontent.com — resolve_scripting_api.txt", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/samuelgursky/davinci-resolve-mcp/main/docs/reference/resolve_scripting_api.txt" }, { "Source ID": "collected-a54724cbb437", "Operation IDs": "72", "Citation": "raw.githubusercontent.com — SKILL.md", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/samuelgursky/davinci-resolve-mcp/main/docs/SKILL.md" } ] }, { "ID": "73", "Software": "Blackmagic Design DaVinci Resolve Studio", "Operation": "AI Set Extender Announced Workflow", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4ad78f2d8f5c\ncollected-7e417f519e51\ncollected-1c9a1ab31eff\ncollected-4136062c9f90\ncollected-9dbf1831a7d9", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolve AI Set Extender is documented as an announced scene-extension workflow.", "Previous record: rationale": "An announcement alone does not establish the available operation or its executed learned mechanism. Preserve the catalogue entry without presenting an installed-operation coordinate; intended synthesis would be a separate conditional comparison.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Blackmagic announced Resolve FX AI Set Extender with DaVinci Resolve 20 (April 2025) as a generative feature to be available via Blackmagic Cloud; it was absent from the 20.0 release (May 2025). The DaVinci Resolve 21 New Features Guide (PDF dated 10 April 2026, 150 pages) contains no mention of Set Extender, and Newsshooter's Resolve 21.1 feature list (7 Sept 2026) contains none; one third-party blog (The Post Flow, August 2026) states it is a text-prompted scene extension in Resolve 21 without citing Blackmagic. Not established: Whether AI Set Extender ships in any Resolve build as of October 2026; if it does, which model performs the extension and whether it runs locally or through Blackmagic Cloud. cgchannel.com, 2025-05-28 • documents.blackmagicdesign.com, 2026-04-10 (PDF; no Set Exte • newsshooter.com, 2026-09-07", "Collected source links": "https://www.blackmagicdesign.com/media/partial/release/20250404-02\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf\nhttps://www.cgchannel.com/2025/05/blackmagic-design-releases-davinci-resolve-20-0/\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf\nhttps://www.newsshooter.com/2026/09/07/blackmagic-design-davinci-resolve-21-1/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4ad78f2d8f5c", "Operation IDs": "73", "Citation": "www.blackmagicdesign.com — 20250404-02", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.blackmagicdesign.com/media/partial/release/20250404-02" }, { "Source ID": "collected-7e417f519e51", "Operation IDs": "73, 74, 75, 76", "Citation": "Resolve 20 guide pp 68–73", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf" }, { "Source ID": "collected-1c9a1ab31eff", "Operation IDs": "73", "Citation": "cgchannel.com, 2025-05-28", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2025/05/blackmagic-design-releases-davinci-resolve-20-0/" }, { "Source ID": "collected-4136062c9f90", "Operation IDs": "73, 74, 75, 77", "Citation": "documents.blackmagicdesign.com, 2026-04-10 (PDF; no Set Exte", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf" }, { "Source ID": "collected-9dbf1831a7d9", "Operation IDs": "73", "Citation": "newsshooter.com, 2026-09-07", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.newsshooter.com/2026/09/07/blackmagic-design-davinci-resolve-21-1/" } ] }, { "ID": "74", "Software": "Blackmagic Design DaVinci Resolve Studio", "Operation": "DaVinci Resolve AI IntelliScript Timeline Assembly", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-138132678af4\ncollected-7e417f519e51\ncollected-4136062c9f90\ncollected-0145e3687789", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolve IntelliScript uses neural transcription and dialogue matching to assemble an initial timeline with alternatives.", "Previous record: rationale": "Learned recognition and alignment determine retained clip relationships. The operation spans fidelity and relational estimation, supporting 8–10 before manual take changes.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). The DaVinci Resolve 21 New Features Guide (April 2026) adds Final Draft (.fdx) import to IntelliScript, which compares script text with transcribed audio and generates a timeline from media pool clips; Resolve 21 also adds word-timing detection for imported transcripts. The documented learned output is transcription and alignment, the assembled timeline consists of existing clips, and no generative stage is documented. Not established: The transcription model and whether script matching is learned or rule-based text matching are not published. Documentation does not place the route within 8–10. documents.blackmagicdesign.com, April 2026, p. 80 • blackmagicdesign.com, DaVinci Resolve 21", "Collected source links": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf\nhttps://www.blackmagicdesign.com/products/davinciresolve/whatsnew", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-138132678af4", "Operation IDs": "71, 72, 74, 75, 76, 77, 401", "Citation": "Resolve21 manual, p3204", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf" }, { "Source ID": "collected-7e417f519e51", "Operation IDs": "73, 74, 75, 76", "Citation": "Resolve 20 guide pp 68–73", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf" }, { "Source ID": "collected-4136062c9f90", "Operation IDs": "73, 74, 75, 77", "Citation": "documents.blackmagicdesign.com, 2026-04-10 (PDF; no Set Exte", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf" }, { "Source ID": "collected-0145e3687789", "Operation IDs": "74, 78", "Citation": "blackmagicdesign.com, DaVinci Resolve 21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.blackmagicdesign.com/products/davinciresolve/whatsnew" } ] }, { "ID": "75", "Software": "Blackmagic Design DaVinci Resolve Studio", "Operation": "DaVinci Resolve Depth Map v2 with Map Adjustment", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-138132678af4\ncollected-7e417f519e51\ncollected-4136062c9f90", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolve Depth Map v2 estimates scene depth; conventional map adjustments transform the retained depth-derived alpha.", "Previous record: rationale": "Depth inference crosses learned spatial relations and signal reconstruction. Adjustment arithmetic retains that responsibility; retain 8–10 without an adjustment-strength weight.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). The Resolve 21 New Features Guide (April 2026) reports the Depth Map tool in Fusion as up to six times faster and introduces CineFocus, which builds on an AI generated depth map; Blackmagic does not name the depth model or say whether the speed-up reflects a new network. The learned output is a dense depth estimate used as an alpha for grading rather than a discrete recognition result. Not established: Model identity, training, and whether Resolve 21 changed the Depth Map v2 network are not published. Documentation does not place dense depth estimation within 8–10. documents.blackmagicdesign.com, April 2026, p. 144", "Collected source links": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-138132678af4", "Operation IDs": "71, 72, 74, 75, 76, 77, 401", "Citation": "Resolve21 manual, p3204", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf" }, { "Source ID": "collected-7e417f519e51", "Operation IDs": "73, 74, 75, 76", "Citation": "Resolve 20 guide pp 68–73", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf" }, { "Source ID": "collected-4136062c9f90", "Operation IDs": "73, 74, 75, 77", "Citation": "documents.blackmagicdesign.com, 2026-04-10 (PDF; no Set Exte", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf" } ] }, { "ID": "76", "Software": "Blackmagic Design DaVinci Resolve Studio", "Operation": "DaVinci Resolve Magic Mask", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7e417f519e51\ncollected-138132678af4\ncollected-d5bff0bfa72b\ncollected-76d945f03040\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolve Magic Mask v2 predicts a subject mask before tracking or direct paint correction.", "Previous record: rationale": "Neural membership determination matches the region reference. Click quantity and the subject's appearance do not determine t.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). DaVinci Resolve 21.0 (4 June 2026) adds a Render in Place option that caches a tracked Magic Mask as a traveling matte node and brings Magic Mask to the new Photo page. The documented task remains person and object isolation; no generative step is attached to Magic Mask. Not established: Blackmagic does not disclose whether the Magic Mask network changed in version 21. cgchannel.com, 2026-06-04 (v21.0) • newsshooter.com, 2026-04-13", "Collected source links": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf\nhttps://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf\nhttps://www.cgchannel.com/2026/06/blackmagic-design-releases-davinci-resolve-21-0/\nhttps://www.newsshooter.com/2026/04/13/blackmagic-design-announces-davinci-resolve-21-major-update-adds-new-photo-page/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7e417f519e51", "Operation IDs": "73, 74, 75, 76", "Citation": "Resolve 20 guide pp 68–73", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_20_New_Features_Guide.pdf" }, { "Source ID": "collected-138132678af4", "Operation IDs": "71, 72, 74, 75, 76, 77, 401", "Citation": "Resolve21 manual, p3204", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf" }, { "Source ID": "collected-d5bff0bfa72b", "Operation IDs": "76", "Citation": "cgchannel.com, 2026-06-04 (v21.0)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/06/blackmagic-design-releases-davinci-resolve-21-0/" }, { "Source ID": "collected-76d945f03040", "Operation IDs": "76", "Citation": "newsshooter.com, 2026-04-13", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.newsshooter.com/2026/04/13/blackmagic-design-announces-davinci-resolve-21-major-update-adds-new-photo-page/" } ] }, { "ID": "77", "Software": "Blackmagic Design DaVinci Resolve Studio", "Operation": "DaVinci Resolve Relight FX", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-138132678af4\ncollected-c00d2d8de157\ncollected-4136062c9f90", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolve Relight estimates surface orientation, derives illumination and applies a grade using the resulting alpha.", "Previous record: rationale": "The learned geometry estimate governs the final illumination. Conventional Gain application transmits that dependency; spatial reconstruction supports 8–10.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). The Resolve 21 New Features Guide (April 2026) reports the Relight tool in Fusion as up to two times faster and gives no model or training detail. The learned output remains a surface-orientation estimate that drives a conventional lighting grade; no generative stage is documented. Not established: Model identity, training, and whether Resolve 21 changed the Relight network are not published. Documentation does not place the route within 8–10. documents.blackmagicdesign.com, April 2026, p. 144", "Collected source links": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_18.5_New_Features_Guide.pdf\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-138132678af4", "Operation IDs": "71, 72, 74, 75, 76, 77, 401", "Citation": "Resolve21 manual, p3204", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf" }, { "Source ID": "collected-c00d2d8de157", "Operation IDs": "77, 78", "Citation": "18.5 guide p11", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_18.5_New_Features_Guide.pdf" }, { "Source ID": "collected-4136062c9f90", "Operation IDs": "73, 74, 75, 77", "Citation": "documents.blackmagicdesign.com, 2026-04-10 (PDF; no Set Exte", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_21_New_Features_Guide.pdf" } ] }, { "ID": "78", "Software": "Blackmagic Design DaVinci Resolve Studio", "Operation": "DaVinci Resolve Super Scale 2x Enhanced", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1965245173d6\ncollected-c00d2d8de157\ncollected-12aa6dd96200\ncollected-0145e3687789", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolve Super Scale 2x Enhanced uses its neural enlargement route with sharpness and noise controls.", "Previous record: rationale": "Learned detail reconstruction spans perceptual reconstruction and fidelity restoration. The undisclosed training objective supports 5–8 without a precise point.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Blackmagic describes Super Scale 2x Enhanced (added in Resolve 18.5) as a DaVinci Neural Engine mode for extremely high quality 2x output with its own sharpness and noise reduction sliders. The DaVinci Resolve 21 What's New page adds AI UltraSharpen as a separate Neural Engine sharpening and upscaling tool and describes no change to Super Scale. Not established: Network architecture and training objective of Super Scale 2x Enhanced, including whether it is trained with pixel loss or with perceptual/adversarial loss. redsharknews.com, Resolve 18.5 beta 5 • blackmagicdesign.com, DaVinci Resolve 21", "Collected source links": "https://forum.blackmagicdesign.com/viewtopic.php?f=21&t=179277\nhttps://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_18.5_New_Features_Guide.pdf\nhttps://www.redsharknews.com/davinci-resolve-18.5-beta-5-adds-yet-more-features\nhttps://www.blackmagicdesign.com/products/davinciresolve/whatsnew", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1965245173d6", "Operation IDs": "78", "Citation": "Resolve 18.5 release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://forum.blackmagicdesign.com/viewtopic.php?f=21&t=179277" }, { "Source ID": "collected-c00d2d8de157", "Operation IDs": "77, 78", "Citation": "18.5 guide p11", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_18.5_New_Features_Guide.pdf" }, { "Source ID": "collected-12aa6dd96200", "Operation IDs": "78", "Citation": "redsharknews.com, Resolve 18.5 beta 5", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.redsharknews.com/davinci-resolve-18.5-beta-5-adds-yet-more-features" }, { "Source ID": "collected-0145e3687789", "Operation IDs": "74, 78", "Citation": "blackmagicdesign.com, DaVinci Resolve 21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.blackmagicdesign.com/products/davinciresolve/whatsnew" } ] }, { "ID": "79", "Software": "Descript", "Operation": "Transcript-directed cuts on generated video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-44ac135e0ca5\ncollected-19ffbb7d828a\ncollected-2ee6e1a80f82\ncollected-2882799347e0\ncollected-c243c0efaa3d\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Descript deletes or moves an exactly selected span using an already supplied aligned transcript.", "Previous record: rationale": "This isolated command applies an existing time mapping conventionally. Transcription, inferred range selection and separately invoked Regenerate are outside its scope.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: text selection in script; Delete; Ignore (strikethrough); cut and paste to move; edit boundaries. Returns one result; editable structure: script-linked timeline. “Deleting removes text and its associated media from view.” Source • Screenshot", "Collected source links": "https://help.descript.com/getting-started/edit-like-a-doc\nhttps://help.descript.com/script-editing/delete-vs-ignore\nhttps://mintcdn.com/descript-5bf56f3f/1OcToTDEN_1-oRIu/images/external/descript-24762062673933-2987de5601.png\nhttps://help.descript.com/working-in-projects/duplicate\nhttps://help.descript.com/visuals/captions", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-44ac135e0ca5", "Operation IDs": "79", "Citation": "help.descript.com — edit-like-a-doc", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.descript.com/getting-started/edit-like-a-doc" }, { "Source ID": "collected-19ffbb7d828a", "Operation IDs": "79", "Citation": "help.descript.com — delete-vs-ignore", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.descript.com/script-editing/delete-vs-ignore" }, { "Source ID": "collected-2ee6e1a80f82", "Operation IDs": "79", "Citation": "mintcdn.com — descript-24762062673933-2987de5601.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://mintcdn.com/descript-5bf56f3f/1OcToTDEN_1-oRIu/images/external/descript-24762062673933-2987de5601.png" }, { "Source ID": "collected-2882799347e0", "Operation IDs": "79", "Citation": "help.descript.com — duplicate", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.descript.com/working-in-projects/duplicate" }, { "Source ID": "collected-c243c0efaa3d", "Operation IDs": "79", "Citation": "help.descript.com — captions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.descript.com/visuals/captions" } ] }, { "ID": "80", "Software": "Descript", "Operation": "Underlord rough-cut assembly of generated video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d9ef7828eeaf\ncollected-27dabfba83a9\ncollected-982e9690679f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Underlord uses an LLM to choose media ranges and apply an initial rough cut through editing tools.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. In the rough cut the model determines which ranges of the supplied media remain and in what order, with frame and audio content taken from the source; this is the REF10 kind of determination and places the operation nearer REF10 than REF02. The interval stays open below REF10 because the LLM authors the edit order and selection from a prompt.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. Descript Help (undated, current October 2026) shows Underlord's model is user-selectable among Claude, Gemini, GPT and Grok variants; Anthropic's Descript case study names Claude Sonnet 4.5 as the January 2026 self-serve default. Rough-cut steps cut ranges from supplied recordings; image, video, avatar and speech generation is a separate request path (Descript primer, updated 29 July 2026). Not established: Whether current rough-cut requests insert generated B-roll or speech by default. Previous reason: Learned editorial range selection survives the cuts. The 2–10 comparison reflects semantic and temporal construction; supplying an aligned transcript does not remove that responsibility. help.descript.com, undated • claude.com, January 2026 • descript.com, updated 2026-07-29", "Collected source links": "https://help.descript.com/getting-started/underlord-beta-your-ai-co-editor-in-descript\nhttps://www.descript.com/blog/article/underlord-ai-video-editor-primer\nhttps://claude.com/customers/descript", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d9ef7828eeaf", "Operation IDs": "80", "Citation": "help.descript.com, undated", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.descript.com/getting-started/underlord-beta-your-ai-co-editor-in-descript" }, { "Source ID": "collected-27dabfba83a9", "Operation IDs": "80", "Citation": "descript.com, updated 2026-07-29", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.descript.com/blog/article/underlord-ai-video-editor-primer" }, { "Source ID": "collected-982e9690679f", "Operation IDs": "80", "Citation": "claude.com, January 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://claude.com/customers/descript" } ] }, { "ID": "81", "Software": "Dreamina by CapCut", "Operation": "Seedance 2.5 Text to Video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-78d1ddb7f457\ncollected-06eaab2c235d\ncollected-57297216286a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Dreamina invokes Seedance 2.5 to generate an initial video clip through its learned multimodal generation framework.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Seedance 2.5). Designer controls: prompt; model choice; aspect ratio; duration (to 30 s; 180 s beta long-video mode); multimodal references (images, video, audio); R2V motion control. Returns count not documented; a flat video. “adjust aspect ratio and duration based on your platform requirements” Source", "Collected source links": "https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5\nhttps://lf3-static.bytednsdoc.com/obj/eden-cn/lapzild-tss/ljhwZthlaukjlkulzlp/transparency/20260803/Public%20Summary%20of%20Training%20Data%20Content%20for%20Seedance%202.5.pdf\nhttps://dreamina.capcut.com/seedance/seedance-2-5", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-78d1ddb7f457", "Operation IDs": "81", "Citation": "seed.bytedance.com — one-take-creation-flexible-referencing-introducing-seedance-2-5", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5" }, { "Source ID": "collected-06eaab2c235d", "Operation IDs": "81", "Citation": "lf3-static.bytednsdoc.com — Public Summary of Training Data Content for Seedance 2.5.pdf", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://lf3-static.bytednsdoc.com/obj/eden-cn/lapzild-tss/ljhwZthlaukjlkulzlp/transparency/20260803/Public%20Summary%20of%20Training%20Data%20Content%20for%20Seedance%202.5.pdf" }, { "Source ID": "collected-57297216286a", "Operation IDs": "81", "Citation": "dreamina.capcut.com — seedance-2-5", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dreamina.capcut.com/seedance/seedance-2-5" } ] }, { "ID": "82", "Software": "DxO PureRAW", "Operation": "PureRAW 6 DeepPRIME XD3 Denoising and Demosaicing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-661d614b9178\ncollected-28540c4fdf61\ncollected-2a7e3d66738c\ncollected-c089d72b9232\ncollected-ef5934b10f82\ncollected-456db53686b4\ncollected-f5d2ba4373d4\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "PureRAW 6 DeepPRIME XD3 jointly estimates denoised and demosaiced image information for the selected RAW route.", "Previous record: rationale": "Trained photographic signal recovery matches fidelity restoration. Separate geometric corrections are excluded; current losses and weights are not inferred.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). DxO PureRAW 6 (March 2026) ships DeepPRIME XD3 as a single convolutional network that performs denoising, demosaicing and chromatic aberration correction together, trained on matched pairs of raw sensor data and high-fidelity reference images. Chromatic aberration correction now sits inside the learned stage, whereas the record lists the trained task as joint denoising and demosaicing. Not established: Loss terms are not published. DxO pages refused fetch (HTTP 403); DxO wording comes from search results and a third-party review. dxo.com, 2026 • toddmarsh.com, 2026 • photoworkout.com, 2026", "Collected source links": "https://www.dxo.com/en/news/deepprime-xd3-fourth-generation/\nhttps://www.dxo.com/en/technology/deepprime/\nhttps://www.dxo.com/news/deepprime-xd3-fourth-generation/\nhttps://toddmarsh.com/dxo-pureraw-6-review/\nhttps://www.photoworkout.com/dxo-pureraw-6-deal/\nhttps://web.archive.org/web/20250915145205id_/https://support.dxo.com/hc/en-us/articles/4433333661201-Should-I-delete-my-original-image-after-processing-it-with-DxO-PureRAW\nhttps://download-center.dxo.com/Support/docs/PureRAW_v6/user-guide/PR6_manual_EN.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-661d614b9178", "Operation IDs": "82", "Citation": "www.dxo.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.dxo.com/en/news/deepprime-xd3-fourth-generation/" }, { "Source ID": "collected-28540c4fdf61", "Operation IDs": "82", "Citation": "www.dxo.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.dxo.com/en/technology/deepprime/" }, { "Source ID": "collected-2a7e3d66738c", "Operation IDs": "82", "Citation": "dxo.com, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.dxo.com/news/deepprime-xd3-fourth-generation/" }, { "Source ID": "collected-c089d72b9232", "Operation IDs": "82", "Citation": "toddmarsh.com, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://toddmarsh.com/dxo-pureraw-6-review/" }, { "Source ID": "collected-ef5934b10f82", "Operation IDs": "82", "Citation": "photoworkout.com, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.photoworkout.com/dxo-pureraw-6-deal/" }, { "Source ID": "collected-456db53686b4", "Operation IDs": "82", "Citation": "web.archive.org — 4433333661201-Should-I-delete-my-original-image-after-processing-it-with-DxO-PureRAW", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20250915145205id_/https://support.dxo.com/hc/en-us/articles/4433333661201-Should-I-delete-my-original-image-after-processing-it-with-DxO-PureRAW" }, { "Source ID": "collected-f5d2ba4373d4", "Operation IDs": "82", "Citation": "download-center.dxo.com — PR6_manual_EN.pdf", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://download-center.dxo.com/Support/docs/PureRAW_v6/user-guide/PR6_manual_EN.pdf" } ] }, { "ID": "83", "Software": "DxO PureRAW 6", "Operation": "Geometric lens-distortion correction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-303a8435bf39\ncollected-19a99c3c7ecf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "PureRAW geometric correction applies a measured camera/lens module to correct distortion.", "Previous record: rationale": "The isolated calibrated geometric transformation is conventional. DeepPRIME's separate reconstruction is not counted inside this stage.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Lens distortion correction on/off; crop option (Image cropped to original ratio, Maximum Rectangle, Complete image area); DxO Module download; output format and destination. Returns one result; a flat image. “the image is cropped so that just the bleed is eliminated, regardless of the proportions of the original” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.dxo.com/en/dxo-module/\nhttps://userguides.dxo.com/pureraw/en/processing-tools-and-settings/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-303a8435bf39", "Operation IDs": "83", "Citation": "www.dxo.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.dxo.com/en/dxo-module/" }, { "Source ID": "collected-19a99c3c7ecf", "Operation IDs": "83", "Citation": "userguides.dxo.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://userguides.dxo.com/pureraw/en/processing-tools-and-settings/" } ] }, { "ID": "84", "Software": "Figma", "Operation": "Agent Image Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3ac496766d7e\ncollected-a328dfd59f1e\ncollected-5b784b91c5c0\ncollected-6c7c1f797ccf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Figma's agent invokes image generation and places the generated raster or image fill in the design.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented for the Figma agent specifically. Figma's image generation across products uses Google models: Gemini 2.5 Flash (Make Image and Edit Image, October 2025), Imagen 4 named among models leveraged, and Nano Banana Pro (Gemini 3 Pro Image) in all Figma products that generate images (November 2025). Model choice: Yes for Make or edit image (dropdown of available models). Not established: Which image model the Figma agent (launched 2026-05-20) calls by default and whether the user can override it from the agent. help.figma.com, page date not shown • figma.com, 2025-11-20 • googlecloudpresscorner.com, 2025-10-09\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (model not named on agent page). Designer controls: prompt (typed or dictated); selected layer (opens on-canvas prompt box); added context (design URLs, images, files). Returns count not documented; form not documented. “Generate images, remove backgrounds, boost resolution, expand existing images, isolate and erase objects” Source • Screenshot", "Collected source links": "https://help.figma.com/hc/en-us/articles/37998629035799-Work-with-the-Figma-agent-in-design-files\nhttps://help.figma.com/hc/en-us/articles/24004542669463-Make-or-edit-an-image-with-AI\nhttps://www.figma.com/blog/creativity-meets-precision-with-gemini-3-pro-with-nano-banana/\nhttps://www.googlecloudpresscorner.com/2025-10-09-Figma-and-Google-Cloud-Team-Up-to-Empower-Next-Generation-of-Creative-Builders", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3ac496766d7e", "Operation IDs": "84", "Citation": "help.figma.com — 37998629035799-Work-with-the-Figma-agent-in-design-files", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.figma.com/hc/en-us/articles/37998629035799-Work-with-the-Figma-agent-in-design-files" }, { "Source ID": "collected-a328dfd59f1e", "Operation IDs": "84", "Citation": "help.figma.com, page date not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.figma.com/hc/en-us/articles/24004542669463-Make-or-edit-an-image-with-AI" }, { "Source ID": "collected-5b784b91c5c0", "Operation IDs": "84", "Citation": "figma.com, 2025-11-20", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.figma.com/blog/creativity-meets-precision-with-gemini-3-pro-with-nano-banana/" }, { "Source ID": "collected-6c7c1f797ccf", "Operation IDs": "84", "Citation": "googlecloudpresscorner.com, 2025-10-09", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.googlecloudpresscorner.com/2025-10-09-Figma-and-Google-Cloud-Team-Up-to-Empower-Next-Generation-of-Creative-Builders" } ] }, { "ID": "85", "Software": "Figma", "Operation": "Weave Replace Background", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1c914c09a07c\ncollected-74a22d56eb08\ncollected-d40600681a1c\ncollected-349e8b77c0d6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Figma's embedded Weave Replace Background tool synthesizes a replacement background for an image.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=5; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented for the Weave tool in Figma Design. Inside Figma Weave (formerly Weavy) the edit-image node list includes \"Bria Replace Background\" (Bria AI) and a separate \"Replace Background\" node whose provider is not named. Model choice: In Figma Weave, yes: the user picks the node (Bria Replace Background, Replace Background, or general edit models such as Gemini, FLUX, SeedEdit, Ideogram). Not established: Which node or model the Replace background Weave tool runs inside Figma Design; the provider behind the unbranded Replace Background node in Weave. help.figma.com, page date not shown • help.weavy.ai\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Weave tool; model not named). Designer controls: selected canvas image; text prompt describing the new background; number of runs. Returns several to choose from (designer sets number of runs; maximum not stated); form not documented. “Describe a new background and replace the existing one in your image.” Source • Screenshot", "Collected source links": "https://help.figma.com/hc/en-us/articles/40779260614935-Use-Weave-tools-in-Figma\nhttps://www.figma.com/blog/connecting-figma-and-weave/\nhttps://help.weavy.ai/en/articles/12343904-edit-image-models-comparison\nhttps://cdn.sanity.io/images/599r6htc/regionalized/4822bdbeaac6cd01051e0ef32b3415bf311714b9-3264x1836.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1c914c09a07c", "Operation IDs": "85", "Citation": "help.figma.com, page date not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.figma.com/hc/en-us/articles/40779260614935-Use-Weave-tools-in-Figma" }, { "Source ID": "collected-74a22d56eb08", "Operation IDs": "85", "Citation": "www.figma.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.figma.com/blog/connecting-figma-and-weave/" }, { "Source ID": "collected-d40600681a1c", "Operation IDs": "85", "Citation": "help.weavy.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.weavy.ai/en/articles/12343904-edit-image-models-comparison" }, { "Source ID": "collected-349e8b77c0d6", "Operation IDs": "85", "Citation": "cdn.sanity.io — 4822bdbeaac6cd01051e0ef32b3415bf311714b9-3264x1836.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.sanity.io/images/599r6htc/regionalized/4822bdbeaac6cd01051e0ef32b3415bf311714b9-3264x1836.png" } ] }, { "ID": "86", "Software": "Final Cut Pro (Apple)", "Operation": "Magnetic Mask Neural Selection", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-42dce1606523\ncollected-2a1c21787ca7\ncollected-777d379490b3\ncollected-88906445fe75", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Final Cut Pro Magnetic Mask predicts and tracks foreground membership over a video clip.", "Previous record: rationale": "The retained mask matches learned region determination. Temporal analysis is recorded without treating repeated execution as greater X magnitude.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Final Cut Pro 12.3 (30 June 2026) added Auto Mask, a separate on-device tool that isolates recognized elements such as skin, sky and clothing; Magnetic Mask continues alongside it. Apple release notes since 11.1 (27 March 2025) list only fixes and performance work for Magnetic Mask. Not established: Apple does not disclose the Magnetic Mask model or whether 12.x updated it. support.apple.com, 2026-06-30 (FCP 12.3) • cined.com, 2026-06/07", "Collected source links": "https://support.apple.com/guide/final-cut-pro/ver1d67e3a53/mac\nhttps://support.apple.com/en-ca/guide/final-cut-pro/ver43e886e74/mac\nhttps://support.apple.com/en-us/102825\nhttps://www.cined.com/apple-creator-studio-update-adds-generate-captions-edit-detection-and-auto-mask-to-final-cut-pro/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-42dce1606523", "Operation IDs": "86", "Citation": "Apple Add Magnetic Masks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.apple.com/guide/final-cut-pro/ver1d67e3a53/mac" }, { "Source ID": "collected-2a1c21787ca7", "Operation IDs": "86", "Citation": "Apple Edit Magnetic Masks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.apple.com/en-ca/guide/final-cut-pro/ver43e886e74/mac" }, { "Source ID": "collected-777d379490b3", "Operation IDs": "86", "Citation": "support.apple.com, 2026-06-30 (FCP 12.3)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.apple.com/en-us/102825" }, { "Source ID": "collected-88906445fe75", "Operation IDs": "86", "Citation": "cined.com, 2026-06/07", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cined.com/apple-creator-studio-update-adds-generate-captions-edit-detection-and-auto-mask-to-final-cut-pro/" } ] }, { "ID": "87", "Software": "Final Cut Pro (Apple)", "Operation": "Magnetic Timeline Video Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-32619d07e7f6\ncollected-e7d59b8d14e1\ncollected-228d06c33031", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Final Cut Pro performs a manually chosen insert move and the Magnetic Timeline's prescribed ripple consequences.", "Previous record: rationale": "Automatic timing updates follow editing rules without learned content selection. Their convenience does not establish a trained predictor.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: drag clip; Position tool (P); Edit > Move (Left, Right, Up, Down). Returns one result; editable structure: timeline clips. “Final Cut Pro ripples clips (and any connected clips) in the timeline to make room for the repositioned clips” Source • Screenshot", "Collected source links": "https://support.apple.com/guide/final-cut-pro/intro-to-the-magnetic-timeline-verb8fcfc133/mac\nhttps://support.apple.com/guide/final-cut-pro/arrange-clips-in-the-timeline-verc147f195/12.3/mac/15.6\nhttps://help.apple.com/assets/69433400CD7B104B7D071A52/694334052F74CB432B0CC82C/en_US/378c00d9719e7ce4281fb3a1dd15759d.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-32619d07e7f6", "Operation IDs": "87", "Citation": "Apple Magnetic Timeline", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.apple.com/guide/final-cut-pro/intro-to-the-magnetic-timeline-verb8fcfc133/mac" }, { "Source ID": "collected-e7d59b8d14e1", "Operation IDs": "87", "Citation": "Apple Arrange Clips", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.apple.com/guide/final-cut-pro/arrange-clips-in-the-timeline-verc147f195/12.3/mac/15.6" }, { "Source ID": "collected-228d06c33031", "Operation IDs": "87", "Citation": "help.apple.com — 378c00d9719e7ce4281fb3a1dd15759d.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.apple.com/assets/69433400CD7B104B7D071A52/694334052F74CB432B0CC82C/en_US/378c00d9719e7ce4281fb3a1dd15759d.png" } ] }, { "ID": "88", "Software": "Final Cut Pro (Apple)", "Operation": "Transcribe to Captions / Generate Closed Captions", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-abbc057848e4\ncollected-43382f9ec703\ncollected-9769dc172b45", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Final Cut Pro recognizes speech and produces timed closed-caption clips before manual correction.", "Previous record: rationale": "Transcription and temporal alignment cross fidelity and relational estimation, giving 8–10 with no selected midpoint.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). The Final Cut Pro for Mac 12.4 user guide names the feature Generate Captions, with a Closed Captions output and a new Subtitles output; both transcribe each clip with a downloaded language model (U.S. English, Apple silicon) and connect timed clips to their source. Apple's November 2024 announcement describes the model as an Apple-trained large language model that transcribes spoken audio; the documented output is a time-aligned transcript with no other generated text. Not established: Apple does not publish the model's training or decoding constraints. Documentation does not place language-model transcription within 8–10. support.apple.com, Final Cut Pro for Mac 12.4 guide • support.apple.com, Final Cut Pro for Mac 12.4 guide • apple.com, 2024-11", "Collected source links": "https://www.apple.com/newsroom/2024/11/final-cut-pro-11-begins-a-new-chapter-for-video-editing-on-mac/\nhttps://support.apple.com/guide/final-cut-pro/create-closed-captions-vere399dab5e/mac\nhttps://support.apple.com/guide/final-cut-pro/generate-subtitles-automatically-verd6d52c4e8/mac", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-abbc057848e4", "Operation IDs": "88", "Citation": "Apple Final Cut Pro 11 launch", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.apple.com/newsroom/2024/11/final-cut-pro-11-begins-a-new-chapter-for-video-editing-on-mac/" }, { "Source ID": "collected-43382f9ec703", "Operation IDs": "88", "Citation": "Apple Create Closed Captions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.apple.com/guide/final-cut-pro/create-closed-captions-vere399dab5e/mac" }, { "Source ID": "collected-9769dc172b45", "Operation IDs": "88", "Citation": "support.apple.com, Final Cut Pro for Mac 12.4 guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.apple.com/guide/final-cut-pro/generate-subtitles-automatically-verd6d52c4e8/mac" } ] }, { "ID": "89", "Software": "Finch", "Operation": "Enterprise AI Unit Plan Generation and Linked Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a00556afe1a2\ncollected-1980ce0ee37a\ncollected-241ad235e4ad\ncollected-76987abf4bf6\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Finch Generated Plans constructs a unit layout using the documented generative plan route and its design relationships.", "Previous record: rationale": "Learned plan construction retains semantic and spatial choices. The current references support a broad 2–10 comparison rather than a specific architectural-generation point.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). Finch documentation (undated, current October 2026) describes two routes in Enterprise Generate Unit Plan: similar library plans adjusted by 'finch algorithms', and 'AI' generation based on the firm's plan library style and rules. AEC Magazine (2023) attributes generation to optimisation algorithms over graph rules and Architosh (September 2024) to algorithmic-aided design; no source located names a trained model. Not established: Whether the generative route uses a trained model or rule-based optimisation; if no trained model is shown, retained learning would become unresolved. docs.finch3d.com, undated • aecmag.com, 2023-01-30", "Collected source links": "https://docs.finch3d.com/docs/projects-and-variants/unit-editor/enterprise-generate-unit-plan\nhttps://docs.finch3d.com/courses/finch-101/finch-101-generating-unit-plans\nhttps://aecmag.com/ai/finch3d-starts-to-sing/\nhttps://docs.finch3d.com/docs/projects-and-variants/iterate-with-variants", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a00556afe1a2", "Operation IDs": "89", "Citation": "Finch Enterprise Generate Unit Plan", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.finch3d.com/docs/projects-and-variants/unit-editor/enterprise-generate-unit-plan" }, { "Source ID": "collected-1980ce0ee37a", "Operation IDs": "89", "Citation": "Finch 101 Generating Unit Plans", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.finch3d.com/courses/finch-101/finch-101-generating-unit-plans" }, { "Source ID": "collected-241ad235e4ad", "Operation IDs": "89", "Citation": "aecmag.com, 2023-01-30", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://aecmag.com/ai/finch3d-starts-to-sing/" }, { "Source ID": "collected-76987abf4bf6", "Operation IDs": "89", "Citation": "docs.finch3d.com — iterate-with-variants", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.finch3d.com/docs/projects-and-variants/iterate-with-variants" } ] }, { "ID": "90", "Software": "Adobe Firefly (web app)", "Operation": "AI Assistant Image Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-447dfad387aa\ncollected-ed78daf2df97\ncollected-53ad144e76bd", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Firefly AI Assistant invokes an image generator through its model-routing and tool orchestration process.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (assistant routes the request to a generator). Designer controls: prompt (conversation); Creative Skill choice. Returns count not documented; form not documented. “you describe what you want to create and the assistant orchestrates and executes complex, multi-step workflows” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://blog.adobe.com/en/publish/2026/04/15/introducing-firefly-ai-assistant-new-way-create-with-our-creative-agent\nhttps://helpx.adobe.com/firefly/web/firefly-ai-assistant/firefly-ai-assistant-overview.html\nhttps://helpx.adobe.com/firefly/web/firefly-ai-assistant/ai-assistant-faq.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-447dfad387aa", "Operation IDs": "90", "Citation": "Announcement and lineage", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2026/04/15/introducing-firefly-ai-assistant-new-way-create-with-our-creative-agent" }, { "Source ID": "collected-ed78daf2df97", "Operation IDs": "90", "Citation": "operational overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/firefly/web/firefly-ai-assistant/firefly-ai-assistant-overview.html" }, { "Source ID": "collected-53ad144e76bd", "Operation IDs": "90, 252", "Citation": "Firefly AI Assistant FAQ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/firefly/web/firefly-ai-assistant/ai-assistant-faq.html" } ] }, { "ID": "91", "Software": "Adobe Firefly", "Operation": "Image 5 Text to Image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-40f5576f0721\ncollected-6920b126bb67\ncollected-e99771b07545\ncollected-b270e65948fb", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Firefly Image 5 generates the image through the explicitly selected Adobe model.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Firefly Image 5, designer-selected). Designer controls: prompt; model choice; aspect ratio; style reference; composition reference; visual intensity; effects. Returns one result; a flat image. “Firefly Image 5: A single image is generated” Source • Screenshot Quote taken from a search result for the vendor page.", "Collected source links": "https://news.adobe.com/news/2025/10/adobe-max-2025-firefly\nhttps://helpx.adobe.com/ca/firefly/web/work-with-images/generate-images/generate-images-from-text-descriptions.html\nhttps://helpx.adobe.com/firefly/web/work-with-images/generate-images/generate-images-from-text-descriptions.html\nhttps://helpx.adobe.com/content/dam/help/en/community/firefly/firefly-5-guide/Screenshot-2025-04-30-at-4-11-18-PM.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-40f5576f0721", "Operation IDs": "91", "Citation": "Adobe · Text to Image", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://news.adobe.com/news/2025/10/adobe-max-2025-firefly" }, { "Source ID": "collected-6920b126bb67", "Operation IDs": "91", "Citation": "Adobe · Image generation models and controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/ca/firefly/web/work-with-images/generate-images/generate-images-from-text-descriptions.html" }, { "Source ID": "collected-e99771b07545", "Operation IDs": "91", "Citation": "helpx.adobe.com — generate-images-from-text-descriptions.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/firefly/web/work-with-images/generate-images/generate-images-from-text-descriptions.html" }, { "Source ID": "collected-b270e65948fb", "Operation IDs": "91", "Citation": "helpx.adobe.com — Screenshot-2025-04-30-at-4-11-18-PM.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/content/dam/help/en/community/firefly/firefly-5-guide/Screenshot-2025-04-30-at-4-11-18-PM.png" } ] }, { "ID": "92", "Software": "Adobe Firefly (web app)", "Operation": "Text to Video with Firefly Video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f1084ea3e858\ncollected-6d48d5397e80", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Firefly Video generates a new clip through the explicitly selected Adobe text-to-video route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Firefly Video, default model). Designer controls: prompt; model choice (Firefly Video or partner models); aspect ratio; resolution; camera motion presets; camera angle; shot size; motion reference video; start and end keyframe images. Returns count not documented; a flat video. “camera angle, shot angle, and motion” Source • Screenshot Quote taken from a search result for the vendor page.", "Collected source links": "https://blog.adobe.com/en/publish/2025/02/12/meet-firefly-video-model-ai-powered-creation-with-unparalleled-creative-control\nhttps://helpx.adobe.com/firefly/web/work-with-audio-and-video/work-with-video/generate-videos-using-text-prompts.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f1084ea3e858", "Operation IDs": "92", "Citation": "Adobe · Text to Video / Generate Video", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2025/02/12/meet-firefly-video-model-ai-powered-creation-with-unparalleled-creative-control" }, { "Source ID": "collected-6d48d5397e80", "Operation IDs": "92", "Citation": "Adobe · Firefly Video generation and output", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/firefly/web/work-with-audio-and-video/work-with-video/generate-videos-using-text-prompts.html" } ] }, { "ID": "93", "Software": "Adobe Firefly Boards", "Operation": "Generative Fill on Board Images", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5aafa6d18fc5\ncollected-97c0d296f5d6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Firefly Boards invokes Generative Fill on one board image.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=4; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Firefly fill). Designer controls: brush (size, hardness); select more / select less; prompt; Insert or Remove. Returns several to choose from (count not stated; More adds variations); a flat image. “add new elements or remove existing ones from images non-destructively using simple brush strokes and text prompts” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/firefly/web/create-mood-boards/firefly-boards/edit-images.html\nhttps://helpx.adobe.com/firefly/web/work-with-images/edit-images/generative-fill.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5aafa6d18fc5", "Operation IDs": "93", "Citation": "Adobe: Edit images in Boards", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/firefly/web/create-mood-boards/firefly-boards/edit-images.html" }, { "Source ID": "collected-97c0d296f5d6", "Operation IDs": "93", "Citation": "Adobe: Use generative fill", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/firefly/web/work-with-images/edit-images/generative-fill.html" } ] }, { "ID": "94", "Software": "Autodesk Flame", "Operation": "Inference Node LaMa Masked Inpainting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.4", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a8f6aefd8773\ncollected-3c2986b5b4ed\ncollected-d3338b5e2b7e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Flame's documented LaMa Inference-node configuration synthesizes missing image content inside a mask.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 4 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.0–4.4; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=4; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (6 − 1) = 4.4. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (LaMa ONNX model loaded by the artist). Designer controls: model file (ONNX plus JSON sidecar); Front and Matte inputs; Grey To White matte shader; CPU or GPU; bit depth. Returns one result; a flat video: Result and OutMatte outputs. “The Inference node doesn't offer any controls over the models themselves.” Source • Screenshot", "Collected source links": "https://help.autodesk.com/cloudhelp/2025/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2025/What-s-New-in-2025-2/wn-2025-2-inference.html\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2026/What-s-New-in-2026/wn-2026-machine-learning.html\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Flame-EffectsToolsReference/files/inference_node.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a8f6aefd8773", "Operation IDs": "94", "Citation": "Autodesk · Inference Node", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2025/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2025/What-s-New-in-2025-2/wn-2025-2-inference.html" }, { "Source ID": "collected-3c2986b5b4ed", "Operation IDs": "94, 98", "Citation": "Flame 2026 · Inference models and masks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2026/What-s-New-in-2026/wn-2026-machine-learning.html" }, { "Source ID": "collected-d3338b5e2b7e", "Operation IDs": "94", "Citation": "help.autodesk.com — inference_node.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Flame-EffectsToolsReference/files/inference_node.html" } ] }, { "ID": "95", "Software": "Autodesk Flame", "Operation": "ML Depth Map Active Node", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-78cb5ef5965d\ncollected-f8616786622a\ncollected-cfe6e9d0e37a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Flame's active ML Depth Map route predicts relative scene depth.", "Previous record: rationale": "Spatial inference crosses region relations and signal reconstruction, supporting 8–10. MiDaS or DPT is not attributed to Flame without product evidence.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). Flame 2027 (April 2026) replaces the ML Depth Map Matchbox with two internal models, ML Depth Indoor and ML Depth Outdoor, and adds Load Model for an external ONNX depth model. Autodesk names neither internal model nor its training; the learned output is a dense depth estimate used for depth-derived mattes rather than a discrete recognition result, and no generative stage is documented. Not established: Model identity, relative or metric output, and training objective for the Indoor and Outdoor models are not published. Documentation does not place dense depth estimation within 8–10. help.autodesk.com, Flame 2027 • digitalproduction.com, 2026-04-20", "Collected source links": "https://digitalproduction.com/2026/04/20/autodesk-flame-2027-metadata-and-depth-maps/\nhttps://help.autodesk.com/cloudhelp/2025/ENU/Flame-WhatsNew/files/Previous-Releases/2020/GUID-FE433726-ACE9-4610-B427-A89D79177C14.html\nhttps://help.autodesk.com/cloudhelp/2027/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2027/What-s-New-in-2027/wn-2027-ml.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-78cb5ef5965d", "Operation IDs": "95", "Citation": "Digital Production · ML Depth Map", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://digitalproduction.com/2026/04/20/autodesk-flame-2027-metadata-and-depth-maps/" }, { "Source ID": "collected-f8616786622a", "Operation IDs": "95", "Citation": "Flame · Machine learning maps", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2025/ENU/Flame-WhatsNew/files/Previous-Releases/2020/GUID-FE433726-ACE9-4610-B427-A89D79177C14.html" }, { "Source ID": "collected-cfe6e9d0e37a", "Operation IDs": "95", "Citation": "help.autodesk.com, Flame 2027", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2027/What-s-New-in-2027/wn-2027-ml.html" } ] }, { "ID": "96", "Software": "Autodesk Flame", "Operation": "ML Face Normal Map Active Node", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9fe731185ef3\ncollected-774f3f4ef585\ncollected-4d47e49d3bff", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Flame Face Analysis predicts facial surface normals for the active normal map.", "Previous record: rationale": "Learned geometric estimation supports an 8–10 comparison. The region of interest is an execution control, not a fraction of inference.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). Flame 2026 Action help states that Normal map face analysis uses machine learning to detect a face and generate its normals, with a Region of Interest control. The only training description is the 2020 account of about a hundred thousand synthetic beauty images paired with normal maps, which describes supervised estimation of surface normals; no generative stage is documented. Not established: Whether the 2020 model has been retrained or replaced in Flame 2026 or 2027 is not documented. Documentation does not place dense normal estimation within 8–10. help.autodesk.com, Flame 2026 • beforesandafters.com, 2020-04-17", "Collected source links": "https://beforesandafters.com/2020/04/17/inside-flames-new-machine-learning-powered-human-face-segmentation/\nhttps://help.autodesk.com/cloudhelp/2020/ENU/Flame-Compositingin3d/files/GUID-D7D7C9B7-FE85-497A-BBC2-455BEB112FC4.htm\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Flame-Action/files/Action-Shading-and-Textures/GUID-1A2F6A6D-F2A1-4BE7-8B36-1D2F1773DB2F.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9fe731185ef3", "Operation IDs": "96", "Citation": "befores & afters · Face Normal Map", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://beforesandafters.com/2020/04/17/inside-flames-new-machine-learning-powered-human-face-segmentation/" }, { "Source ID": "collected-774f3f4ef585", "Operation IDs": "96", "Citation": "Flame · Face normal map controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2020/ENU/Flame-Compositingin3d/files/GUID-D7D7C9B7-FE85-497A-BBC2-455BEB112FC4.htm" }, { "Source ID": "collected-4d47e49d3bff", "Operation IDs": "96", "Citation": "help.autodesk.com, Flame 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Flame-Action/files/Action-Shading-and-Textures/GUID-1A2F6A6D-F2A1-4BE7-8B36-1D2F1773DB2F.html" } ] }, { "ID": "97", "Software": "Autodesk Flame", "Operation": "ML Sky Extraction Active Selective", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c378adc29af6\ncollected-b62fa709cc2c\ncollected-13a21ac8ddd0", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Flame ML Sky Extraction predicts sky membership for the active selective.", "Previous record: rationale": "Neural recognition and boundary estimation match region determination. The current model is not assumed identical to its documented predecessor.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Flame 2026 help lists Sky among the ML layer types of the Action Selective, described as extracting a matte for sky on the image horizon. Flame 2027 (21 April 2026) added ML Depth Indoor and Outdoor models, and Flame 2027.1 (27 July 2026) introduced the Isolate tool, which replaces GMask Tracer for creating mattes with Selective objects or ML models. Not established: Flame 2027 documentation for the Sky model was not checked directly; whether the Sky model weights changed is not documented. help.autodesk.com, Flame 2026 • cgchannel.com, 2026-07-27 (Flame 2027.1)", "Collected source links": "https://help.autodesk.com/cloudhelp/2024/ENU/Flame-WhatsNew/files/GUID-EB3FAC1F-1F55-4A5D-A977-F2408729234E.htm\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Flame-Action/files/Action-Selectives/GUID-A4A91A5E-A2F2-48CE-8405-0E4D6CF436CC.html\nhttps://www.cgchannel.com/2026/07/autodesk-releases-flame-2027/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c378adc29af6", "Operation IDs": "97", "Citation": "Flame · Neural sky recognition", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2024/ENU/Flame-WhatsNew/files/GUID-EB3FAC1F-1F55-4A5D-A977-F2408729234E.htm" }, { "Source ID": "collected-b62fa709cc2c", "Operation IDs": "97", "Citation": "Flame · Machine learning selectives", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Flame-Action/files/Action-Selectives/GUID-A4A91A5E-A2F2-48CE-8405-0E4D6CF436CC.html" }, { "Source ID": "collected-13a21ac8ddd0", "Operation IDs": "97", "Citation": "cgchannel.com, 2026-07-27 (Flame 2027.1)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/07/autodesk-releases-flame-2027/" } ] }, { "ID": "98", "Software": "Autodesk Flame", "Operation": "ML Timewarp 2026 Active Effect", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-08110a4a7e5b\ncollected-d07dfcdfe0db\ncollected-3c2986b5b4ed\ncollected-151c4f73d11f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Flame Machine Learning 2026 Timewarp predicts intermediate video frames.", "Previous record: rationale": "Temporal signal reconstruction crosses fidelity restoration and perceptual completion, supporting 5–8. Speed and frame counts do not determine t.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Autodesk's AI Transparency Card for Flame Machine Learning Timewarp lists the model source as open-source, the data source as internally generated, and the model as trained locally. Flame 2026 added a new frame interpolation model to the Timewarp tool; the 2025.1 help page states that inference uses two frames to generate an in-between frame. Not established: Which open-source model the 2026 interpolation model derives from, its architecture, and whether it is trained with a pixel reconstruction loss or a perceptual/adversarial loss. damassets.autodesk.net, not shown • help.autodesk.com, Flame 2026 • help.autodesk.com, Flame 2025.1", "Collected source links": "https://help.autodesk.com/cloudhelp/2026/ENU/Flame-EffectsToolsReference/files/GUID-09DE74B2-7668-41FD-BB58-6885800F5BAE.html\nhttps://damassets.autodesk.net/content/dam/autodesk/www/pdfs/flame-machine-learning-timewarp.pdf\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2026/What-s-New-in-2026/wn-2026-machine-learning.html\nhttps://help.autodesk.com/cloudhelp/2025/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2025/What-s-New-in-2025-1/wn-2025-1-timewarp.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-08110a4a7e5b", "Operation IDs": "98", "Citation": "Flame · ML Timewarp models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Flame-EffectsToolsReference/files/GUID-09DE74B2-7668-41FD-BB58-6885800F5BAE.html" }, { "Source ID": "collected-d07dfcdfe0db", "Operation IDs": "98", "Citation": "damassets.autodesk.net, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://damassets.autodesk.net/content/dam/autodesk/www/pdfs/flame-machine-learning-timewarp.pdf" }, { "Source ID": "collected-3c2986b5b4ed", "Operation IDs": "94, 98", "Citation": "Flame 2026 · Inference models and masks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2026/What-s-New-in-2026/wn-2026-machine-learning.html" }, { "Source ID": "collected-151c4f73d11f", "Operation IDs": "98", "Citation": "help.autodesk.com, Flame 2025.1", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2025/ENU/Flame-WhatsNew/files/What-s-New-in-Flame-Family-2025/What-s-New-in-2025-1/wn-2025-1-timewarp.html" } ] }, { "ID": "99", "Software": "FLUX through BFL API or Playground", "Operation": "FLUX.2 Pro Targeted Image Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d537936fa3bf\ncollected-0bc062d9d055\ncollected-c7d5abc9df3c\ncollected-323b4d27328d\ncollected-818b18469565\ncollected-27d6ab4dfdd4\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "FLUX.2 pro performs the selected targeted image edit through its learned image-editing route.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=2; conditional=3; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (FLUX.2 [pro], editing route). Designer controls: prompt naming the edit; input image; additional reference images; seed; safety_tolerance; output_format. Returns count not documented; a flat image: jpeg, png or webp. “Change backgrounds, swap textures, edit text—all while maintaining photorealism that matches professional photography.” Source", "Collected source links": "https://bfl.ai/blog/flux-2\nhttps://raw.githubusercontent.com/black-forest-labs/flux2/main/src/flux2/sampling.py\nhttps://docs.bfl.ai/flux_2/flux2_image_editing\nhttps://help.bfl.ai/articles/8667153955-what-is-the-bfl-playground\nhttps://docs.bfl.ai/guides/prompting_editing_single_reference\nhttps://docs.bfl.ai/api-reference/models/generate-or-edit-an-image-with-flux2-%5Bpro%5D-preview", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d537936fa3bf", "Operation IDs": "99, 100, 101", "Citation": "BFL FLUX.2 launch", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://bfl.ai/blog/flux-2" }, { "Source ID": "collected-0bc062d9d055", "Operation IDs": "99, 100, 101", "Citation": "BFL public FLUX.2 sampler", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/black-forest-labs/flux2/main/src/flux2/sampling.py" }, { "Source ID": "collected-c7d5abc9df3c", "Operation IDs": "99, 100", "Citation": "docs.bfl.ai — flux2_image_editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.bfl.ai/flux_2/flux2_image_editing" }, { "Source ID": "collected-323b4d27328d", "Operation IDs": "99", "Citation": "help.bfl.ai — 8667153955-what-is-the-bfl-playground", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.bfl.ai/articles/8667153955-what-is-the-bfl-playground" }, { "Source ID": "collected-818b18469565", "Operation IDs": "99", "Citation": "docs.bfl.ai — prompting_editing_single_reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.bfl.ai/guides/prompting_editing_single_reference" }, { "Source ID": "collected-27d6ab4dfdd4", "Operation IDs": "99", "Citation": "docs.bfl.ai — generate-or-edit-an-image-with-flux2-[pro]-preview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.bfl.ai/api-reference/models/generate-or-edit-an-image-with-flux2-%5Bpro%5D-preview" } ] }, { "ID": "100", "Software": "FLUX (Black Forest Labs API / Playground)", "Operation": "FLUX.2 pro multi-reference scene synthesis", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d537936fa3bf\ncollected-0bc062d9d055\ncollected-c7d5abc9df3c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "FLUX.2 pro synthesizes a scene using its multi-reference generation route.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=2; conditional=3; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (FLUX.2 [pro], multi-reference route). Designer controls: prompt naming references as image 1, image 2; reference images input_image to input_image_8 (up to 8 via API, 10 in playground); seed; width/height; safety_tolerance; output_format; prompt upsampling toggle. Returns count not documented; a flat image: jpeg, png or webp. “Reference multiple images simultaneously - up to 8 via API, up to 10 in the playground.” Source", "Collected source links": "https://bfl.ai/blog/flux-2\nhttps://raw.githubusercontent.com/black-forest-labs/flux2/main/src/flux2/sampling.py\nhttps://docs.bfl.ai/flux_2/flux2_image_editing", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d537936fa3bf", "Operation IDs": "99, 100, 101", "Citation": "BFL FLUX.2 launch", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://bfl.ai/blog/flux-2" }, { "Source ID": "collected-0bc062d9d055", "Operation IDs": "99, 100, 101", "Citation": "BFL public FLUX.2 sampler", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/black-forest-labs/flux2/main/src/flux2/sampling.py" }, { "Source ID": "collected-c7d5abc9df3c", "Operation IDs": "99, 100", "Citation": "docs.bfl.ai — flux2_image_editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.bfl.ai/flux_2/flux2_image_editing" } ] }, { "ID": "101", "Software": "FLUX (Black Forest Labs API / Playground)", "Operation": "FLUX.2 pro text-to-image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d537936fa3bf\ncollected-0bc062d9d055\ncollected-aeef0de4175a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "FLUX.2 pro generates a raster through its text-to-image route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (FLUX.2 [pro]). Designer controls: prompt (hex color codes and JSON structured prompts supported); width/height; seed; safety_tolerance (0-5); output_format; prompt upsampling toggle (disable_pup). Returns count not documented; a flat image: jpeg, png or webp. “Generate photorealistic images with precise control over colors, poses, and composition” Source", "Collected source links": "https://bfl.ai/blog/flux-2\nhttps://raw.githubusercontent.com/black-forest-labs/flux2/main/src/flux2/sampling.py\nhttps://docs.bfl.ai/flux_2/flux2_overview", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d537936fa3bf", "Operation IDs": "99, 100, 101", "Citation": "BFL FLUX.2 launch", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://bfl.ai/blog/flux-2" }, { "Source ID": "collected-0bc062d9d055", "Operation IDs": "99, 100, 101", "Citation": "BFL public FLUX.2 sampler", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/black-forest-labs/flux2/main/src/flux2/sampling.py" }, { "Source ID": "collected-aeef0de4175a", "Operation IDs": "101", "Citation": "docs.bfl.ai — flux2_overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.bfl.ai/flux_2/flux2_overview" } ] }, { "ID": "102", "Software": "Foundry Katana", "Operation": "Katana Node Lookdev/Lighting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-be93874bf22f\ncollected-df9f3b50f714", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Katana's scoped GafferThree edit changes one existing light-exposure attribute before rendering.", "Previous record: rationale": "The native attribute edit is conventional. The broader look-development application and a later renderer do not enter this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: GafferThree object table; Object tab; Material tab; Exposure; Intensity. Returns one result; editable structure: scene graph light attributes. “The GafferThree node type allows you to create lights under an arbitrary hierarchy of rigs.” Source", "Collected source links": "https://learn.foundry.com/katana/Content/rg/misc_nodes/gafferthree.html\nhttps://learn.foundry.com/katana/Content/ug/adding_assigning_materials/adding_assigning_materials.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-be93874bf22f", "Operation IDs": "102", "Citation": "Foundry GafferThree reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/katana/Content/rg/misc_nodes/gafferthree.html" }, { "Source ID": "collected-df9f3b50f714", "Operation IDs": "102", "Citation": "Foundry material assignment", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/katana/Content/ug/adding_assigning_materials/adding_assigning_materials.html" } ] }, { "ID": "103", "Software": "Foundry Mari", "Operation": "Mari Layer/Node Painting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-845aedca9287\ncollected-097d605cc5e8\ncollected-0a7bb8f4d1b7\ncollected-264b3f7df639\ncollected-c44e9a43561e\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Mari bakes an already supplied paint buffer into an existing texture layer using the recorded view and UV mapping.", "Previous record: rationale": "The bounded projection and bake apply prescribed image-to-surface mapping. Learned image creation and other paint tools are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: brush; paint buffer move/transform (M); bake (B); bake behavior (AutoBakeAndClear, ClearOnly, Manual); erase (E); masking; clone stamp; mirror projection; symmetry. Returns one result; editable structure: texture layer. “Then the yellow paint is baked down onto the object (press B key). The paint is added to the current layer.” Source • Screenshot", "Collected source links": "https://learn.foundry.com/mari/Content/user_guide/painting/paint_buffer.html\nhttps://www.foundry.com/products/mari\nhttps://learn.foundry.com/mari/Content/Resources/images/user_guide/paintbuffer_concept_1.jpg\nhttps://learn.foundry.com/mari/content/tutorials/tutorial_3/tutorial_painting.html\nhttps://learn.foundry.com/mari/Content/user_guide/controlling_paint/erase_paint.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-845aedca9287", "Operation IDs": "103", "Citation": "Foundry Mari paint buffer", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/mari/Content/user_guide/painting/paint_buffer.html" }, { "Source ID": "collected-097d605cc5e8", "Operation IDs": "103", "Citation": "www.foundry.com — mari", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.foundry.com/products/mari" }, { "Source ID": "collected-0a7bb8f4d1b7", "Operation IDs": "103", "Citation": "learn.foundry.com — paintbuffer_concept_1.jpg", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/mari/Content/Resources/images/user_guide/paintbuffer_concept_1.jpg" }, { "Source ID": "collected-264b3f7df639", "Operation IDs": "103", "Citation": "learn.foundry.com — tutorial_painting.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/mari/content/tutorials/tutorial_3/tutorial_painting.html" }, { "Source ID": "collected-c44e9a43561e", "Operation IDs": "103", "Citation": "learn.foundry.com — erase_paint.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/mari/Content/user_guide/controlling_paint/erase_paint.html" } ] }, { "ID": "104", "Software": "Foundry Nuke", "Operation": "Cattery MiDaS Depth Estimation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c63370a5f3eb\ncollected-e1e64b4dccb9\ncollected-056114a4ed68\ncollected-bd49a7f7e87d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Nuke Cattery MiDaS predicts relative depth, followed by a retained affine depth-map adjustment.", "Previous record: rationale": "Learned scene geometry remains consequential after remapping. The 8–10 spatial-estimation comparison is unaffected by adding arithmetic application.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). No current Foundry documentation adds to the record: the Cattery MiDaS page renders no text for fetch, and the Nuke 17.0 release notes list no depth-estimation change. The MiDaS paper describes a training objective invariant to depth range and scale across mixed datasets; the output is a dense relative depth estimate rather than a discrete recognition result, with no generative stage. Not established: The MiDaS version packaged in the current Cattery node is not confirmed. Documentation does not place dense depth estimation within 8–10. arxiv.org, arXiv 1907.01341 • learn.foundry.com, Nuke 17.0", "Collected source links": "https://community.foundry.com/cattery/37770/midas\nhttps://github.com/isl-org/MiDaS\nhttps://arxiv.org/abs/1907.01341\nhttps://learn.foundry.com/nuke/17.1v1/content/release_notes/nuke_17.0.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c63370a5f3eb", "Operation IDs": "104", "Citation": "Foundry Cattery MiDaS model entry, Downloads and Instructions inspected", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.foundry.com/cattery/37770/midas" }, { "Source ID": "collected-e1e64b4dccb9", "Operation IDs": "104", "Citation": "MiDaS original repository, README and model comparison", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/isl-org/MiDaS" }, { "Source ID": "collected-056114a4ed68", "Operation IDs": "104", "Citation": "arxiv.org, arXiv 1907.01341", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://arxiv.org/abs/1907.01341" }, { "Source ID": "collected-bd49a7f7e87d", "Operation IDs": "104, 106", "Citation": "learn.foundry.com, Nuke 17.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/17.1v1/content/release_notes/nuke_17.0.html" } ] }, { "ID": "105", "Software": "Foundry Nuke", "Operation": "Nuke Node Compositing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-867a5eae0bd7\ncollected-4f20cda362dc\ncollected-4fd88a4c7feb\ncollected-dcecd7844c3c\ncollected-3de94ed09a78\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Nuke evaluates a defined graph of conventional transforms, grades and merges.", "Previous record: rationale": "The selected graph contains no trained node. Supplied images and their generation history do not change these compositing operations.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: nodes/parameters; operation (over, plus, multiply, screen and others); mix; mask input; set bbox to; A/B channels; output channels; Transform and Grade node knobs. Returns one result; editable structure: nodes. “Merge allows you to layer multiple images together.” Source • Screenshot", "Collected source links": "https://learn.foundry.com/nuke/content/reference_guide/merge_nodes/merge.html\nhttps://learn.foundry.com/nuke/content/reference_guide/transform_nodes/transform.html\nhttps://learn.foundry.com/nuke/content/resources/images/ug_images/node_properties.png\nhttps://learn.foundry.com/nuke/content/getting_started/managing_scripts/loading_scripts.html\nhttps://learn.foundry.com/nuke/content/getting_started/using_interface/viewers.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-867a5eae0bd7", "Operation IDs": "105", "Citation": "Foundry Merge node reference and training transcript", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/reference_guide/merge_nodes/merge.html" }, { "Source ID": "collected-4f20cda362dc", "Operation IDs": "105", "Citation": "Foundry Transform node reference and training transcript", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/reference_guide/transform_nodes/transform.html" }, { "Source ID": "collected-4fd88a4c7feb", "Operation IDs": "105", "Citation": "learn.foundry.com — node_properties.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/resources/images/ug_images/node_properties.png" }, { "Source ID": "collected-dcecd7844c3c", "Operation IDs": "105", "Citation": "learn.foundry.com — loading_scripts.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/getting_started/managing_scripts/loading_scripts.html" }, { "Source ID": "collected-3de94ed09a78", "Operation IDs": "105, 109", "Citation": "learn.foundry.com — viewers.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/getting_started/using_interface/viewers.html" } ] }, { "ID": "106", "Software": "Foundry Nuke (NukeX)", "Operation": "CopyCat Matte Extraction and Correction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a4e16e586fb1\ncollected-e74215415af6\ncollected-bd49a7f7e87d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "CopyCat learns matte extraction from paired examples; its prediction is followed by Grade and RotoPaint cleanup with learned regions retained.", "Previous record: rationale": "The endpoint still contains learned region determination, supporting a 10 match. Training-example count and the amount painted supply no numerical dilution rule.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Nuke 17.0 (February 2026) adds BigCat for training across larger datasets, with data augmentation, validation input and selectable L1, L2 and perceptual (LPIPS) losses; CopyCat remains the per-shot trainer. The matte-extraction task trained from paired examples is unchanged. Not established: Whether the perceptual-loss option applies to CopyCat itself is not established from the release notes. learn.foundry.com, Nuke 17.0", "Collected source links": "https://learn.foundry.com/nuke/content/reference_guide/air_nodes/copycat.html\nhttps://learn.foundry.com/nuke/content/comp_environment/air_tools/cc-data.html\nhttps://learn.foundry.com/nuke/17.1v1/content/release_notes/nuke_17.0.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a4e16e586fb1", "Operation IDs": "106", "Citation": "Foundry CopyCat node reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/reference_guide/air_nodes/copycat.html" }, { "Source ID": "collected-e74215415af6", "Operation IDs": "106", "Citation": "Foundry Create a Data Set for Training", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/comp_environment/air_tools/cc-data.html" }, { "Source ID": "collected-bd49a7f7e87d", "Operation IDs": "104, 106", "Citation": "learn.foundry.com, Nuke 17.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/17.1v1/content/release_notes/nuke_17.0.html" } ] }, { "ID": "107", "Software": "Foundry Nuke (NukeX)", "Operation": "Nuke AIR Deblur", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-28d9858e2c3f\ncollected-d18a58eb1590\ncollected-725944f4f4e2\ncollected-2dfaadcd5fff", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Nuke AIR Deblur uses a pretrained network to reconstruct motion-blurred image detail.", "Previous record: rationale": "The trained restoration task supports a fidelity-restoration match. Mask and mix controls do not directly supply t; exact cancellation would require separate proof.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). The Nuke 17.1v2 reference guide describes Deblur as removing motion blur with a pre-trained machine learning network, with mix and mask controls. The 2021 AIR launch coverage says Foundry trained the node with the AIR system but gives no data or objective; Nuke 17.0 (February 2026) adds BigCat and an LPIPS perceptual loss for user training in CopyCat and lists no change to Deblur. Not established: Training data, loss, and whether any perceptual or adversarial term shaped the shipped Deblur weights are not published. learn.foundry.com, Nuke 17.1v2 • cgchannel.com, 2021-03 • digitalproduction.com, 2026-02-26", "Collected source links": "https://learn.foundry.com/nuke/content/reference_guide/air_nodes/deblur.html\nhttps://campaigns.foundry.com/products/nuke-family/releases/13-0\nhttps://www.cgchannel.com/2021/03/nukes-air-toolset-lets-you-train-your-own-neural-networks/\nhttps://digitalproduction.com/2026/02/26/nuke-17-0-rewires-3d-and-adds-gaussian-splats/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-28d9858e2c3f", "Operation IDs": "107", "Citation": "Foundry Deblur node reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/reference_guide/air_nodes/deblur.html" }, { "Source ID": "collected-d18a58eb1590", "Operation IDs": "107, 108", "Citation": "Foundry Nuke 13.0 release overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://campaigns.foundry.com/products/nuke-family/releases/13-0" }, { "Source ID": "collected-725944f4f4e2", "Operation IDs": "107", "Citation": "cgchannel.com, 2021-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2021/03/nukes-air-toolset-lets-you-train-your-own-neural-networks/" }, { "Source ID": "collected-2dfaadcd5fff", "Operation IDs": "107", "Citation": "digitalproduction.com, 2026-02-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://digitalproduction.com/2026/02/26/nuke-17-0-rewires-3d-and-adds-gaussian-splats/" } ] }, { "ID": "108", "Software": "Foundry Nuke (NukeX)", "Operation": "Nuke AIR Upscale", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-174a52fc8931\ncollected-d18a58eb1590\ncollected-cf97f2da5c72", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Nuke AIR Upscale predicts enlarged image detail through its pretrained network.", "Previous record: rationale": "The published task supports 5–8 between perceptual reconstruction and fidelity restoration; undisclosed objective details do not justify a precise midpoint.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Nuke 17.1v2 documentation keeps the Upscale description as a pre-trained network that doubles the input format and publishes no training details. Foundry's Nuke 13.0 release page states the Upscale and Deblur models were optimized using the ML methodology behind CopyCat, and the Nuke 17.1v2 BigCat page documents L2 (mean squared error) as CopyCat's default loss. Not established: Which loss trained the shipped Upscale weights. The link from 'CopyCat methodology' to CopyCat's default L2 loss is an inference; Foundry names no objective for the Upscale model itself. learn.foundry.com, Nuke 17.1v2 • campaigns.foundry.com, Nuke 13.0 • learn.foundry.com, Nuke 17.1v2", "Collected source links": "https://learn.foundry.com/nuke/content/reference_guide/air_nodes/upscale.html\nhttps://campaigns.foundry.com/products/nuke-family/releases/13-0\nhttps://learn.foundry.com/nuke/content/comp_environment/air_tools/using-bigcat.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-174a52fc8931", "Operation IDs": "108", "Citation": "Foundry Upscale node reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/reference_guide/air_nodes/upscale.html" }, { "Source ID": "collected-d18a58eb1590", "Operation IDs": "107, 108", "Citation": "Foundry Nuke 13.0 release overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://campaigns.foundry.com/products/nuke-family/releases/13-0" }, { "Source ID": "collected-cf97f2da5c72", "Operation IDs": "108", "Citation": "learn.foundry.com, Nuke 17.1v2", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/comp_environment/air_tools/using-bigcat.html" } ] }, { "ID": "109", "Software": "Foundry Nuke (NukeX)", "Operation": "Nuke SmartVector Paint Propagation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5d4bf6c4f598\ncollected-125c113c6377\ncollected-295bb39910a4\ncollected-6ae6b802097e\ncollected-fb4bb03c87b7\ncollected-3de94ed09a78\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SmartVector estimates image motion and VectorDistort propagates reference paint through that motion field.", "Previous record: rationale": "The recorded algorithm account does not conclusively identify whether the active estimator contains a trained mechanism. Retain the dependency description while leaving the equation branch unresolved.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Foundry's Nuke 17.1v2 SmartVector page lists Vector Detail (default 0.3), Strength (default 1.5), Flicker Compensation, luminance channel weights and nearest-vector infill of matte regions, without naming its estimator. The Nuke 17.1v2 VectorGenerator page documents a Local block-matching method and a Regularized semi-global motion estimation method whose Strength control also defaults to 1.5; neither page mentions a trained model. Not established: Foundry documentation does not state that SmartVector uses the VectorGenerator Regularized estimator or that its estimator holds no trained component; the link rests on shared controls and defaults. learn.foundry.com, Nuke 17.1v2 • learn.foundry.com, Nuke 17.1v2 • learn.foundry.com, Nuke 17.1v2", "Collected source links": "https://learn.foundry.com/nuke/content/reference_guide/time_nodes/smartvector.html\nhttps://learn.foundry.com/nuke/content/comp_environment/smartpaint/generating_vectors.html\nhttps://learn.foundry.com/nuke/content/reference_guide/time_nodes/vectorgenerator.html\nhttps://learn.foundry.com/nuke/content/comp_environment/smartpaint/generating_distortion.html\nhttps://learn.foundry.com/nuke/content/comp_environment/rotopaint/editing_existing_attributes.html\nhttps://learn.foundry.com/nuke/content/getting_started/using_interface/viewers.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5d4bf6c4f598", "Operation IDs": "109", "Citation": "Foundry SmartVector node reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/reference_guide/time_nodes/smartvector.html" }, { "Source ID": "collected-125c113c6377", "Operation IDs": "109", "Citation": "Foundry Generating Motion Vectors", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/comp_environment/smartpaint/generating_vectors.html" }, { "Source ID": "collected-295bb39910a4", "Operation IDs": "109", "Citation": "learn.foundry.com, Nuke 17.1v2", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/reference_guide/time_nodes/vectorgenerator.html" }, { "Source ID": "collected-6ae6b802097e", "Operation IDs": "109", "Citation": "learn.foundry.com, Nuke 17.1v2", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/comp_environment/smartpaint/generating_distortion.html" }, { "Source ID": "collected-fb4bb03c87b7", "Operation IDs": "109", "Citation": "learn.foundry.com — editing_existing_attributes.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/comp_environment/rotopaint/editing_existing_attributes.html" }, { "Source ID": "collected-3de94ed09a78", "Operation IDs": "105, 109", "Citation": "learn.foundry.com — viewers.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.foundry.com/nuke/content/getting_started/using_interface/viewers.html" } ] }, { "ID": "110", "Software": "Freepik AI Suite / Magnific", "Operation": "Mystic 2.5 text-to-image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2cf56f6a40fc\ncollected-c9cc9d821f80\ncollected-15752d35cb6e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Mystic 2.5 generates a raster through the specifically selected Freepik or Magnific image-generation route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Mystic; model options zen, flexible, fluid, realism, super_real, editorial_portraits). Designer controls: prompt (@character syntax); structure reference; structure_strength; style reference; adherence; hdr; creative_detailing; resolution (1k/2k/4k); aspect ratio; model; engine; fixed_generation; styling (styles, characters, color palette). Returns count not documented; a flat image: PNG. “Convert descriptive text input into images using AI. This endpoint accepts a variety of parameters to customize the generated images.” Source", "Collected source links": "https://www.magnific.com/ai/docs/image-ai-models\nhttps://www.freepik.com/ai/image-generator\nhttps://docs.magnific.com/api-reference/mystic/post-mystic", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2cf56f6a40fc", "Operation IDs": "110", "Citation": "Magnific image model catalogue", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.magnific.com/ai/docs/image-ai-models" }, { "Source ID": "collected-c9cc9d821f80", "Operation IDs": "110", "Citation": "Magnific image generator", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.freepik.com/ai/image-generator" }, { "Source ID": "collected-15752d35cb6e", "Operation IDs": "110", "Citation": "docs.magnific.com — post-mystic", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.magnific.com/api-reference/mystic/post-mystic" } ] }, { "ID": "111", "Software": "Autodesk Fusion", "Operation": "Automated Modeling with parametric solid output", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ae184598e633\ncollected-a1a41969f573\ncollected-fcad97d62e7e\ncollected-93e6ca3327a4\ncollected-ce0ff9d1b94f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Fusion Automated Modeling constructs connecting shapes using its documented geometric synthesis technology.", "Previous record: rationale": "The documented geometry engine is a prescribed level-set shape-synthesis optimization that determines the connecting shape from selected faces and avoid bodies, and no trained predictor is documented in that path; under the comparison policy no learned determination is retained, so X = 15.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Changed 3 Oct 2026 X Unresolved → 15.0. Autodesk's Fusion blog states Automated Modeling combines basic inputs with patented generative design technologies, and Engineering.com (5 Dec 2022) reports it uses the same underlying technology as Autodesk's generative design tool, with no load or manufacturing criteria. An Autodesk University article describes that technology as a level-set approach to topology synthesis, and Autodesk patent US12488539B2 (priority 26 Jun 2023, published 2 Dec 2025) describes level-set shape synthesis with connectivity filtering of preserve bodies and contains no trained model. Not established: No Autodesk page names the algorithm of the 2026 Automated Modeling build directly; the patent does not name the feature; the January 2026 product update page refused retrieval. Previous reason: AI and generative-design labels do not establish a trained predictor. The current learned contribution remains unresolved, preventing an honest numerical branch choice. autodesk.com • engineering.com, 2022-12-05 • autodesk.com", "Collected source links": "https://help.autodesk.com/cloudhelp/ENU/Fusion-Model/files/SLD-AUTOMATED-MODELING.htm\nhttps://help.autodesk.com/cloudhelp/ENU/Fusion-Model/files/SLD-GENERATE-DESIGN-ALTERNATIVES.htm\nhttps://www.autodesk.com/products/fusion-360/blog/getting-started-automated-modeling-basics/\nhttps://www.engineering.com/generative-design-for-everyone-meet-automated-modeling-from-autodesk/\nhttps://www.autodesk.com/autodesk-university/article/Fusion-360-Introduction-Generative-Design", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ae184598e633", "Operation IDs": "111", "Citation": "help.autodesk.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/Fusion-Model/files/SLD-AUTOMATED-MODELING.htm" }, { "Source ID": "collected-a1a41969f573", "Operation IDs": "111", "Citation": "help.autodesk.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/Fusion-Model/files/SLD-GENERATE-DESIGN-ALTERNATIVES.htm" }, { "Source ID": "collected-fcad97d62e7e", "Operation IDs": "111", "Citation": "www.autodesk.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/products/fusion-360/blog/getting-started-automated-modeling-basics/" }, { "Source ID": "collected-93e6ca3327a4", "Operation IDs": "111", "Citation": "engineering.com, 2022-12-05", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.engineering.com/generative-design-for-everyone-meet-automated-modeling-from-autodesk/" }, { "Source ID": "collected-ce0ff9d1b94f", "Operation IDs": "111, 113", "Citation": "www.autodesk.com — Fusion-360-Introduction-Generative-Design", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/autodesk-university/article/Fusion-360-Introduction-Generative-Design" } ] }, { "ID": "112", "Software": "Autodesk Fusion", "Operation": "Drawing Automation with AI Planning and Native Views", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9fa370a20dd4\ncollected-77ed69794034\ncollected-11b0e14354b0\ncollected-96cc9b9e3454", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Fusion Drawing Automation uses trained AI planning to choose drawing views, arrangement and dimensions before native drawing construction.", "Previous record: rationale": "The documented learned determinations are membership and arrangement decisions over existing model geometry (which components are omitted, which dimension arrangement is offered), the kind REF10 makes, while sheets and views come from templates and heuristics; the learned share of the drawing is smaller than in REF10, giving an interval from REF10 toward REF12, where learned computation is retained inside a prescribed procedure.", "Previous record: calculation": "X = 10 + t × (12 − 10); t = 0–1; X = 10–12. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 10.0–12.0. Engineering.com (27 January 2025) states the core of Automated Drawings is based on user templates and heuristics, with an added AI model that classifies standard fasteners and excludes them, and an Auto Dimension tool that offers a list of dimension arrangements in the manner of Sketch AutoConstrain. Autodesk Help lists a 'Detect and Omit Fasteners' preference and user-chosen dimension strategies (baseline, chain, ordinate); no source located describes a trained model choosing views, so the record's description of AI planning of views is not supported. Not established: Whether Auto Dimension in drawings uses a trained model (AutoConstrain is a transformer trained on constrained sketches, January 2025), and whether 2026 updates (Autodesk Assistant drawing support, May 2026) added learned view selection. Previous reason: Learned structural planning remains in conventional drawing objects. The synthesis-to-relational 2–10 envelope preserves the unresolved division of responsibility among internal steps. engineering.com, 2025-01-27 • help.autodesk.com, undated • adsknews.autodesk.com, 2024-10-15", "Collected source links": "https://help.autodesk.com/cloudhelp/ENU/Fusion-Drawing/files/DWG-REF-AUTO-PREFERENCES.htm\nhttps://www.autodesk.com/products/fusion-360/blog/drawing-automation-autodesk-fusion/\nhttps://www.engineering.com/ai-takes-on-sketching-and-drawings-in-autodesk-fusion/\nhttps://adsknews.autodesk.com/en/news/autodesk-ai-strengthens-fusion-and-alias/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9fa370a20dd4", "Operation IDs": "112", "Citation": "Documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/Fusion-Drawing/files/DWG-REF-AUTO-PREFERENCES.htm" }, { "Source ID": "collected-77ed69794034", "Operation IDs": "112", "Citation": "Autodesk · Drawing Automation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/products/fusion-360/blog/drawing-automation-autodesk-fusion/" }, { "Source ID": "collected-11b0e14354b0", "Operation IDs": "112", "Citation": "engineering.com, 2025-01-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.engineering.com/ai-takes-on-sketching-and-drawings-in-autodesk-fusion/" }, { "Source ID": "collected-96cc9b9e3454", "Operation IDs": "112", "Citation": "adsknews.autodesk.com, 2024-10-15", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://adsknews.autodesk.com/en/news/autodesk-ai-strengthens-fusion-and-alias/" } ] }, { "ID": "113", "Software": "Autodesk Fusion", "Operation": "Generative Design Outcome to Editable Solid", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c7f45733a29e\ncollected-64bd98a13fbb\ncollected-ce0ff9d1b94f\ncollected-66415846c404\ncollected-c14fb53d19f3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Fusion Generative Design produces optimized design outcomes and converts a selected result to an editable solid.", "Previous record: rationale": "Outcome geometry follows from a prescribed physics-driven level-set solver and the editable solid from geometric surface fitting; the documented learned components sort and recommend finished outcomes and carry no computational responsibility for the geometry, so under the comparison policy X = 15.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Changed 3 Oct 2026 X Unresolved → 15.0. An Autodesk University article states that the core mechanism of Fusion generative design is a level-set approach to topology synthesis and that its synthesis methods are algorithmic. Autodesk patent US11886165B2 (priority 9 Nov 2018, granted 30 Jan 2024) describes conversion of generative design output to editable B-rep through mesh partitioning, T-spline fitting and gradient-descent boundary adjustment, with no trained model. Autodesk's Fusion blog (27 Nov 2024, mirrored) assigns machine learning to clustering and recommending outcomes, which organizes results without computing outcome geometry. Not established: The Autodesk University article shows no date, and no 2026 Autodesk document restates the solver method. Previous reason: The active solver and any learned geometry contribution are not established by the broad workflow record. Neither the word generative nor later editability supplies X. autodesk.com • patents.google.com, 2024-01-30 • dailycadcam.com, 2024-11-27", "Collected source links": "https://help.autodesk.com/cloudhelp/ENU/Fusion-GenerativeDesign/files/GD-OVERVIEW.htm\nhttps://help.autodesk.com/view/fusion360/ENU/?contextId=DESIGN_HISTORY\nhttps://www.autodesk.com/autodesk-university/article/Fusion-360-Introduction-Generative-Design\nhttps://patents.google.com/patent/US11886165B2/en\nhttps://dailycadcam.com/generative-design-in-autodesk-fusion-revolutionizing-design-with-ai/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c7f45733a29e", "Operation IDs": "113", "Citation": "Autodesk · Generative Design", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/Fusion-GenerativeDesign/files/GD-OVERVIEW.htm" }, { "Source ID": "collected-64bd98a13fbb", "Operation IDs": "113", "Citation": "Autodesk · Parametric and direct modeling modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/view/fusion360/ENU/?contextId=DESIGN_HISTORY" }, { "Source ID": "collected-ce0ff9d1b94f", "Operation IDs": "111, 113", "Citation": "www.autodesk.com — Fusion-360-Introduction-Generative-Design", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/autodesk-university/article/Fusion-360-Introduction-Generative-Design" }, { "Source ID": "collected-66415846c404", "Operation IDs": "113", "Citation": "patents.google.com, 2024-01-30", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://patents.google.com/patent/US11886165B2/en" }, { "Source ID": "collected-c14fb53d19f3", "Operation IDs": "113", "Citation": "dailycadcam.com, 2024-11-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dailycadcam.com/generative-design-in-autodesk-fusion-revolutionizing-design-with-ai/" } ] }, { "ID": "114", "Software": "Autodesk Fusion", "Operation": "Parametric Feature Modeling", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "Specified sketch profile, direction, extent and operation; return parametric solid feature. Exclude inferred AutoConstrain relations upstream.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Specified sketch profile, direction, extent and solid operation; no inferred AutoConstrain relations included upstream.", "Learned determination": "None in the scoped extrusion.", "Conventional processing": "The geometric kernel constructs a parametric solid and updates dependencies.", "Model contribution in the result": "No learned prediction contributes to this standalone operation.", "Why this scope score (S)": "S0/U0/J0: parameter dependence and geometric complexity do not establish a fitted predictive relationship.", "Why this latitude score (U)": "S0/U0/J0: parameter dependence and geometric complexity do not establish a fitted predictive relationship.", "Why this coordination score (J)": "S0/U0/J0: parameter dependence and geometric complexity do not establish a fitted predictive relationship.", "Comparisons and other modes": "X15. If AutoConstrain predictions are included within the declared endpoint, their contribution remains; this standalone conventional comparator cannot erase it.", "Citations": "Fusion Extrude reference — https://help.autodesk.com/cloudhelp/ENU/Fusion-Model/files/SLD-REF-EXTRUDE.htm", "Source IDs": "CAL2-FUSION", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human provides geometry and parameterized rule. Conventional geometric computation constructs the solid and updates dependents.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X15. If AutoConstrain predictions are included within the declared endpoint, their contribution remains; this standalone conventional comparator cannot erase it.", "Previous record: scope": "Fusion evaluates a specified extrusion and its parametric dependencies in the native feature timeline.", "Previous record: rationale": "Geometric construction and recomputation implement prescribed functions without a trained predictor in this scope.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: Profiles; Start; Direction; Extent Type; Distance; Taper Angle; Operation (Join, Cut, Intersect, New Body, New Component). Returns one result; editable structure: solid body plus timeline feature. “Extrudes the entire area within selected closed profiles” Source • Screenshot", "Collected source links": "https://www.autodesk.com/products/fusion-360/overview\nhttps://help.autodesk.com/view/fusion360/ENU/?contextId=SLD-EXTRUDE\nhttps://web.archive.org/web/20251204225730id_/https://help.autodesk.com/cloudhelp/ENU/Fusion-Assemble/files/ASM-DESIGN-MODELING-MODES.htm\nhttps://help.autodesk.com/cloudhelp/ENU/Fusion-Model/files/SLD-MODIFY-PARAMETERS.htm", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-114", "Sources": [ { "Source ID": "CAL2-FUSION", "Operation IDs": "114", "Citation": "Fusion Extrude reference", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Fusion Extrude reference", "Finding used in this review": "Specified profile and extent determine conventional feature creation.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/Fusion-Model/files/SLD-REF-EXTRUDE.htm" } ] }, { "ID": "115", "Software": "Autodesk Fusion", "Operation": "Sketch AutoConstrain", "Proposed X": "8.1", "Review status": "Proposed X calculated", "Previous X": "8.0–10.0", "S": "2", "U": "3", "J": "3", "Reviewed on": "2026-10-14", "Configuration": "Supplied sketch geometry and existing relations; Modify Geometry off; learned constraints/dimensions selected and applied.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Supplied sketch entities and existing relations; Modify Geometry off; candidate selection and acceptance.", "Learned determination": "The predictor proposes missing constraints and dimensions expressing design intent.", "Conventional processing": "The conventional solver applies the accepted predicted relations.", "Model contribution in the result": "Predicted relations remain in the sketch after acceptance; conventional execution retains them.", "Why this scope score (S)": "S2 structured constraints/dimensions; U3 unresolved design intent; J3 relationships among supplied primitives. Selection and application retain model-predicted relations.", "Why this latitude score (U)": "S2 structured constraints/dimensions; U3 unresolved design intent; J3 relationships among supplied primitives. Selection and application retain model-predicted relations.", "Why this coordination score (J)": "S2 structured constraints/dimensions; U3 unresolved design intent; J3 relationships among supplied primitives. Selection and application retain model-predicted relations.", "Comparisons and other modes": "X8.1 remains unchanged. Deleting some proposals does not erase retained ones. Removing/replacing every learned relation independently and retaining no other learned effect reaches X15. Modify Geometry on requires a separate dependency account.", "Citations": "Fusion AutoConstrain instructions — https://help.autodesk.com/cloudhelp/ENU/Fusion-Sketch/files/SKT-AUTO-CONSTRAIN.htm\nAutodesk AutoConstrain technical account — https://www.research.autodesk.com/blog/using-ai-to-power-autoconstrain-in-fusion-automated-sketching/\nFusion AutoConstrain access discussion — https://forums.autodesk.com/t5/fusion-design-validate-document/where-is-the-automate-tab-when-making-a-sketch/m-p/14267848", "Source IDs": "CAL2-AUTOCON\nCAL2-AUTOAI\nCAL2-REPORTFUSION", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human specifies initial geometry and relations and reviews a candidate. Learned prediction proposes missing design-intent relations. Conventional solver applies the accepted relations.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X8.1 remains unchanged. Deleting some proposals does not erase retained ones. Removing/replacing every learned relation independently and retaining no other learned effect reaches X15. Modify Geometry on requires a separate dependency account.", "Previous record: scope": "Fusion AutoConstrain predicts constraints and dimensions from sketch geometry through the documented Transformer lineage.", "Previous record: rationale": "Learned relational completion crosses region determination and fidelity estimation, supporting 8–10. The conventional constraint solver retains the predicted relationships.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). Fusion help describes AutoConstrain as generating candidate sets of constraints and dimensions from which the user selects one before the sketch updates. Autodesk Research's 14 October 2025 post (ICCV 2025) describes aligning the constraint-generation model with supervised fine-tuning, DPO, expert iteration and reinforcement learning toward fully constrained sketches; the learned output is a set of relations and dimensions on existing geometry, and no new geometry is generated. Not established: Whether the aligned ICCV 2025 model is the one deployed in Fusion is not stated. Documentation does not place relational prediction within 8–10. help.autodesk.com, accessed 2026-10-03 • research.autodesk.com, 2025-10-14", "Collected source links": "https://www.research.autodesk.com/blog/using-ai-to-power-autoconstrain-in-fusion-automated-sketching/\nhttps://help.autodesk.com/cloudhelp/ENU/Fusion-Sketch/files/SKT-AUTO-CONSTRAIN.htm\nhttps://www.research.autodesk.com/blog/ai-alignment-in-cad-design-teaching-machines-to-understand-design-intent-in-autoconstrain/\nhttps://help.autodesk.com/cloudhelp/ENU/Fusion-360-API/files/fusion_Sketch_autoConstrain.htm\nhttps://www.reddit.com/r/Fusion360/comments/1ryioog/why_cant_they_just_list_constraints_in_a_list/\nhttps://help.autodesk.com/cloudhelp/ENU/Fusion-360-API/files/fusion_AutoConstrainResult.htm", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "121/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-115", "Sources": [ { "Source ID": "CAL2-AUTOCON", "Operation IDs": "115", "Citation": "Fusion AutoConstrain instructions", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Fusion AutoConstrain instructions", "Finding used in this review": "Generate/review/customize/apply constraints and dimensions; optional geometry modification is a separate mode.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/ENU/Fusion-Sketch/files/SKT-AUTO-CONSTRAIN.htm" }, { "Source ID": "CAL2-AUTOAI", "Operation IDs": "115", "Citation": "Autodesk AutoConstrain technical account", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Autodesk AutoConstrain technical account", "Finding used in this review": "Product-linked learned relationship prediction over supplied sketch primitives.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://www.research.autodesk.com/blog/using-ai-to-power-autoconstrain-in-fusion-automated-sketching/" }, { "Source ID": "CAL2-REPORTFUSION", "Operation IDs": "115", "Citation": "Fusion AutoConstrain access discussion", "Type": "user_report_or_forum", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Fusion AutoConstrain access discussion", "Finding used in this review": "September 2026 access/licensing issue; availability does not change the assessed retained function.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://forums.autodesk.com/t5/fusion-design-validate-document/where-is-the-automate-tab-when-making-a-sketch/m-p/14267848" } ] }, { "ID": "116", "Software": "Gamma", "Operation": "AI visual presentation generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4a5166cc921e\ncollected-63cb9bc5933b\ncollected-069a29118a22", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Gamma generates an illustrated native presentation through managed planning and selected AI image generation.", "Previous record: rationale": "Learned composition and image synthesis remain in the presentation. The mixed responsibilities support 2–10 without a stage-count average or invented controller specification.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). Gamma Help (undated, current October 2026) lists AI images as the default image source when generating a presentation, with automatic model selection among 40+ image models (Flux, GPT Image, Gemini, Ideogram, Recraft, DALL-E 3 and others in the API reference). Synthesized images and planned card layout are both retained in the default result, which keeps the 2–10 envelope. Not established: Which model writes text and chooses card layout. help.gamma.app, undated • developers.gamma.app, undated", "Collected source links": "https://help.gamma.app/en/articles/7838093-how-do-i-create-a-new-presentation-document-or-webpage-in-gamma\nhttps://help.gamma.app/en/articles/11856101-how-do-i-use-the-visuals-menu-in-gamma\nhttps://developers.gamma.app/reference/image-model-accepted-values", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4a5166cc921e", "Operation IDs": "116", "Citation": "Gamma creation modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.gamma.app/en/articles/7838093-how-do-i-create-a-new-presentation-document-or-webpage-in-gamma" }, { "Source ID": "collected-63cb9bc5933b", "Operation IDs": "116", "Citation": "Gamma visuals menu", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.gamma.app/en/articles/11856101-how-do-i-use-the-visuals-menu-in-gamma" }, { "Source ID": "collected-069a29118a22", "Operation IDs": "116", "Citation": "developers.gamma.app, undated", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developers.gamma.app/reference/image-model-accepted-values" } ] }, { "ID": "117", "Software": "Glyphs 4", "Operation": "Master interpolation of AI-informed glyph drawings", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a5d4cd5c8d0d\ncollected-1fd74d9854e9\ncollected-79bf11e5dcfb\ncollected-35433173f86d\ncollected-b689da52ea67\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Glyphs interpolates compatible vector masters at a specified instance coordinate.", "Previous record: rationale": "Linear outline interpolation is conventional. Earlier AI-informed drawing does not change the interpolation mechanism.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: axes; master axis coordinates; instance axis coordinates (Font Info > Exports); Add Instance; interpolation type (variable or default); preview. Returns one result; editable structure: vector paths, exported as font instances. “A small number of masters can result in a large number of interpolated instances.” Source", "Collected source links": "https://handbook.glyphsapp.com/interpolation/outline-compatibility/\nhttps://handbook.glyphsapp.com/higher-order-interpolation/\nhttps://handbook.glyphsapp.com/interpolation/\nhttps://handbook.glyphsapp.com/working-with-multiple-fonts/\nhttps://handbook.glyphsapp.com/custom-parameter-descriptions/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a5d4cd5c8d0d", "Operation IDs": "117", "Citation": "Glyphs outline compatibility", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://handbook.glyphsapp.com/interpolation/outline-compatibility/" }, { "Source ID": "collected-1fd74d9854e9", "Operation IDs": "117", "Citation": "Glyphs higher-order interpolation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://handbook.glyphsapp.com/higher-order-interpolation/" }, { "Source ID": "collected-79bf11e5dcfb", "Operation IDs": "117", "Citation": "handbook.glyphsapp.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://handbook.glyphsapp.com/interpolation/" }, { "Source ID": "collected-35433173f86d", "Operation IDs": "117", "Citation": "handbook.glyphsapp.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://handbook.glyphsapp.com/working-with-multiple-fonts/" }, { "Source ID": "collected-b689da52ea67", "Operation IDs": "117", "Citation": "handbook.glyphsapp.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://handbook.glyphsapp.com/custom-parameter-descriptions/" } ] }, { "ID": "118", "Software": "Godot Engine", "Operation": "MCP LLM 3D Scene Node Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-00bdca061f65\ncollected-6a6fbfe5fd9b\ncollected-70ec396b0a4a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "A Godot MCP LLM selects a scene-node insertion and invokes native construction in the existing scene.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The model determines which native node type is inserted, where in the tree and with which property values, with node behaviour prescribed by Godot; this is the which-and-where kind of determination REF10 makes and places the operation nearer REF10 than REF02. The interval stays open below REF10 because the LLM authors the placement from a prompt.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. Coding-Solo godot-mcp (npm 0.1.1; listing shows last update 16 April 2026) runs scene operations through one bundled GDScript file that accepts an operation type and JSON parameters; add_node adds a chosen node type with properties to an existing scene. No tool calls a generative model or executes LLM-written GDScript. Not established: The LLM in use is chosen by the client. Previous reason: The chosen node and placement retain learned structural decisions, supporting a 2–10 envelope. The isolated native add_node command has a separate conventional scope; this entry includes LLM planning. raw.githubusercontent.com, main branch", "Collected source links": "https://github.com/Coding-Solo/godot-mcp/blob/1209744fad78f3998f98c7394fd0f6ef50da5281/README.md\nhttps://github.com/Coding-Solo/godot-mcp/blob/1209744fad78f3998f98c7394fd0f6ef50da5281/src/index.ts#L484-L544\nhttps://raw.githubusercontent.com/Coding-Solo/godot-mcp/main/README.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-00bdca061f65", "Operation IDs": "118", "Citation": "Coding-Solo Godot MCP README", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Coding-Solo/godot-mcp/blob/1209744fad78f3998f98c7394fd0f6ef50da5281/README.md" }, { "Source ID": "collected-6a6fbfe5fd9b", "Operation IDs": "118", "Citation": "Godot MCP command dispatch", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Coding-Solo/godot-mcp/blob/1209744fad78f3998f98c7394fd0f6ef50da5281/src/index.ts#L484-L544" }, { "Source ID": "collected-70ec396b0a4a", "Operation IDs": "118", "Citation": "raw.githubusercontent.com, main branch", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/Coding-Solo/godot-mcp/main/README.md" } ] }, { "ID": "119", "Software": "Google Docs", "Operation": "Gemini image and diagram generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-350b7a144a95\ncollected-037c822af719\ncollected-7fc25530c241", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Google Docs invokes Gemini image or diagram-image generation and inserts the resulting image.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Gemini; version not named on help page). Designer controls: prompt; style (Add a style, for cover images); natural-language refinement including aspect ratio and style; Edit image prompt after insert. Returns several to choose from (number not stated); a flat image: inserted image; diagram editability not stated. “To get several suggested images, click Create.” Source • Screenshot", "Collected source links": "https://workspaceupdates.googleblog.com/2026/07/generate-and-edit-visuals-with-gemini-in-Google-Docs.html\nhttps://workspace.google.com/blog/product-announcements/july-2026-workspace-feature-drop\nhttps://support.google.com/docs/answer/15559738?hl=en", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-350b7a144a95", "Operation IDs": "119", "Citation": "Google Docs contextual visual generation release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://workspaceupdates.googleblog.com/2026/07/generate-and-edit-visuals-with-gemini-in-Google-Docs.html" }, { "Source ID": "collected-037c822af719", "Operation IDs": "119", "Citation": "July Google Workspace feature release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://workspace.google.com/blog/product-announcements/july-2026-workspace-feature-drop" }, { "Source ID": "collected-7fc25530c241", "Operation IDs": "119", "Citation": "support.google.com — 15559738?hl=en", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.google.com/docs/answer/15559738?hl=en" } ] }, { "ID": "120", "Software": "Google Flow", "Operation": "Insert and Remove edits to generated video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9ddfa4fc6056\ncollected-b52178d08899\ncollected-74428fad396a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Google Flow's documented edit route inserts or removes scene content through learned video revision.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 5 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=5; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Flow video edit; model not named for this mode). Designer controls: prompt describing the object to insert or remove; box drawn on the frame (Rectangle). Returns count not documented; a flat video. “Tap the pencil icon on any clip to insert objects directly into videos or remove elements, without changing anything else.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://blog.google/innovation-and-ai/products/veo-updates-flow/\nhttps://support.google.com/labs/answer/16935718?hl=en\nhttps://blog.google/technology/google-labs/flow-refine-videos/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9ddfa4fc6056", "Operation IDs": "120, 121", "Citation": "Google Veo 3.1 and Flow release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.google/innovation-and-ai/products/veo-updates-flow/" }, { "Source ID": "collected-b52178d08899", "Operation IDs": "120", "Citation": "Google Flow video editing help", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.google.com/labs/answer/16935718?hl=en" }, { "Source ID": "collected-74428fad396a", "Operation IDs": "120", "Citation": "blog.google — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.google/technology/google-labs/flow-refine-videos/" } ] }, { "ID": "121", "Software": "Google Flow", "Operation": "Veo 3.1 video generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–4.4", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f2d40a76ac91\ncollected-6a0f80fc52e6\ncollected-9ddfa4fc6056\ncollected-aa875e8dbe31\ncollected-e283d62b5ebc\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Google Flow invokes Veo 3.1 Fast in the recorded reference-conditioned video-generation mode.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–4.4; no midpoint is used.", "Previous record: calculation": "K=1; M=2; unknown=3; conditional=0; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 4 ÷ (6 − 1) = 4.4. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Veo 3.1 Fast, Ingredients to Video). Designer controls: prompt; ingredients (reference images or video); model choice; aspect ratio; number of outputs. Returns several to choose from (designer sets number of outputs; maximum not stated); a flat video: 8 s in Ingredients mode for Veo 3.1 Fast; generated audio. “Use multiple reference images to control the characters, objects and style.” Source", "Collected source links": "https://storage.googleapis.com/deepmind-media/Model-Cards/Veo-3-Model-Card.pdf\nhttps://storage.googleapis.com/deepmind-media/veo/Veo-3-Tech-Report.pdf\nhttps://blog.google/innovation-and-ai/products/veo-updates-flow/\nhttps://support.google.com/flow/answer/16935718?hl=en\nhttps://support.google.com/flow/answer/16353334?hl=en", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f2d40a76ac91", "Operation IDs": "121", "Citation": "Veo 3 family model card", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://storage.googleapis.com/deepmind-media/Model-Cards/Veo-3-Model-Card.pdf" }, { "Source ID": "collected-6a0f80fc52e6", "Operation IDs": "121", "Citation": "Veo technical report", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://storage.googleapis.com/deepmind-media/veo/Veo-3-Tech-Report.pdf" }, { "Source ID": "collected-9ddfa4fc6056", "Operation IDs": "120, 121", "Citation": "Google Veo 3.1 and Flow release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.google/innovation-and-ai/products/veo-updates-flow/" }, { "Source ID": "collected-aa875e8dbe31", "Operation IDs": "121", "Citation": "support.google.com — 16935718?hl=en", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.google.com/flow/answer/16935718?hl=en" }, { "Source ID": "collected-e283d62b5ebc", "Operation IDs": "121", "Citation": "support.google.com — 16353334?hl=en", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.google.com/flow/answer/16353334?hl=en" } ] }, { "ID": "122", "Software": "Google Slides", "Operation": "Gemini 3 Pro image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d635b6934cdf\ncollected-99bdf956b0e1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Google Slides invokes Gemini 3 Pro Image to create an image through Help me visualize.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Gemini 3 Pro Image). Designer controls: prompt; style (Add a style); aspect ratio (Square 1:1, Wide 16:9, Tall 9:16). Returns count not documented; a flat image: inserted image. “You can generate and edit an image with Gemini 3 Pro image in:” Source", "Collected source links": "https://support.google.com/docs/answer/16443280?hl=en\nhttps://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Image-Model-Card.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d635b6934cdf", "Operation IDs": "122", "Citation": "Google Slides image-generation help", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.google.com/docs/answer/16443280?hl=en" }, { "Source ID": "collected-99bdf956b0e1", "Operation IDs": "15, 122, 128", "Citation": "Gemini 3 Pro Image model card", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Image-Model-Card.pdf" } ] }, { "ID": "123", "Software": "Grasshopper + LunchBox plugin", "Operation": "LunchBoxML LightGBM Surface Regression", "Proposed X": "10.7", "Review status": "Proposed X calculated", "Previous X": "12.0", "S": "2", "U": "3", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "Vendor XY-to-Z surface example from random sampled points, fitted regressor followed by predictions at supplied XY locations; return plotted predicted surface.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Training XY/Z samples, target labels, regressor settings and test XY coordinates in the vendor surface example.", "Learned determination": "Data-fitted LightGBM trees predict Z at the supplied test coordinates.", "Conventional processing": "Plotting joins the predicted samples into the displayed surface.", "Model contribution in the result": "The returned surface retains the fitted predictive relationship. Fit and use are inside this declared endpoint.", "Why this scope score (S)": "Proposed inclusion: S2 coordinate field governing supplied coordinates; U3 unobserved values without a documented faithful source-surface recovery target; J0 separable point predictions. Shared model parameters do not by themselves make the predictions jointly spatial.", "Why this latitude score (U)": "Proposed inclusion: S2 coordinate field governing supplied coordinates; U3 unobserved values without a documented faithful source-surface recovery target; J0 separable point predictions. Shared model parameters do not by themselves make the predictions jointly spatial.", "Why this coordination score (J)": "Proposed inclusion: S2 coordinate field governing supplied coordinates; U3 unobserved values without a documented faithful source-surface recovery target; J0 separable point predictions. Shared model parameters do not by themselves make the predictions jointly spatial.", "Comparisons and other modes": "v1 applicability was pending; proposed X10.7. The same example-derived predictive role includes linear regression and fitted spline/RBF predictors. A designer-authored equation or direct fixed-kernel resampling without fitted predictive coefficients remains conventional. Reusable training does not imply a higher J.", "Citations": "LunchBoxML regression trainers and testers — https://apps.provingground.io/docs/lunchbox-documentation/lunchbox-ml/regression-trainers-and-testers/\nLightGBM project — https://github.com/lightgbm-org/LightGBM", "Source IDs": "CAL2-LUNCH\nCAL2-LGBM", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies training samples, target labels and test coordinates. Fitted trees predict Z values; conventional plotting connects them. Fit and use are included in this declared operation.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. v1 applicability was pending; proposed X10.7. The same example-derived predictive role includes linear regression and fitted spline/RBF predictors. A designer-authored equation or direct fixed-kernel resampling without fitted predictive coefficients remains conventional. Reusable training does not imply a higher J.", "Previous record: scope": "LunchBoxML trains a LightGBM regressor to predict surface height and constructs the resulting surface in Grasshopper.", "Previous record: rationale": "A learned numerical predictor approximating a target relation supports the fixed-function reference at 12. Its tree architecture is non-neural but remains learned; non-neural must not be confused with conventional no-learning computation.", "Previous record: calculation": "X = 12 + 0 × (12 − 12) = 12.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (12.0). LunchBoxML documentation names ML.NET LightGBM regression as the trainer in its regression-surface example. No newer LunchBox release changing the trainers was found; the food4rhino listing found shows 2024.2.9.0 (16 April 2024), and Proving Ground reported a Grasshopper 2 port in alpha in 2025. Not established: Whether LunchBoxML components are part of the Grasshopper 2 port is not documented. apps.provingground.io, undated • grasshopperdocs.com, undated", "Collected source links": "https://apps.provingground.io/docs/lunchbox-documentation/lunchbox-ml/regression-trainers-and-testers/\nhttps://github.com/microsoft/LightGBM\nhttps://grasshopperdocs.com/addons/lunchbox.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "32/3", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-123", "Sources": [ { "Source ID": "CAL2-LUNCH", "Operation IDs": "123", "Citation": "LunchBoxML regression trainers and testers", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: LunchBoxML regression trainers and testers", "Finding used in this review": "Saved reusable regressors; surface example maps XY to Z from random samples, not a documented faithful source-surface recovery task.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://apps.provingground.io/docs/lunchbox-documentation/lunchbox-ml/regression-trainers-and-testers/" }, { "Source ID": "CAL2-LGBM", "Operation IDs": "123", "Citation": "LightGBM project", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: LightGBM project", "Finding used in this review": "Tree-based supervised prediction; point predictions do not imply joint spatial synthesis.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://github.com/lightgbm-org/LightGBM" } ] }, { "ID": "124", "Software": "gsplat", "Operation": "Training Pipeline", "Proposed X": "7.2", "Review status": "Proposed X calculated", "Previous X": "Unresolved", "S": "4", "U": "2", "J": "3", "Reviewed on": "2026-10-14", "Configuration": "Fit and retain explicit Gaussian scene parameters from supplied multiview observations; app_opt=False, post_processing=None; assess representation and subsequent novel-view rendering.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Multiview images and camera information; baseline gsplat with app_opt=False and post_processing=None.", "Learned determination": "Optimization fits Gaussian position, shape, opacity and appearance to supplied observations.", "Conventional processing": "Rasterization renders the retained fitted representation at requested views.", "Model contribution in the result": "The representation predicts appearance at unprovided views. Its data-fitted state remains even without a separate neural appearance network.", "Why this scope score (S)": "Proposed inclusion: S4 reconstruction of an observed scene representation; U2 fidelity-led recovery from sampled observations, using the supplied-view L1 plus SSIM objective; J3 coupled observed-scene fitting. Novel views alone do not establish U4, and static multiview reconstruction is not J5 synthesis of evolving new content.", "Why this latitude score (U)": "Proposed inclusion: S4 reconstruction of an observed scene representation; U2 fidelity-led recovery from sampled observations, using the supplied-view L1 plus SSIM objective; J3 coupled observed-scene fitting. Novel views alone do not establish U4, and static multiview reconstruction is not J5 synthesis of evolving new content.", "Why this coordination score (J)": "Proposed inclusion: S4 reconstruction of an observed scene representation; U2 fidelity-led recovery from sampled observations, using the supplied-view L1 plus SSIM objective; J3 coupled observed-scene fitting. Novel views alone do not establish U4, and static multiview reconstruction is not J5 synthesis of evolving new content.", "Comparisons and other modes": "v1 applicability was pending; proposed X7.2. Calling all optimization learned would wrongly include specified CAD/physics. The distinguishing premise is fitted observed-instance predictive state. An earlier draft U4/X5.5 was rejected: incomplete observations also occur in ordinary fidelity recovery. A strict reused-model-only boundary would exclude this baseline; that is a scope decision, not evidence that the fit has no inference.", "Citations": "gsplat simple trainer — https://raw.githubusercontent.com/nerfstudio-project/gsplat/main/examples/simple_trainer.py\n3D Gaussian Splatting research — https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/", "Source IDs": "CAL2-GSPLAT\nCAL2-GSRESEARCH", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies images and camera information. Optimization fits Gaussian position, shape, opacity and appearance to observations. Rendering retains this data-fitted scene estimate; no separate neural appearance network is required.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. v1 applicability was pending; proposed X7.2. Calling all optimization learned would wrongly include specified CAD/physics. The distinguishing premise is fitted observed-instance predictive state. An earlier draft U4/X5.5 was rejected: incomplete observations also occur in ordinary fidelity recovery. A strict reused-model-only boundary would exclude this baseline; that is a scope decision, not evidence that the fit has no inference.", "Previous record: scope": "The gsplat baseline directly fits Gaussian scene parameters with optional neural appearance extensions disabled.", "Previous record: rationale": "Direct scene fitting exposes an unresolved boundary between learned representation and conventional estimation. Absence of neural layers alone cannot justify 15 under a learned-determination equation; a representation-consistent branch criterion is needed.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). gsplat v1.6.0 (2026) lists PPISP as an alternative to the bilateral grid (January 2026), faster rasterization (June 2026) and spherical-harmonics updates (August 2026); the baseline still fits Gaussian parameters directly, and no pretrained network is listed. Not established: The remaining gap is a calibration decision: whether per-scene fitting of a scene representation counts as learned determination on this scale, applied consistently across neural fields and explicit Gaussians. Vendor or project documentation cannot settle that criterion. github.com, Jan 2026", "Collected source links": "https://github.com/nerfstudio-project/gsplat\nhttps://github.com/nerfstudio-project/gsplat/blob/main/examples/simple_trainer.py", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "36/5", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-124", "Sources": [ { "Source ID": "CAL2-GSPLAT", "Operation IDs": "124", "Citation": "gsplat simple trainer", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: gsplat simple trainer", "Finding used in this review": "Explicit splat parameters and multiview loss; app_opt=False and post_processing=None baseline. Checkpoints retain fitted state. LPIPS evaluation is not a retained predictor.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/nerfstudio-project/gsplat/main/examples/simple_trainer.py" }, { "Source ID": "CAL2-GSRESEARCH", "Operation IDs": "124", "Citation": "3D Gaussian Splatting research", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: 3D Gaussian Splatting research", "Finding used in this review": "Observed-scene reconstruction and novel-view rendering; inverse graphics rather than new-scene semantic invention.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/" } ] }, { "ID": "125", "Software": "Hailuo by MiniMax", "Operation": "MiniMax H3 video generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-acfeb3278d87\ncollected-643740d8b1d7", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "MiniMax H3 generates audiovisual clips through the Hailuo generation route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (MiniMax H3, video with native audio). Designer controls: prompt; reference images; reference videos (e.g. camera movement); reference audio. Returns count not documented; a flat video: stereo sound, up to 15 s, up to 2K. “generating video with native stereo sound, up to 15 seconds at 2K resolution” Source", "Collected source links": "https://www.minimax.io/news/minimax-hailuo-23\nhttps://www.minimax.io/blog/minimax-h3", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-acfeb3278d87", "Operation IDs": "125", "Citation": "MiniMax · Hailuo overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.minimax.io/news/minimax-hailuo-23" }, { "Source ID": "collected-643740d8b1d7", "Operation IDs": "125", "Citation": "MiniMax H3 release and Hailuo access", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.minimax.io/blog/minimax-h3" } ] }, { "ID": "126", "Software": "Hatch Embroidery 4", "Operation": "PhotoStitch conversion with editable visual preview", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f00aed6c1a11\ncollected-cad5ad09bcc3\ncollected-25498dd77230", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Hatch PhotoStitch converts an image into a visual stitch design and preview.", "Previous record: rationale": "The reviewed conversion description does not establish a trained predictor or a wholly conventional algorithm. Automatic conversion cannot choose the equation branch.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). The Hatch Embroidery 4 Digitizer page describes Color PhotoStitch with smart color reduction and thread matching and stitch-direction analysis, alongside Reef PhotoStitch and PhotoFlash; the Hatch FAQ calls PhotoStitch a set of photographic auto-digitizing tools. Neither page names an algorithm or a trained model; the AI feature named for Hatch 4 is the AI Answers assistant. Not established: Whether color reduction, stitch-direction analysis and stitch generation in PhotoStitch use trained models or prescribed image-processing rules. hatchembroidery.com, Hatch Embroidery 4 • hatchembroidery.com", "Collected source links": "https://hatchembroidery.com/\nhttps://hatchembroidery.com/products/hatch-embroidery/digitizer\nhttps://hatchembroidery.com/faq/features", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f00aed6c1a11", "Operation IDs": "126", "Citation": "Hatch · Hatch Auto-Digitize/PhotoStitch", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://hatchembroidery.com/" }, { "Source ID": "collected-cad5ad09bcc3", "Operation IDs": "126", "Citation": "Hatch current PhotoStitch and editing scope", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://hatchembroidery.com/products/hatch-embroidery/digitizer" }, { "Source ID": "collected-25498dd77230", "Operation IDs": "126", "Citation": "hatchembroidery.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://hatchembroidery.com/faq/features" } ] }, { "ID": "127", "Software": "Higgsfield", "Operation": "Cinema Studio 4.0 directed shot generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-02dc636d3588\ncollected-7c6eeabafb44\ncollected-83e752be8826", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Cinema Studio 4.0 synthesizes a directed video shot through its selected generative route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Seedance 2.5, made by ByteDance. The Higgsfield help center states that Cinema Studio 4.0 runs on Seedance 2.5. Generates: Whole clip: every frame of a new video of 4 to 30 seconds. Without references the endpoint runs text-to-video; with any image, video or audio reference it runs reference-to-video. Kept from supplied media: Generate mode: no supplied frames are documented as kept in the output; up to 50 references condition the new frames (a face, a product's shape, a style). Edit Video mode: the help center states the edit changes an object or a face without regenerating the whole clip, so the rest of the supplied footage is kept. Not established: The identity of the 'native Cinema Studio models' named in the launch blog. higgsfield.ai • higgsfield.ai, 2026-08-12 • docs.higgsfield.ai\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Seedance 2.5 inside Cinema Studio 4.0). Designer controls: prompt; camera body; lens; focal length; aperture; camera movement presets (30+); genre; character emotion slider and @character tags; up to 50 references (images, videos, Soul Cast characters); duration up to 30 s. Returns count not documented; a flat video: up to 1080p; audio from the same pass. “Cinema Studio 4.0 runs on Seedance 2.5 and adds the Director's Panel” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-cinema-studio\nhttps://higgsfield.ai/blog/cinema-studio-4-0\nhttps://docs.higgsfield.ai/docs/llms-full.txt", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-02dc636d3588", "Operation IDs": "127, 146", "Citation": "Higgsfield Cinema Studio help and Seedance 2.5 attribution", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-cinema-studio" }, { "Source ID": "collected-7c6eeabafb44", "Operation IDs": "127, 146", "Citation": "Higgsfield Cinema Studio 4.0 feature guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/cinema-studio-4-0" }, { "Source ID": "collected-83e752be8826", "Operation IDs": "127, 146", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/llms-full.txt" } ] }, { "ID": "128", "Software": "Higgsfield", "Operation": "Nano Banana Pro masked image editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–3.5", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b4a9a14c2f84\ncollected-ac937f148647\ncollected-99bdf956b0e1\ncollected-5d0514efd100", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield invokes Nano Banana Pro for masked image revision.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 3 model-decided, 1 unknown and 1 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–3.5; no midpoint is used.", "Previous record: calculation": "K=0; M=3; unknown=1; conditional=1; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 2 ÷ (5 − 1) = 3.5. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Nano Banana Pro Inpaint (Higgsfield's name; editor route /edit?model=nano_banana_pro_inpaint), which Higgsfield describes as built on top of Nano Banana Pro. Nano Banana Pro is Gemini 3 Pro Image, made by Google DeepMind. Generates: Higgsfield documents new content generated inside the painted mask: removal with background reconstruction, replacement of the object, or modification per the instruction, matched to lighting, color palette, optics and perspective. Per its model card, the underlying model outputs a complete generated image from text and image inputs. Kept from supplied media: Higgsfield states that everything outside the mask is protected with pixel-level precision and that unmasked areas are preserved with zero distortions. No document describes the mechanism. Not established: Whether Higgsfield composites source pixels outside the mask after generation; how the painted mask is encoded for the model; whether Nano Banana Pro Inpaint is a Higgsfield-tuned variant; output resolution; whether the four-step sequence describes model internals or product behavior. higgsfield.ai • storage.googleapis.com, 2025-11 • ai.google.dev\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Nano Banana Pro / Gemini 3 Pro Image). Designer controls: mask brush; prompt; output quality (Low, Medium, High). Returns count not documented; a flat image. “Use the mask brush to highlight the exact object or region.” Source", "Collected source links": "https://higgsfield.ai/blog/Top-Editing-Tool-in-2025-Nano-Banana-Pro-Inpaint\nhttps://higgsfield.ai/image-editing\nhttps://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Image-Model-Card.pdf\nhttps://ai.google.dev/gemini-api/docs/image-generation", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b4a9a14c2f84", "Operation IDs": "128", "Citation": "Higgsfield Nano Banana Pro Inpaint workflow guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/Top-Editing-Tool-in-2025-Nano-Banana-Pro-Inpaint" }, { "Source ID": "collected-ac937f148647", "Operation IDs": "128", "Citation": "Higgsfield image editing product description and entry link", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/image-editing" }, { "Source ID": "collected-99bdf956b0e1", "Operation IDs": "15, 122, 128", "Citation": "Gemini 3 Pro Image model card", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Image-Model-Card.pdf" }, { "Source ID": "collected-5d0514efd100", "Operation IDs": "128", "Citation": "ai.google.dev", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://ai.google.dev/gemini-api/docs/image-generation" } ] }, { "ID": "129", "Software": "Higgsfield", "Operation": "Soul 2.0 image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b629b6b09a31\ncollected-9992eb6e373f\ncollected-49bd58d964e9\ncollected-78f4d564ff27\ncollected-3b9099e1ca4a\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Soul 2.0 synthesizes image content through Higgsfield's trained image model.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Soul 2.0, Higgsfield's in-house foundation image model (API id higgsfield-ai/soul/v2/standard, released on the Higgsfield API 02/01/2026). Generates: Whole image: every pixel of each output frame, from a text prompt or from a reference image. API resolution 720p to 1080p (CLI quality tiers 1.5k and 2k); up to 4 images per generation. Kept from supplied media: No pixels from supplied media are documented as kept. A reference image guides style and composition; when a reference image is supplied the prompt field is disabled and the reference becomes the generation's direction. Not established: Architecture, base model, training data and whether Soul 2.0 is trained from scratch or adapted from an existing open model. higgsfield.ai, 2026-09-08 • designboom.com, 2026-04-03 • higgsfield.ai\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Soul 2.0). Designer controls: prompt; preset or Moodboard; Soul ID character; Soul HEX color palette; aspect ratio; quality (1.5k, 2k); batch size (1-4); seed and Soul ID strength (API). Returns several to choose from (up to 4); a flat image. “Up to 4 per generation. Set the batch size in the generation bar before clicking Generate.” Source", "Collected source links": "https://higgsfield.ai/soul-intro\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-soul-to-generate-images\nhttps://higgsfield.ai/blog/soul-2-vs-nano-banana-pro\nhttps://www.designboom.com/technology/higgsfield-art-directed-quality-ai-image-generation/\nhttps://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b629b6b09a31", "Operation IDs": "129", "Citation": "Higgsfield Soul product and model positioning", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/soul-intro" }, { "Source ID": "collected-9992eb6e373f", "Operation IDs": "129, 135", "Citation": "Higgsfield Soul family help and reference restrictions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-soul-to-generate-images" }, { "Source ID": "collected-49bd58d964e9", "Operation IDs": "129", "Citation": "higgsfield.ai, 2026-09-08", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/soul-2-vs-nano-banana-pro" }, { "Source ID": "collected-78f4d564ff27", "Operation IDs": "129", "Citation": "designboom.com, 2026-04-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.designboom.com/technology/higgsfield-art-directed-quality-ai-image-generation/" }, { "Source ID": "collected-3b9099e1ca4a", "Operation IDs": "129, 130, 138, 142, 152, 154, 156", "Citation": "higgsfield.ai — whats-in-my-higgsfield-account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account" } ] }, { "ID": "130", "Software": "Higgsfield / Kling AI", "Operation": "Kling 3.0 Motion Control performance transfer", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-aa31801357c3\ncollected-455a856bc45b\ncollected-cc7bb8095cd6\ncollected-e51998bf6799\ncollected-3b9099e1ca4a\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Kling 3.0 Motion Control synthesizes character footage using a performance reference.", "Previous record: rationale": "Current placement unresolved. K=3; M=0; unknown=1; conditional=2; A=6. Lower: 2 + 3 × 3 ÷ (6 − 1) = 3.8. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.8; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Kling 3.0 Motion Control (Kling VIDEO 3.0 Motion Control), made by Kuaishou (Kling AI). The Higgsfield API exposes it as POST /kling-video/v3/motion-control/pro and /kling-video/v3/motion-control/std. Generates: Whole video clip. The model generates every frame: the character from the image performing the movement, facial expression, hand gestures and pacing of the motion video. Kept from supplied media: Character identity and visual appearance come from the character image, used as a conditioning reference. Motion, facial expression and timing come from the motion video. Not established: Documentation does not establish: whether the Kling report describes the 3.0 product or an earlier Motion Control version; how Higgsfield's background choice maps to the API (the API schema lists no background parameter, so it may correspond to character_orientation); whether any image background or clothing pixels survive unchanged; whether a pose or keypoint extraction step runs before generation. docs.higgsfield.ai • higgsfield.ai, 2026-03-05 • kling.ai\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Kling 3.0 Motion Control). Designer controls: character image; motion reference video (3-30 s, or Motion Library); prompt (background elements); character orientation (match video or image); keep original sound; mode (Standard/Professional). Returns count not documented; a flat video: length matches reference video. “Kling 3.0 Motion Control (Pro) generates videos by animating a source image according to the motion in a provided video.” Source • Screenshot", "Collected source links": "https://open.higgsfield.ai/models/kling-video/v3/motion-control/pro/api-reference\nhttps://kling.ai/quickstart/motion-control-user-guide\nhttps://docs.higgsfield.ai/docs/models/kling-3-motion-control\nhttps://higgsfield.ai/blog/kling-motion-control-3\nhttps://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-aa31801357c3", "Operation IDs": "130", "Citation": "Higgsfield Kling 3.0 Motion Control Pro API schema", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://open.higgsfield.ai/models/kling-video/v3/motion-control/pro/api-reference" }, { "Source ID": "collected-455a856bc45b", "Operation IDs": "130", "Citation": "Kling AI Motion Control guide dated 03052026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://kling.ai/quickstart/motion-control-user-guide" }, { "Source ID": "collected-cc7bb8095cd6", "Operation IDs": "130", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/models/kling-3-motion-control" }, { "Source ID": "collected-e51998bf6799", "Operation IDs": "130", "Citation": "higgsfield.ai, 2026-03-05", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/kling-motion-control-3" }, { "Source ID": "collected-3b9099e1ca4a", "Operation IDs": "129, 130, 138, 142, 152, 154, 156", "Citation": "higgsfield.ai — whats-in-my-higgsfield-account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account" } ] }, { "ID": "131", "Software": "Higgsfield / Kling AI", "Operation": "Kling O3 targeted video appearance edit", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0ff01f27b96e\ncollected-eb26e4671417\ncollected-290b8483835a\ncollected-9a91dd0479cf\ncollected-96c7956a2916\ncollected-7ced4b964f4b\ncollected-22f911c609a1\ncollected-22f5b9391d6c\ncollected-7744bb050459\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Kling O3 performs the documented targeted video appearance edit.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=1; conditional=5; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Kling O3, the Kling VIDEO 3.0 Omni model made by Kuaishou (Kling AI). Higgsfield labels it Kling 3.0 Omni Edit in the web app; the API endpoint is kling-video/o3/video-edit with modes std, pro and 4k. Generates: The Omni-Generator synthesizes the output video from the source video, the instruction and optional references. The model selects the edited element (object, subject appearance, outfit, background, style) from the instruction; no mask is supplied. Kept from supplied media: Higgsfield states that motion and composition stay intact and that the change sits inside the original shot. Output duration is derived from the source video. Not established: No O3-specific technical report located; docs.higgsfield.ai • higgsfield.ai • higgsfield.ai\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Kling O3 Video Edit). Designer controls: input video; prompt naming the change; optional reference images or elements; mode (standard, pro, 4K). Returns count not documented; a flat video. “Kling O3 Video Edit generates edited videos from a single input video URL and optional images or elements, guided by a text prompt.” Source", "Collected source links": "https://higgsfield.ai/blog/Kling-01-is-Here-A-Complete-Guide-to-Video-Model\nhttps://app.klingai.com/cn/quickstart/klingai-video-3-model-user-guide\nhttps://docs.higgsfield.ai/docs/models/kling-o3/video-edit\nhttps://higgsfield.ai/ai-video-editor\nhttps://higgsfield.ai/video-to-video\nhttps://open.higgsfield.ai/models/kling-video/o3/video-edit/playground\nhttps://open.higgsfield.ai/terms-of-service\nhttps://kling.ai/quickstart/klingai-video-o1-user-guide\nhttps://docs.higgsfield.ai/docs/concepts/requests", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0ff01f27b96e", "Operation IDs": "131", "Citation": "Higgsfield Kling O1 generation versus edit modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/Kling-01-is-Here-A-Complete-Guide-to-Video-Model" }, { "Source ID": "collected-eb26e4671417", "Operation IDs": "131", "Citation": "Kling Video 3.0 guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://app.klingai.com/cn/quickstart/klingai-video-3-model-user-guide" }, { "Source ID": "collected-290b8483835a", "Operation IDs": "131", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/models/kling-o3/video-edit" }, { "Source ID": "collected-9a91dd0479cf", "Operation IDs": "131", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/ai-video-editor" }, { "Source ID": "collected-96c7956a2916", "Operation IDs": "131", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/video-to-video" }, { "Source ID": "collected-7ced4b964f4b", "Operation IDs": "131", "Citation": "open.higgsfield.ai — playground", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://open.higgsfield.ai/models/kling-video/o3/video-edit/playground" }, { "Source ID": "collected-22f911c609a1", "Operation IDs": "131", "Citation": "open.higgsfield.ai — terms-of-service", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://open.higgsfield.ai/terms-of-service" }, { "Source ID": "collected-22f5b9391d6c", "Operation IDs": "131", "Citation": "kling.ai — klingai-video-o1-user-guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://kling.ai/quickstart/klingai-video-o1-user-guide" }, { "Source ID": "collected-7744bb050459", "Operation IDs": "131", "Citation": "docs.higgsfield.ai — requests", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/concepts/requests" } ] }, { "ID": "132", "Software": "Higgsfield / Kling AI", "Operation": "Kling 3.0 prompt-directed clip generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9b80e3f50aa4\ncollected-effdb7b60b67\ncollected-effeb6cd38f3\ncollected-58ce1859c0d3\ncollected-64a7c4dd2d13", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Kling VIDEO 3.0 generates a clip through the specified generation route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Kling 3.0 (Kling VIDEO 3.0), made by Kuaishou Technology. Higgsfield also offers Kling 3.0 Turbo (changelog 2026-06-16) and Kling 3.0 Omni. Generates: Whole clip: every frame of a new video of 3 to 15 seconds at 720p, 1080p or 4K, with multi-shot structure (scene cuts) inside one generation. In auto multi-shot mode the model splits the prompt into shots; in custom mode the creator defines each shot. Kept from supplied media: Text-to-video: nothing from supplied media. With a start frame, the supplied image is the opening frame and its composition is the base of the clip; with start and end frames the model generates the path between the two images. Not established: Architecture and training data of Kling 3.0. higgsfield.ai • kling.ai, 2026-02-06 • arxiv.org, 2025-12-18\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Kling VIDEO 3.0). Designer controls: prompt; start and end frames; element reference; multi-shot (auto) or custom multi-shot (shot count, per-shot durations and prompts); duration (3-15 s); resolution (720p/1080p); native audio. Returns count not documented; a flat video. “The model will automatically plan scene transitions, shot framing, and camera angle changes based on the prompts” Source", "Collected source links": "https://www.prnewswire.com/news-releases/kling-ai-launches-3-0-model-ushering-in-an-era-where-everyone-can-be-a-director-302679944.html\nhttps://ir.kuaishou.com/news-releases/news-release-details/kling-ai-launches-30-model-ushering-era-where-everyone-can-be\nhttps://higgsfield.ai/blog/Kling-3.0-is-on-Higgsfield-User-Guide-AI-Video-Generation\nhttps://kling.ai/quickstart/klingai-video-3-model-user-guide\nhttps://arxiv.org/abs/2512.16776", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9b80e3f50aa4", "Operation IDs": "132", "Citation": "PR Newswire · Kling 3.0 Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.prnewswire.com/news-releases/kling-ai-launches-3-0-model-ushering-in-an-era-where-everyone-can-be-a-director-302679944.html" }, { "Source ID": "collected-effdb7b60b67", "Operation IDs": "132, 188", "Citation": "Kuaishou Kling 3.0 model family", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://ir.kuaishou.com/news-releases/news-release-details/kling-ai-launches-30-model-ushering-era-where-everyone-can-be" }, { "Source ID": "collected-effeb6cd38f3", "Operation IDs": "132", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/Kling-3.0-is-on-Higgsfield-User-Guide-AI-Video-Generation" }, { "Source ID": "collected-58ce1859c0d3", "Operation IDs": "132", "Citation": "kling.ai, 2026-02-06", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://kling.ai/quickstart/klingai-video-3-model-user-guide" }, { "Source ID": "collected-64a7c4dd2d13", "Operation IDs": "132", "Citation": "arxiv.org, 2025-12-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://arxiv.org/abs/2512.16776" } ] }, { "ID": "133", "Software": "Higgsfield", "Operation": "AI Influencer attribute-conditioned character image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b01e4d4c9304\ncollected-93c17245b97f\ncollected-b47c34d237db\ncollected-5391443a80ee\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "AI Influencer generates a character image through its attribute-conditioned Builder.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. No Higgsfield page names the model that generates the AI Influencer character image. Generates: Whole image of a character (every pixel of the frame), built from 19 menu settings (identity, body, face, hair, style), each tuned on a three-level scale from Average to Extreme; output up to 4K. Human and non-human characters are supported. Kept from supplied media: An optional uploaded face photo: the documentation states the character keeps the facial features (identity) of that photo while age, body, hair and wardrobe change. Pixel-level reuse of the photo is not documented. Not established: Which image model generates the character. higgsfield.ai, 2026-10-03 • higgsfield.ai • higgsfield.ai\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Higgsfield character model; not named). Designer controls: attribute menus (character type, gender, ethnicity/origin base, skin color, eye color, skin conditions, age); advanced Face, Body, Style tabs; Type to customize character field (Edit); regenerate. Returns count not documented; a flat image: up to 4K; saved to library. “Character Type, Gender, Ethnicity / Origin Base, Skin Color, Eye Color, Skin Conditions, and Age.” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-ai-influencer\nhttps://higgsfield.ai/blog/how-to-create-ai-influencer\nhttps://higgsfield.ai/blog/new-ai-influencer\nhttps://higgsfield.ai/ai-influencer", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b01e4d4c9304", "Operation IDs": "133, 135", "Citation": "Higgsfield influencer", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-ai-influencer" }, { "Source ID": "collected-93c17245b97f", "Operation IDs": "133", "Citation": "Higgsfield AI Influencer guide, 06132026, retrieved September 22", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/how-to-create-ai-influencer" }, { "Source ID": "collected-b47c34d237db", "Operation IDs": "133", "Citation": "higgsfield.ai, 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/new-ai-influencer" }, { "Source ID": "collected-5391443a80ee", "Operation IDs": "133", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/ai-influencer" } ] }, { "ID": "134", "Software": "Higgsfield", "Operation": "Soul 2.0 Photodump coordinated image set", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.75–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4a2ed22ab443\ncollected-4fb1fc4b07f2\ncollected-8dae98cc4ec5", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Soul Photodump generates coordinated images using an established Soul ID and preset.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 3 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.75–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 1 ÷ (5 − 1) = 2.75. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Soul 2.0 (Higgsfield in-house) with a Soul ID character. The 2026 Soul Photodump is described as built on Soul ID inside Soul 2.0. Generates: A set of 15 whole images (2026 Soul Photodump), each fully generated, showing the same Soul ID character across varied scenes, settings, outfits and moods (selfies, candid moments, golden hour, urban, interiors). Two presets set the style of the set. Kept from supplied media: Face identity of the Soul ID character, learned from 20 or more uploaded photos (25 or more for the 2025 Photodump Studio). No pixels from the training photos are documented as kept. Not established: The coordination mechanism behind the set: per-image prompts, their author (template or language model), and whether any image conditions the next. higgsfield.ai, 2026-02-26 • higgsfield.ai, 2025-09-18 • higgsfield.ai, 2026-06-29\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Soul 2.0 with Soul ID). Designer controls: Soul ID character; preset (Male Archive, Cool Girl Dump). Returns several to choose from (15 photos per set; guide suggests curating to 10); a flat image. “a full set of 15 photos” Source • Screenshot", "Collected source links": "https://higgsfield.ai/blog/how-to-create-ai-photodump-higgsfield-soul\nhttps://higgsfield.ai/blog/Higgsfield-Photodump-Studio-Your-Viral-Photoshoot\nhttps://higgsfield.ai/blog/sould-id-best-character-consistency", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4a2ed22ab443", "Operation IDs": "134", "Citation": "Higgsfield photo", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/how-to-create-ai-photodump-higgsfield-soul" }, { "Source ID": "collected-4fb1fc4b07f2", "Operation IDs": "134", "Citation": "Earlier Photodump Studio guide, 09182025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/Higgsfield-Photodump-Studio-Your-Viral-Photoshoot" }, { "Source ID": "collected-8dae98cc4ec5", "Operation IDs": "134", "Citation": "higgsfield.ai, 2026-06-29", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/sould-id-best-character-consistency" } ] }, { "ID": "135", "Software": "Higgsfield", "Operation": "Soul Cinema cinematic still generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-504fc4d7d001\ncollected-b01e4d4c9304\ncollected-fdbd3ce2a51b\ncollected-9c734cf66ab6\ncollected-9992eb6e373f\ncollected-0eeba424b4a2", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Soul Cinema generates a cinematic still through its named image-generation route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Soul Cinema, a Higgsfield in-house image model in the Soul family (Soul Cinema Preview launched 2026-03-04). Generates: Whole still image with a built-in cinematic grade (film grain, era-specific color, lens character, foreground-background separation). Third-party API documentation lists 720p and 1080p; the Higgsfield CLI lists quality tiers 1.5k and 2k. Kept from supplied media: No pixels from supplied media are documented as kept. The CLI accepts at most one image reference and an optional Soul ID; the Soul ID carries a learned face identity into the image. Not established: Architecture and training data. higgsfield.ai • higgsfield.ai, 2026-03-04 • higgsfield.ai\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Soul Cinema image model, Soul family). Designer controls: prompt; reference image (disables prompt); Soul ID character; Soul HEX color palette; aspect ratio; quality; batch size. Returns several to choose from (up to 4); a flat image. “Up to 4 per generation. Set the batch size in the generation bar before clicking Generate.” Source", "Collected source links": "https://higgsfield.ai/blog/higgsfield-3d-jutsu\nhttps://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-ai-influencer\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/which-ai-model-should-i-use\nhttps://higgsfield.ai/blog/soul-cinema-preview\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-soul-to-generate-images\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-soul-cinema", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-504fc4d7d001", "Operation IDs": "135, 162, 163, 164", "Citation": "Higgsfield 3d", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-3d-jutsu" }, { "Source ID": "collected-b01e4d4c9304", "Operation IDs": "133, 135", "Citation": "Higgsfield influencer", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-ai-influencer" }, { "Source ID": "collected-fdbd3ce2a51b", "Operation IDs": "135, 160", "Citation": "higgsfield.ai, modified 2026-08-28", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/which-ai-model-should-i-use" }, { "Source ID": "collected-9c734cf66ab6", "Operation IDs": "135", "Citation": "higgsfield.ai, 2026-03-04", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/soul-cinema-preview" }, { "Source ID": "collected-9992eb6e373f", "Operation IDs": "129, 135", "Citation": "Higgsfield Soul family help and reference restrictions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-soul-to-generate-images" }, { "Source ID": "collected-0eeba424b4a2", "Operation IDs": "135", "Citation": "higgsfield.ai — how-do-i-use-soul-cinema", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-soul-cinema" } ] }, { "ID": "136", "Software": "Higgsfield", "Operation": "Layers automatic image decomposition", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e777837e3172\ncollected-ec333beb7aeb\ncollected-b8aa18da134a\ncollected-fbaad63d3a07\ncollected-7ccb344292c1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Layers decomposes an image into separately accessible elements.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=5; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Higgsfield's Layers page states that Layers is powered by Seedream 5.0, made by ByteDance Seed. Higgsfield's Layer Decomposition guide names no model for the decomposition tool itself. Generates: A stack of separate layers: a background base plate plus transparent element layers (subject, text blocks, decorations, other objects). ByteDance documents that background areas hidden behind extracted elements are inpainted, so those pixels are model-generated content absent from the source image. Kept from supplied media: Higgsfield and ByteDance describe the layers as separated from the input image, so visible layer content corresponds to the source. Neither documents whether the visible pixels of each layer are copied from the source with a predicted alpha matte or re-rendered by the model; Not established: Which model and version runs on Higgsfield (Seedream 5.0 or Seedream 5.0 Pro); whether element layers are pixel copies with mattes or regenerations; whether segmentation runs as a separate model; how Standard and Fast modes differ. higgsfield.ai • higgsfield.ai • higgsfield.ai, 2026-08-11\nChanged 3 Oct 2026 X 2.0–10.0 → 2.0. Higgsfield's Layers page (October 2026) states the product is powered by Seedream 5.0, and ByteDance's Seedream 5.0 Pro announcement (8 July 2026) states that layer separation inpaints background areas hidden by the extracted subject. Runway describes Seedream 5.0 Layers as a ByteDance image generation model (20 August 2026). Higgsfield's own changelog (11 August 2026) and blog (17 August 2026) describe subject, background and element layers without naming the fill step. Not established: Higgsfield's decomposition text does not state hidden-region completion; the link rests on its 'Powered by Seedream 5.0' statement. Whether Higgsfield calls the Seedream 5.0 Pro decomposition endpoint or another variant is undocumented. The new evidence also bears on IRM_OP_137 (manual recomposition after decomposition). Previous reason: Learned element separation can combine membership estimation with completion of hidden regions. The present evidence supports a 2–10 envelope rather than a point inferred from layer count. higgsfield.ai, page as of 3 Oct 2026 • seed.bytedance.com, 2026-07-08 • segmind.com", "Collected source links": "https://higgsfield.ai/higgsfield-layers\nhttps://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://higgsfield.ai/creator-hub/changelog\nhttps://seed.bytedance.com/en/blog/beyond-generation-it-understands-design-introducing-seedream-5-0-pro\nhttps://www.segmind.com/models/seedream-5-pro-layer-decomposition", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e777837e3172", "Operation IDs": "136, 137, 138, 139, 159", "Citation": "Higgsfield layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/higgsfield-layers" }, { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-b8aa18da134a", "Operation IDs": "136, 137, 138, 148, 153", "Citation": "higgsfield.ai, 2026-08-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/changelog" }, { "Source ID": "collected-fbaad63d3a07", "Operation IDs": "136", "Citation": "seed.bytedance.com, 2026-07-08", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://seed.bytedance.com/en/blog/beyond-generation-it-understands-design-introducing-seedream-5-0-pro" }, { "Source ID": "collected-7ccb344292c1", "Operation IDs": "136", "Citation": "www.segmind.com — seedream-5-pro-layer-decomposition", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.segmind.com/models/seedream-5-pro-layer-decomposition" } ] }, { "ID": "137", "Software": "Higgsfield", "Operation": "Layers manual recomposition after AI decomposition", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e777837e3172\ncollected-ec333beb7aeb\ncollected-b8aa18da134a\ncollected-32f0da9768eb", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Layers permits movement, resizing and deletion after automatic decomposition.", "Previous record: rationale": "The public account does not establish whether movement uses completed layers alone or invokes fresh learned filling. Clarify the active recomposition path before assigning conventional 15 or retained-learning X.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Higgsfield's changelog (11 Aug 2026) describes Layers as turning a flat image into separate editable elements to adjust and rebuild individually, and the vendor guide (28 Aug 2026) says decomposition returns background, subject and other elements as independently editable layers. A third-party walkthrough (Creative Pad Media, 17 Aug 2026) states that regenerating the image uses Seedream 5.0 Pro after text correction and lists move, resize and delete as edits; no source states whether moving a layer triggers that regeneration or whether the background layer is completed at decomposition. Not established: Whether move, resize and delete recompose completed layers alone or invoke Seedream 5.0 Pro or another model to fill or regenerate; whether background completion happens during decomposition. higgsfield.ai, 2026-08-11 • higgsfield.ai, 2026-08-28 • creativepadmedia.com, 2026-08-17 (third-party)", "Collected source links": "https://higgsfield.ai/higgsfield-layers\nhttps://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://higgsfield.ai/creator-hub/changelog\nhttps://www.creativepadmedia.com/fix-ai-generated-text-with-1-click-higgsfield-layers-edit/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e777837e3172", "Operation IDs": "136, 137, 138, 139, 159", "Citation": "Higgsfield layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/higgsfield-layers" }, { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-b8aa18da134a", "Operation IDs": "136, 137, 138, 148, 153", "Citation": "higgsfield.ai, 2026-08-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/changelog" }, { "Source ID": "collected-32f0da9768eb", "Operation IDs": "137", "Citation": "creativepadmedia.com, 2026-08-17 (third-party)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.creativepadmedia.com/fix-ai-generated-text-with-1-click-higgsfield-layers-edit/" } ] }, { "ID": "138", "Software": "Higgsfield", "Operation": "Layers Edit Text replacement in a raster", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "4.25–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec333beb7aeb\ncollected-b8aa18da134a\ncollected-e777837e3172\ncollected-3b9099e1ca4a\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Layers Edit Text detects text and regenerates the raster with replacement wording and matched styling.", "Previous record: rationale": "Working v5 documentation coding: 3 kept, 1 model-decided, 1 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 4.25–5.0; no midpoint is used.", "Previous record: calculation": "K=3; M=1; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 3 ÷ (5 − 1) = 4.25. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not named in Higgsfield's Edit Text documentation. Edit Text is a Higgsfield Layers feature, and the Layers page states that Layers is powered by Seedream 5.0 (ByteDance Seed). Generates: Newly rendered text in place of each detected line the user rewrites, set into the design with matching font, color and placement. Higgsfield describes the result as a new image returned with the replacement text. Kept from supplied media: Higgsfield states the tool works instead of rebuilding the image from scratch and places the new text into the original design. No document states whether pixels outside the replaced text are copied from the source or regenerated with the rest of the image; the tutorial wording ('Regenerating the image') suggests a full-image generation pass, held provisionally. Not established: Model identity for text detection and for generation; whether non-text pixels are composited from the source; whether the text region is masked before generation; output resolution. higgsfield.ai • higgsfield.ai, 2026-08-11 • higgsfield.ai\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Layers regeneration; model not named). Designer controls: Detect text; new wording per detected line. Returns count not documented; a flat image: result drops into the Layers project. “Detect text first, which finds every line of text on the image and lists it.” Source • Screenshot", "Collected source links": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://higgsfield.ai/creator-hub/changelog\nhttps://higgsfield.ai/higgsfield-layers\nhttps://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-b8aa18da134a", "Operation IDs": "136, 137, 138, 148, 153", "Citation": "higgsfield.ai, 2026-08-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/changelog" }, { "Source ID": "collected-e777837e3172", "Operation IDs": "136, 137, 138, 139, 159", "Citation": "Higgsfield layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/higgsfield-layers" }, { "Source ID": "collected-3b9099e1ca4a", "Operation IDs": "129, 130, 138, 142, 152, 154, 156", "Citation": "higgsfield.ai — whats-in-my-higgsfield-account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account" } ] }, { "ID": "139", "Software": "Higgsfield", "Operation": "Layers Remove Background on generated imagery", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e777837e3172\ncollected-ec333beb7aeb\ncollected-46dfc2d3deec\ncollected-1c684698859f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Remove Background extracts a foreground cutout with transparency.", "Previous record: rationale": "The documented extraction task supports a region-determination match. A new background is not included, and exact segmentation architecture remains undisclosed.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Higgsfield Layers documents Remove Background as cutting out the subject with a transparent alpha for compositing, and the stand-alone remover outputs a transparent PNG. Background generation and relighting are offered as separate Higgsfield tools outside this operation. Not established: The model behind the image_background_remover utility remains undisclosed. higgsfield.ai, undated, fetched 2026-10-03 • higgsfield.ai, undated, fetched 2026-10-03 • github.com, fetched 2026-10-03", "Collected source links": "https://higgsfield.ai/higgsfield-layers\nhttps://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://higgsfield.ai/apps/image-background-remover\nhttps://github.com/higgsfield-ai/cli/blob/main/MODELS.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e777837e3172", "Operation IDs": "136, 137, 138, 139, 159", "Citation": "Higgsfield layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/higgsfield-layers" }, { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-46dfc2d3deec", "Operation IDs": "139", "Citation": "higgsfield.ai, undated, fetched 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/apps/image-background-remover" }, { "Source ID": "collected-1c684698859f", "Operation IDs": "139, 158", "Citation": "github.com, fetched 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/higgsfield-ai/cli/blob/main/MODELS.md" } ] }, { "ID": "140", "Software": "Higgsfield", "Operation": "Expand image outpainting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec333beb7aeb\ncollected-fbd4761a3d62\ncollected-586b7a2c275a\ncollected-86ea0a93dc80\ncollected-5f463a541d2b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Expand synthesizes content beyond image boundaries.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: For the prompt-free Expand image app (higgsfield.ai/apps/expand-image) the model is not documented. Higgsfield's AI Image Extender page offers a choice of models (Nano Banana and Nano Banana Pro by Google, Seedream by ByteDance, FLUX.2 Pro by Black Forest Labs, Kling O1 Image by Kuaishou, Higgsfield Soul, GPT Image by OpenAI, Recraft) and shows Nano Banana Pro preselected in its generation bar. Generates: New content in the canvas area outside the original borders (background, scenery, context) to reach the chosen aspect ratio, matched to lighting, perspective and style. Kept from supplied media: Higgsfield states that the original photo stays pixel-perfect and that the AI paints outward from the existing edges without modifying the subject. No mechanism is documented. Not established: Model in the prompt-free Expand app; whether a compositing step keeps source pixels; output resolution; whether automatic model routing applies. higgsfield.ai • higgsfield.ai • higgsfield.ai\nDesigner side, 3 Oct 2026 Other. Decided by a trained model (Expand model; not named). Designer controls: uploaded image (PNG or JPG); target aspect ratio (16:9, 9:16, 4:3, 1:1). Returns count not documented; a flat image. “Your original photo stays pixel-perfect. The AI only paints outward from the existing edges.” Source • Screenshot", "Collected source links": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://higgsfield.ai/plugins/figma\nhttps://higgsfield.ai/apps/expand-image\nhttps://higgsfield.ai/ai-image-extender\nhttps://static.higgsfield.ai/expand-image/expand-step-2.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-fbd4761a3d62", "Operation IDs": "140, 144", "Citation": "Higgsfield figma", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/plugins/figma" }, { "Source ID": "collected-586b7a2c275a", "Operation IDs": "140", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/apps/expand-image" }, { "Source ID": "collected-86ea0a93dc80", "Operation IDs": "140", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/ai-image-extender" }, { "Source ID": "collected-5f463a541d2b", "Operation IDs": "140", "Citation": "static.higgsfield.ai — expand-step-2.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://static.higgsfield.ai/expand-image/expand-step-2.png" } ] }, { "ID": "141", "Software": "Higgsfield", "Operation": "Relight image lighting revision", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5321f31d645a\ncollected-ec333beb7aeb\ncollected-bb497f6e3883", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Relight combines a documented generative/depth-based process to reconstruct illumination.", "Previous record: rationale": "Learned lighting realization crosses fixed mapping and perceptual reconstruction, supporting 5–12. Interface light controls do not establish a recovered physical scene.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Higgsfield's Relight page (published 6 January 2026, modified 29 August 2026) still describes generative AI combined with depth mapping and names no model. A third-party site says the image is re-rendered with the new lighting; a search-summary claim that GPT Image 1.5 powers Relight does not appear on the vendor pages opened. Not established: Model identity, and whether the whole frame is regenerated or a lighting change is composited over preserved pixels; that distinction would move the comparison toward REF05 or toward REF12. higgsfield.ai, 2026-01-06; modified 2026-08-29 • aitoolcurator.com, not shown", "Collected source links": "https://higgsfield.ai/blog/Relight-Director-Style-Cinematic-Lighting\nhttps://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://www.aitoolcurator.com/blog/higgsfield-relight/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5321f31d645a", "Operation IDs": "141", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/Relight-Director-Style-Cinematic-Lighting" }, { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-bb497f6e3883", "Operation IDs": "141", "Citation": "aitoolcurator.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.aitoolcurator.com/blog/higgsfield-relight/" } ] }, { "ID": "142", "Software": "Higgsfield", "Operation": "Angles image viewpoint synthesis", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.75–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-62c095f0424b\ncollected-ec333beb7aeb\ncollected-84ea3e536d65\ncollected-e275a7dd08e5\ncollected-3b9099e1ca4a\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Angles synthesizes a new image viewpoint, including views not directly observed.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 2 unknown and 1 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.75–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=2; conditional=1; A=5. Lower: 2 + 3 × 1 ÷ (5 − 1) = 2.75. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Higgsfield Angles (Angles 2.0 / v2). Higgsfield calls it a high-speed model and names no model or maker. Generates: A new still image of the input subject or scene seen from a chosen viewpoint (rotation, tilt, zoom, full 360 degrees including behind the subject). The whole output image is generated. Kept from supplied media: Subject identity and scene content of the input image, used as conditioning; the blog claims character or object consistency across angles. No document states that input pixels are kept unchanged. Not established: Documentation does not establish: the model and its maker; whether any 3D geometry is estimated; how unseen regions are filled; whether Angles 1 and Angles 2.0 run the same model. higgsfield.ai • chasejarvis.com, not shown • creativepadmedia.com\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Angles model; not named). Designer controls: uploaded image; interactive 3D rotation; manual sliders; quick-select angles; lighting brightness and color. Returns several to choose from (12 via Generate from all angles; one per single-angle Generate); a flat image: aspect ratio matches input. “Click 'Generate from all angles' to instantly get results from different viewpoints.” Source • Screenshot", "Collected source links": "https://higgsfield.ai/blog/Change-the-Angle-of-Any-Image\nhttps://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://chasejarvis.com/blog/higgsfield-angles-2-0-is-here-my-100-honest-review/\nhttps://www.creativepadmedia.com/higgsfield-shots-turns-1-image-into-9-cinematic-angles/\nhttps://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-62c095f0424b", "Operation IDs": "142", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/Change-the-Angle-of-Any-Image" }, { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-84ea3e536d65", "Operation IDs": "142", "Citation": "chasejarvis.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://chasejarvis.com/blog/higgsfield-angles-2-0-is-here-my-100-honest-review/" }, { "Source ID": "collected-e275a7dd08e5", "Operation IDs": "142", "Citation": "creativepadmedia.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.creativepadmedia.com/higgsfield-shots-turns-1-image-into-9-cinematic-angles/" }, { "Source ID": "collected-3b9099e1ca4a", "Operation IDs": "129, 130, 138, 142, 152, 154, 156", "Citation": "higgsfield.ai — whats-in-my-higgsfield-account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account" } ] }, { "ID": "143", "Software": "Higgsfield", "Operation": "AI Skin Enhancer portrait reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec333beb7aeb\ncollected-76db6ba639c9\ncollected-7faf32440dea\ncollected-716b09f4e121", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "AI Skin Enhancer reconstructs portrait surface detail through its selected enhancement mode.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. A Higgsfield blog post on the Skin Enhancer (published 6 Dec 2025, modified 29 Aug 2026) states that the Higgsfield AI Skin Enhancer uses advanced diffusion models together with facial geometry mapping and pixel-level restoration, and a second Higgsfield post with the same dates describes diffusion models that recreate pore textures. A companion guide (modified 28 Aug 2026) lists three enhancement strengths: Soft Skin, Realistic Skin and Imperfect Skin. Not established: Model name; whether the diffusion component runs in each of the three strength settings; the share of the stated pixel-level restoration component. The diffusion statement comes from blog prose, without a technical page or model card. Previous reason: The trained restoration and plausible-detail responsibilities support 5–8. A known depicted identity does not reduce the model's rating through input specificity. higgsfield.ai, published 6 Dec 2025, modified 29 Aug 2026 • higgsfield.ai, published 6 Dec 2025, modified 29 Aug 2026 • higgsfield.ai, published 6 Dec 2025, modified 28 Aug 2026 (R", "Collected source links": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://higgsfield.ai/blog/The-Skin-Enhancer-How-to-Retouch-Pictures-with-AI\nhttps://higgsfield.ai/blog/How-AI-Skin-Enhancer-Retouches-Low-Quality-Photos-Images\nhttps://higgsfield.ai/blog/AI-Skin-Enhancer-How-AI-Retouching-Works-in-Editing", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-76db6ba639c9", "Operation IDs": "143", "Citation": "Higgsfield Skin Enhancer — modes and retouching workflow, 12062025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/The-Skin-Enhancer-How-to-Retouch-Pictures-with-AI" }, { "Source ID": "collected-7faf32440dea", "Operation IDs": "143", "Citation": "higgsfield.ai, published 6 Dec 2025, modified 29 Aug 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/How-AI-Skin-Enhancer-Retouches-Low-Quality-Photos-Images" }, { "Source ID": "collected-716b09f4e121", "Operation IDs": "143", "Citation": "higgsfield.ai, published 6 Dec 2025, modified 29 Aug 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/AI-Skin-Enhancer-How-AI-Retouching-Works-in-Editing" } ] }, { "ID": "144", "Software": "Higgsfield", "Operation": "Face Swap identity replacement in an image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.5–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4b36e9c36886\ncollected-fbd4761a3d62\ncollected-d7554b8a1401\ncollected-bfd6134e6975\ncollected-dad121c5891f\ncollected-ed6c11d7ff4e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Face Swap uses deep learning to realize a replacement identity within a target image.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 2 model-decided, 1 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 3.5–4.25; no midpoint is used.", "Previous record: calculation": "K=2; M=2; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 2 ÷ (5 − 1) = 3.5. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Higgsfield's Face Swap pages describe generative identity embedding and an AI Image Generator that analyzes both images; no model name or maker appears. Generates: A face in the base image replaced by the identity from a reference face photo, with lighting, shadow and facial geometry adjusted to the scene. Whether the model generates the face region alone or the whole frame is not documented; a third-party description says the destination is re-rendered with the new identity. Kept from supplied media: Higgsfield states the base image defines pose, light direction and framing. No document states that hair, body or background pixels stay unchanged; the FAQ says the new face is integrated into the scene, replacing the original. Not established: Model identity and maker; whether the output composites a generated face onto the original image or regenerates the full image; face detection model; output resolution. higgsfield.ai • higgsfield.ai • fluxokit.ai\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (deep learning face swap). Designer controls: target image; source face photo. Returns one result; a flat image. “Our tool uses the same underlying 'deep learning' technology” Source • Screenshot", "Collected source links": "https://higgsfield.ai/blog/AI-Face-Character-Swap-in-Video-Photo-PRO-Guide\nhttps://higgsfield.ai/plugins/figma\nhttps://higgsfield.ai/blog/a-simple-guide-to-face-swap\nhttps://higgsfield.ai/apps/face-swap\nhttps://fluxokit.ai/higgsfield-face-swap\nhttps://static.higgsfield.ai/landing-static/faceswap/step-1.webp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4b36e9c36886", "Operation IDs": "144", "Citation": "Higgsfield swap", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/AI-Face-Character-Swap-in-Video-Photo-PRO-Guide" }, { "Source ID": "collected-fbd4761a3d62", "Operation IDs": "140, 144", "Citation": "Higgsfield figma", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/plugins/figma" }, { "Source ID": "collected-d7554b8a1401", "Operation IDs": "144", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/a-simple-guide-to-face-swap" }, { "Source ID": "collected-bfd6134e6975", "Operation IDs": "144", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/apps/face-swap" }, { "Source ID": "collected-dad121c5891f", "Operation IDs": "144", "Citation": "fluxokit.ai — higgsfield-face-swap", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://fluxokit.ai/higgsfield-face-swap" }, { "Source ID": "collected-ed6c11d7ff4e", "Operation IDs": "144", "Citation": "static.higgsfield.ai — step-1.webp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://static.higgsfield.ai/landing-static/faceswap/step-1.webp" } ] }, { "ID": "145", "Software": "Higgsfield", "Operation": "Image Upscale reconstructed image output", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec333beb7aeb\ncollected-8c21dd614bd0\ncollected-489cd830fcb3\ncollected-27a841ee4940\ncollected-5cccdd590eb9", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield's recorded Topaz image-upscale route reconstructs a higher-resolution raster.", "Previous record: rationale": "The task spans perceptual reconstruction and fidelity restoration, supporting 5–8. Exact Topaz model identity is still required for finer placement.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). PetaPixel (6 Aug 2025) reports that Higgsfield's default image model is Topaz's standard Gigapixel technology, with standard, high-fidelity, text refine, art and CG, and low-res modes, and that Topaz Generative is a separate generative option. The current Higgsfield upscaler page also names generative Topaz Recovery and a ByteDance engine; Topaz classifies its Gigapixel core models as non-generative. Not established: Which Topaz model produced the recorded output. A core Gigapixel model would bring the Standard 2 evidence (IRM_OP_335) to bear; Recover or Topaz Generative would bring the diffusion evidence. petapixel.com, 6 Aug 2025 • higgsfield.ai, not shown • docs.topazlabs.com, not shown", "Collected source links": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield\nhttps://higgsfield.ai/blog/restore-old-photos-ai\nhttps://petapixel.com/2025/08/06/higgsfield-ai-brings-topazs-industry-leading-photo-and-video-upscaling-to-the-web/\nhttps://higgsfield.ai/ai-image-upscaler\nhttps://docs.topazlabs.com/topaz-gigapixel/enhancements/ai-models/core-models", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec333beb7aeb", "Operation IDs": "136, 137, 138, 139, 140, 141, 142, 143, 145", "Citation": "Higgsfield imageedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-photos-with-ai-higgsfield" }, { "Source ID": "collected-8c21dd614bd0", "Operation IDs": "145", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/restore-old-photos-ai" }, { "Source ID": "collected-489cd830fcb3", "Operation IDs": "145", "Citation": "petapixel.com, 6 Aug 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://petapixel.com/2025/08/06/higgsfield-ai-brings-topazs-industry-leading-photo-and-video-upscaling-to-the-web/" }, { "Source ID": "collected-27a841ee4940", "Operation IDs": "145", "Citation": "higgsfield.ai, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/ai-image-upscaler" }, { "Source ID": "collected-5cccdd590eb9", "Operation IDs": "145, 335", "Citation": "docs.topazlabs.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/ai-models/core-models" } ] }, { "ID": "146", "Software": "Higgsfield", "Operation": "Cinema Studio 4.0 temporal extension", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-02dc636d3588\ncollected-7c6eeabafb44\ncollected-83e752be8826\ncollected-9e3932c617cd\ncollected-c2291f6739c6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Cinema Studio's selected Seedance 2.5 extension route generates a continuation of an existing clip.", "Previous record: rationale": "Current placement unresolved. K=2; M=0; unknown=2; conditional=2; A=6. Lower: 2 + 3 × 2 ÷ (6 − 1) = 3.2. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.2; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Seedance 2.5, made by ByteDance, per Higgsfield's help center (Cinema Studio 4.0 runs on Seedance 2.5) and per the API documentation (extension uses the Seedance 2.5 endpoints, POST /bytedance/seedance-2.5/video-extend). Other Higgsfield pages complicate this: the 4.0 launch blog names native Soul Cinema Models with Seedance 2.5 available under multi-model support, and another help page calls Cinema Studio a tool rather than a model. Generates: New seconds of video placed after (forward) or before (backward) the uploaded clip, continuing its motion, lighting and scene; the launch post also names bridge generation between clips. Extension length 4 to 30 seconds through the API; scenes build up to 30 seconds in the app. Kept from supplied media: The uploaded clip serves as context. The API states output framing follows the source video; Not established: Documentation does not establish: whether the app's Extend control calls Seedance 2.5 video-extend or a native Soul Cinema model; whether the output file includes the original frames unchanged; how backward and bridge directions are requested (no direction field in the API schema); how bridge generation is implemented. docs.higgsfield.ai • docs.higgsfield.ai • docs.higgsfield.ai\nDesigner side, 3 Oct 2026 Other. Decided by a trained model (Seedance 2.5 extension route). Designer controls: uploaded video (up to 30 s); direction (forward or backward). Returns count not documented; a flat video. “Forward / Backward Extend takes an uploaded video (up to 30 seconds) and continues it seamlessly in either direction” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-cinema-studio\nhttps://higgsfield.ai/blog/cinema-studio-4-0\nhttps://docs.higgsfield.ai/docs/llms-full.txt\nhttps://docs.higgsfield.ai/docs/models/seedance-2-5\nhttps://docs.higgsfield.ai/docs/models/seedance-2-5/video-extend.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-02dc636d3588", "Operation IDs": "127, 146", "Citation": "Higgsfield Cinema Studio help and Seedance 2.5 attribution", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-cinema-studio" }, { "Source ID": "collected-7c6eeabafb44", "Operation IDs": "127, 146", "Citation": "Higgsfield Cinema Studio 4.0 feature guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/cinema-studio-4-0" }, { "Source ID": "collected-83e752be8826", "Operation IDs": "127, 146", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/llms-full.txt" }, { "Source ID": "collected-9e3932c617cd", "Operation IDs": "146", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/models/seedance-2-5" }, { "Source ID": "collected-c2291f6739c6", "Operation IDs": "146", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/models/seedance-2-5/video-extend.md" } ] }, { "ID": "147", "Software": "Higgsfield", "Operation": "Genjutsu Motion Transfer video reinterpretation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.2", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9e2af48c4907\ncollected-3cc53c497131\ncollected-ec06a4fd816f\ncollected-c588dceb9ac3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Genjutsu Motion Transfer rebuilds character or scene imagery while following a source performance.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 3.2.", "Previous record: calculation": "K=2; M=4; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 2 ÷ (6 − 1) = 3.2. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Genjutsu Motion Transfer, made by Higgsfield. The API endpoint sits in Higgsfield's own namespace (higgsfield/genjutsu/motion-transfer/v1.0), and Layer's model page calls it Higgsfield's video-to-video model. Generates: Whole new video. Characters, location, wardrobe, setting and visual style are regenerated from 1 to 8 reference images (API) and an optional prompt. Kept from supplied media: Motion, camera work and timing (shot blocking and overall shot structure) from the source video. Appearance comes from the reference images. Not established: Documentation does not establish: the underlying model and its maker; how source motion is represented (pose, depth, or direct video conditioning); whether any source pixels pass through unchanged. docs.higgsfield.ai • higgsfield.ai • higgsfield.ai\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Genjutsu Motion Transfer). Designer controls: source video (4-30 s); up to 30 reference images; optional prompt; presets; resolution. Returns count not documented; a flat video: up to 1080p. “Keep the motion, camera and timing while rebuilding the scene from your references.” Source", "Collected source links": "https://open.higgsfield.ai/models/higgsfiled/genjutsu/motion-transfer/v1.0\nhttps://higgsfield.ai/genjutsu\nhttps://docs.higgsfield.ai/docs/models/genjutsu/motion-transfer\nhttps://higgsfield.ai/blog/higgsfield-genjutsu", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9e2af48c4907", "Operation IDs": "147", "Citation": "Higgsfield genmotion", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://open.higgsfield.ai/models/higgsfiled/genjutsu/motion-transfer/v1.0" }, { "Source ID": "collected-3cc53c497131", "Operation IDs": "147, 148", "Citation": "Higgsfield genjutsu", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/genjutsu" }, { "Source ID": "collected-ec06a4fd816f", "Operation IDs": "147", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/models/genjutsu/motion-transfer" }, { "Source ID": "collected-c588dceb9ac3", "Operation IDs": "147, 148", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-genjutsu" } ] }, { "ID": "148", "Software": "Higgsfield", "Operation": "Genjutsu Object Swap in existing footage", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3cc53c497131\ncollected-6fbcf86f090e\ncollected-b8aa18da134a\ncollected-c588dceb9ac3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Genjutsu Object Swap generates a replacement object or character in existing footage.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=4; conditional=0; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Genjutsu, a Higgsfield tool (API endpoint higgsfield/genjutsu/object-swap/v1.0; launched in the web app 2026-09-01). Third-party models on the same API carry their maker's namespace (bytedance/, kling-video/); Generates: The element named by the user or a preset (character, outfit, product, location, object, on-screen text) replaced across the clip, guided by reference images. Output length follows the prepared source video; output up to 1080p. Kept from supplied media: Higgsfield states that the rest of the frame stays untouched, that everything else stays exactly as filmed, and that the change happens without regenerating the rest of the scene. The API page for the sibling Restyle workflow states that instructions do not guarantee exact preservation; no equivalent statement appears for Object Swap. Not established: Model identity; segmentation and tracking steps; whether non-target regions are composited from source frames or generated; whether Higgsfield trained the model or wraps a third-party model. higgsfield.ai, 2026-09-01 • higgsfield.ai, 2026-08-31 • higgsfield.ai\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Genjutsu Object Swap). Designer controls: source video (4-30 s); up to 30 reference images of the replacement; optional prompt; presets (30+). Returns count not documented; a flat video: up to 1080p. “Change a selected character, outfit, product, location or object while preserving the rest of the original shot.” Source", "Collected source links": "https://higgsfield.ai/genjutsu\nhttps://higgsfield.ai/blog/edit-ai-video-without-regenerating\nhttps://higgsfield.ai/creator-hub/changelog\nhttps://higgsfield.ai/blog/higgsfield-genjutsu", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3cc53c497131", "Operation IDs": "147, 148", "Citation": "Higgsfield genjutsu", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/genjutsu" }, { "Source ID": "collected-6fbcf86f090e", "Operation IDs": "148, 149", "Citation": "Higgsfield videoedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-ai-video-without-regenerating" }, { "Source ID": "collected-b8aa18da134a", "Operation IDs": "136, 137, 138, 148, 153", "Citation": "higgsfield.ai, 2026-08-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/changelog" }, { "Source ID": "collected-c588dceb9ac3", "Operation IDs": "147, 148", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-genjutsu" } ] }, { "ID": "149", "Software": "Higgsfield", "Operation": "Seedance 2.5 Draw to Edit video revision", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6fbcf86f090e\ncollected-8c813aa3eb96\ncollected-9dfbd744422d\ncollected-2cd2e4043a87", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Seedance 2.5 Draw to Edit synthesizes revised footage through the named model route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 5 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=5; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Seedance 2.5, made by ByteDance Seed. Higgsfield's help center names Seedance 2.5 Edit as a separate editing model that supports region-level edits marked directly on the frame. Generates: A revised video in which the region marked on a sketched or edited frame (an object, a face, a background detail) changes according to the instruction, carried across the clip. The Seedance video-edit model generates the output video; output framing follows the source. Kept from supplied media: Higgsfield states that the rest of the clip stays untouched. ByteDance describes targeted modifications with continuity before and after the edit, and its Seedance 2.0 report scores how well non-edited regions survive and records unintended modifications to non-edit regions as a failure mode. Not established: Which model, if any, edits the sketched frame before Seedance runs; how the marked region is passed to Seedance 2.5 (mask, reference frame, or prompt); whether unmarked pixels pass through unchanged; no Seedance 2.5 technical report. higgsfield.ai • github.com • higgsfield.ai\nDesigner side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by a trained model (Seedance 2.5 edit). Designer controls: marks drawn on a frame region; prompt describing the change inside the marks; up to 50 reference images; source video (up to 30 s). Returns count not documented; a flat video: up to 1080p. “marking the exact region on the frame that needs to change first, then describing what happens inside that marked area” Source", "Collected source links": "https://higgsfield.ai/blog/edit-ai-video-without-regenerating\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-seedance\nhttps://github.com/higgsfield-ai/skills/blob/main/higgsfield-generate/references/workflows.md\nhttps://higgsfield.ai/blog/seedance-2-5-on-higgsfield-2026", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6fbcf86f090e", "Operation IDs": "148, 149", "Citation": "Higgsfield videoedit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/edit-ai-video-without-regenerating" }, { "Source ID": "collected-8c813aa3eb96", "Operation IDs": "149, 161", "Citation": "higgsfield.ai, modified 2026-09-09", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-seedance" }, { "Source ID": "collected-9dfbd744422d", "Operation IDs": "149, 154", "Citation": "github.com — workflows.md", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/higgsfield-ai/skills/blob/main/higgsfield-generate/references/workflows.md" }, { "Source ID": "collected-2cd2e4043a87", "Operation IDs": "149", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/seedance-2-5-on-higgsfield-2026" } ] }, { "ID": "150", "Software": "Higgsfield", "Operation": "Speak 2.0 image-to-talking-video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1c91751bb7d9\ncollected-a2c9d31ba2b0\ncollected-4a385bb07fa1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Speak 2.0 generates talking-character video from the selected image-to-video route.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 3 unknown and 1 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=3; conditional=1; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Higgsfield Speak 2.0 (also listed as Speak v2), made by Higgsfield. Higgsfield's help center lists it as an image-to-video model in Lipsync Studio, beside seven third-party lipsync models. Generates: Whole talking video generated from one still image: lip movement synced to the speech, facial expression and head motion. Scene templates (General, Selfie, Podcast, Car Talking and more) set the look of the clip. Kept from supplied media: Facial identity and appearance of the person in the input image, used as conditioning (Segmind description). Documentation does not state whether background, framing or the first frame of the image are kept. Not established: Documentation does not establish: the model architecture or any base model; whether scene templates regenerate the background and camera or keep the image's scene; whether the first output frame matches the input image; the share of the frame the model regenerates beyond the face. higgsfield.ai, modified 2026-09-01 • segmind.com • x.com\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (talking-character image-to-video). Designer controls: character image; Audio text (typed script or generated audio); scene template (General, Selfie, Podcast, Car Talking); aspect ratio; model choice. Returns count not documented; a flat video. “Upload a character image or video, type what the character should say in Audio text” Source Includes a third-party source.", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-lipsync-voiceover-and-aspect-ratios\nhttps://www.segmind.com/models/higgsfield-speech2video\nhttps://x.com/higgsfield_ai/status/1968798199248044474", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1c91751bb7d9", "Operation IDs": "150, 151, 154", "Citation": "Higgsfield lipsync", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-lipsync-voiceover-and-aspect-ratios" }, { "Source ID": "collected-a2c9d31ba2b0", "Operation IDs": "150", "Citation": "www.segmind.com — higgsfield-speech2video", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.segmind.com/models/higgsfield-speech2video" }, { "Source ID": "collected-4a385bb07fa1", "Operation IDs": "150", "Citation": "x.com — 1968798199248044474", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://x.com/higgsfield_ai/status/1968798199248044474" } ] }, { "ID": "151", "Software": "Higgsfield", "Operation": "Sync Lipsync 3 mouth revision in existing video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–4.4", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1c91751bb7d9\ncollected-d4dde0e72fa5\ncollected-a50d59b021d4", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Sync Lipsync 3 generates revised mouth imagery and motion in an existing video.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–4.4; no midpoint is used.", "Previous record: calculation": "K=1; M=2; unknown=3; conditional=0; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 4 ÷ (6 − 1) = 4.4. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Sync Lipsync 3, listed in Higgsfield's Lipsync Studio as a video-to-video model for precise lip sync up to 4K. The maker is sync. labs (Synchronicity Labs), whose current model is named sync-3; the match between Higgsfield's label and sync-3 rests on the name and the 4K claim and is held provisionally. Generates: New mouth, lip and lower-face appearance for the speaker in each frame, matched to the target audio. sync-3 generates from a wider spatial window around the face, generates around obstructions such as hands or microphones, and outputs up to 4K with built-in super resolution. Kept from supplied media: sync. labs documents that the generated lip region is composited back into the original frame and that background, body and hair stay untouched, with everything outside the lip and lower-face region preserved exactly. That statement belongs to the general pipeline page; the sync-3 page says the model decides what to generate and what to preserve from scene context and does not state the size of the composited region. Not established: Which sync. labs model version Higgsfield calls; the size of the region sync-3 composites back; whether Higgsfield adds post-processing such as upscaling. higgsfield.ai • sync.so • sync.so\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (sync-3 lip-sync, generates all frames of the shot at once). Designer controls: input video; audio or text (TTS); model choice; sync_mode (bounce/loop/cut_off/silence/remap); speaker selection (coordinates, frame number). Returns count not documented; a flat video. “builds a global understanding of a person across an entire shot, generating all frames at once rather than stitching together isolated segments” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-lipsync-voiceover-and-aspect-ratios\nhttps://sync.so/docs/models/sync-3\nhttps://sync.so/docs/product/how-ai-lip-sync-works", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1c91751bb7d9", "Operation IDs": "150, 151, 154", "Citation": "Higgsfield lipsync", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-lipsync-voiceover-and-aspect-ratios" }, { "Source ID": "collected-d4dde0e72fa5", "Operation IDs": "151", "Citation": "Sync Labs — sync-3 model guide; current native capabilities and options", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://sync.so/docs/models/sync-3" }, { "Source ID": "collected-a50d59b021d4", "Operation IDs": "151", "Citation": "sync.so — how-ai-lip-sync-works", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://sync.so/docs/product/how-ai-lip-sync-works" } ] }, { "ID": "152", "Software": "Higgsfield", "Operation": "Video Relight footage lighting revision", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d86fd0f662d4\ncollected-c2704e0f4c48\ncollected-3b9099e1ca4a\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Video Relight reconstructs video illumination through its documented AI service.", "Previous record: rationale": "The learned appearance transformation supports a 5–12 mapping-to-perceptual envelope. Undisclosed stages prevent a precise comparison with the separately documented Resolve Relight mechanism.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Video Relight became a standalone Higgsfield tool on 20 August 2026; the vendor page says lighting is re-rendered on the existing footage with motion, framing and composition kept. No model, stage or training objective is disclosed beyond the record's source. Not established: Model identity and whether frames are regenerated or a relighting layer is applied to preserved pixels. higgsfield.ai, 2026-08-20; modified 2026-08-28 • omidsaffari.com, 2026", "Collected source links": "https://higgsfield.ai/blog/video-relight-color-palette-higgsfield\nhttps://omidsaffari.com/blog/higgsfield-video-relight-footage-after-recording-2026\nhttps://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d86fd0f662d4", "Operation IDs": "152, 153", "Citation": "Higgsfield vrelight", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/video-relight-color-palette-higgsfield" }, { "Source ID": "collected-c2704e0f4c48", "Operation IDs": "152", "Citation": "omidsaffari.com, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://omidsaffari.com/blog/higgsfield-video-relight-footage-after-recording-2026" }, { "Source ID": "collected-3b9099e1ca4a", "Operation IDs": "129, 130, 138, 142, 152, 154, 156", "Citation": "higgsfield.ai — whats-in-my-higgsfield-account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account" } ] }, { "ID": "153", "Software": "Higgsfield", "Operation": "Color Palette grade of existing video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d86fd0f662d4\ncollected-b8aa18da134a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Color Palette transfers a preset or reference grade to existing video.", "Previous record: rationale": "The feature page does not distinguish a learned transfer from conventional color mapping. The platform's AI branding cannot resolve that branch.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Higgsfield launched Color Palette as a standalone video tool on 20-21 Aug 2026 (previously inside Cinema Studio 4.0), with 50+ presets and an optional reference image used to recolor the uploaded video. The vendor post says it applies grading without touching the shot, prices it the same as Video Relight (135 credits per 15-second clip) and describes the workflow as generating a new version; no model, LUT or transfer method is named. Not established: Whether the grade is produced by a trained video model re-rendering the clip or by a prescribed color mapping such as a LUT or color-statistics transfer. higgsfield.ai, 2026-08-20 • higgsfield.ai, 2026-08-21", "Collected source links": "https://higgsfield.ai/blog/video-relight-color-palette-higgsfield\nhttps://higgsfield.ai/creator-hub/changelog", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d86fd0f662d4", "Operation IDs": "152, 153", "Citation": "Higgsfield vrelight", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/video-relight-color-palette-higgsfield" }, { "Source ID": "collected-b8aa18da134a", "Operation IDs": "136, 137, 138, 148, 153", "Citation": "higgsfield.ai, 2026-08-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/changelog" } ] }, { "ID": "154", "Software": "Higgsfield", "Operation": "Reframe video outpainting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1c91751bb7d9\ncollected-ea3f12dea919\ncollected-939941ea1931\ncollected-528b58b5322b\ncollected-9dfbd744422d\ncollected-45297d0ef84f\ncollected-3b9099e1ca4a\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Reframe synthesizes additional surrounding footage for an expanded frame.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=6; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Higgsfield's agent-skill documentation lists Reframe as the reframe workflow, which it calls a chain, with std and pro modes; the models in the chain are not listed. Generates: New regions outside the original frame, filled to the new aspect ratio (9:16, 16:9, 4:3, 3:4, 21:9, 1:1); Higgsfield states the background is regenerated so subjects at the sides stay in frame. Kept from supplied media: Higgsfield states subjects stay in frame through automatic subject locking. No statement on pixel retention of the original frame area exists. Not established: Model identity for border generation and any first-frame step; whether original pixels pass through; how temporal consistency of generated borders is handled; crop versus regeneration conflict between Higgsfield pages. higgsfield.ai • higgsfield.ai • github.com\nDesigner side, 3 Oct 2026 Other. Decided by a trained model (Reframe model; not named). Designer controls: source video (up to 1 minute); target aspect ratio (9:16, 16:9, 4:3, 3:4, 21:9, 1:1); up to 4 reference images. Returns count not documented; a flat video. “AI regenerates the missing area, so everyone stays in frame” Source • Screenshot", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-lipsync-voiceover-and-aspect-ratios\nhttps://higgsfield.ai/creator-hub/help-center/integrations/external-integrations-higgsfield\nhttps://higgsfield.ai/ai-video-reframer\nhttps://higgsfield.ai/plugins/after-effects\nhttps://github.com/higgsfield-ai/skills/blob/main/higgsfield-generate/references/workflows.md\nhttps://static.higgsfield.ai/reframer/step-2.png\nhttps://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1c91751bb7d9", "Operation IDs": "150, 151, 154", "Citation": "Higgsfield lipsync", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-lipsync-voiceover-and-aspect-ratios" }, { "Source ID": "collected-ea3f12dea919", "Operation IDs": "154, 165, 167", "Citation": "Higgsfield plugins", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/integrations/external-integrations-higgsfield" }, { "Source ID": "collected-939941ea1931", "Operation IDs": "154", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/ai-video-reframer" }, { "Source ID": "collected-528b58b5322b", "Operation IDs": "154", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/plugins/after-effects" }, { "Source ID": "collected-9dfbd744422d", "Operation IDs": "149, 154", "Citation": "github.com — workflows.md", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/higgsfield-ai/skills/blob/main/higgsfield-generate/references/workflows.md" }, { "Source ID": "collected-45297d0ef84f", "Operation IDs": "154", "Citation": "static.higgsfield.ai — step-2.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://static.higgsfield.ai/reframer/step-2.png" }, { "Source ID": "collected-3b9099e1ca4a", "Operation IDs": "129, 130, 138, 142, 152, 154, 156", "Citation": "higgsfield.ai — whats-in-my-higgsfield-account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account" } ] }, { "ID": "155", "Software": "Higgsfield", "Operation": "Video Upscale with Topaz Starlight Precise", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0e61fd704a99\ncollected-114c05dec50c\ncollected-f47701a4c1bc\ncollected-161901231370", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield invokes Topaz Starlight Precise for video detail reconstruction.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Topaz documentation describes Starlight Precise 2.6 (Topaz Video release 1.7 onward, succeeding 2.5) as a diffusion video enhancement model and states that all Starlight models are diffusion-based. Higgsfield's video restoration post (published 3 Sep 2026) describes its Topaz route as diffusion-based enhancement that can add new detail beyond what the source holds. Not established: Training objective details; whether Higgsfield currently serves Starlight Precise 2.5 or 2.6. Previous reason: The restoration task and generative prior support 5–8. The host route does not establish a new inference level relative to its underlying model. developer.topazlabs.com, Starlight Precise 2.6 • docs.topazlabs.com, Precise 2.5/2.6 page • docs.topazlabs.com, not shown", "Collected source links": "https://higgsfield.ai/blog/ai-video-restoration-high-quality\nhttps://docs.topazlabs.com/topaz-video/project-starlight-series/starlight-precise-25\nhttps://developer.topazlabs.com/video-models/starlight/starlight-precise-2.6\nhttps://docs.topazlabs.com/topaz-video/project-starlight-series", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0e61fd704a99", "Operation IDs": "155", "Citation": "Higgsfield video restoration and Topaz hosted controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/ai-video-restoration-high-quality" }, { "Source ID": "collected-114c05dec50c", "Operation IDs": "155, 356", "Citation": "Topaz Starlight Precise 2.5 and 2.6 native documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/project-starlight-series/starlight-precise-25" }, { "Source ID": "collected-f47701a4c1bc", "Operation IDs": "155, 356", "Citation": "developer.topazlabs.com, Starlight Precise 2.6", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/starlight/starlight-precise-2.6" }, { "Source ID": "collected-161901231370", "Operation IDs": "155", "Citation": "docs.topazlabs.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/project-starlight-series" } ] }, { "ID": "156", "Software": "Higgsfield", "Operation": "Marketing Studio campaign image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b8ede620f48e\ncollected-8d7483a934b4\ncollected-0d539e9dc35f\ncollected-f14539865acf\ncollected-3b9099e1ca4a\ncollected-2f6d64b35f26\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Marketing Studio generates campaign imagery through its named image-generation service.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: The 'MS Image model' (Marketing Studio Image), Higgsfield-branded; the maker of the underlying model is not stated. API versions: 2.0 Alpha, 2.5 Flare and 2.5 Sunburst (listed as released 09/01/2026). Generates: Whole image: product shots (Studio, Lifestyle, With Model), ad images to Meta and Google placement specs, marketplace images, at 1k to 4k. The same model also edits supplied images when image URLs are passed. Kept from supplied media: The product from the uploaded product image or product link becomes the subject of the shot; the documentation does not state whether product pixels are composited unchanged or regenerated. Brand kit elements (logo, colors, fonts) extracted from a website are applied. Not established: Maker and base model of MS Image / Marketing Studio Image. higgsfield.ai, 2026-08-18 • open.higgsfield.ai • docs.higgsfield.ai\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Marketing Studio Image 2.0 Alpha). Designer controls: prompt; product or reference images (optional); resolution (1k, 2k, 4k); aspect ratio (auto, 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, 21:9); enhance_prompt with preset_id; quality; moderation. Returns count not documented; a flat image. “Marketing Studio Image generates and edits campaign images.” Source", "Collected source links": "https://open.higgsfield.ai/models/marketing-studio/image/playground\nhttps://higgsfield.ai/blog/new-marketing-studio-higgsfield\nhttps://open.higgsfield.ai/models/marketing-studio/image/api-reference\nhttps://docs.higgsfield.ai/docs/models/marketing-studio-image/generate-and-edit\nhttps://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account\nhttps://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b8ede620f48e", "Operation IDs": "156", "Citation": "Higgsfield marketing API", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://open.higgsfield.ai/models/marketing-studio/image/playground" }, { "Source ID": "collected-8d7483a934b4", "Operation IDs": "156, 157, 158", "Citation": "higgsfield.ai, 2026-08-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/new-marketing-studio-higgsfield" }, { "Source ID": "collected-0d539e9dc35f", "Operation IDs": "156", "Citation": "open.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://open.higgsfield.ai/models/marketing-studio/image/api-reference" }, { "Source ID": "collected-f14539865acf", "Operation IDs": "156", "Citation": "docs.higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.higgsfield.ai/docs/models/marketing-studio-image/generate-and-edit" }, { "Source ID": "collected-3b9099e1ca4a", "Operation IDs": "129, 130, 138, 142, 152, 154, 156", "Citation": "higgsfield.ai — whats-in-my-higgsfield-account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/getting-started/whats-in-my-higgsfield-account" }, { "Source ID": "collected-2f6d64b35f26", "Operation IDs": "156, 157, 158, 159", "Citation": "Higgsfield marketing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads" } ] }, { "ID": "157", "Software": "Higgsfield", "Operation": "Marketing Studio UGC video generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–4.4", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2f6d64b35f26\ncollected-0056726a3496\ncollected-c3d8ee103653\ncollected-8d7483a934b4", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Marketing Studio generates an advertising video through its documented UGC route.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–4.4; no midpoint is used.", "Previous record: calculation": "K=1; M=2; unknown=3; conditional=0; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 4 ÷ (6 − 1) = 4.4. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Seedance 2.0, made by ByteDance, per the undated Higgsfield AI Ad Generator page and Higgsfield's launch post of 2026-04-13 ('Marketing Studio powered by Seedance 2.0'). The relaunched Marketing Studio pages of August 2026 (blog and help center) do not name the video model. Generates: Whole clip up to 15 seconds styled as phone-shot creator content, in Faceless, Talking Head and Silent formats; speech, lip sync and picture are generated in one pass. Kept from supplied media: The product from the product image or link is inserted as the subject; the avatar or Soul ID face is kept as identity across variations. No supplied pixels are documented as kept unchanged in the output. Not established: Whether the August 2026 relaunch moved UGC video to Seedance 2.5 (Higgsfield posted 'Seedance 2.5 for AI UGC. higgsfield.ai • x.com • higgsfield.ai, 2026-08-18\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Marketing Studio; model not named). Designer controls: product images or product URL; website link (brand signals); format (Faceless, Talking Head, Silent); templates; Brand Kit; Soul ID spokesperson (20+ photos). Returns count not documented; a flat video: up to 15 s. “Creator-style ads that look phone-shot, in Faceless, Talking Head, and Silent formats” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads\nhttps://higgsfield.ai/ai-ad-generator\nhttps://x.com/higgsfield/status/2043752265472004424\nhttps://higgsfield.ai/blog/new-marketing-studio-higgsfield", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2f6d64b35f26", "Operation IDs": "156, 157, 158, 159", "Citation": "Higgsfield marketing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads" }, { "Source ID": "collected-0056726a3496", "Operation IDs": "157", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/ai-ad-generator" }, { "Source ID": "collected-c3d8ee103653", "Operation IDs": "157", "Citation": "x.com — 2043752265472004424", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://x.com/higgsfield/status/2043752265472004424" }, { "Source ID": "collected-8d7483a934b4", "Operation IDs": "156, 157, 158", "Citation": "higgsfield.ai, 2026-08-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/new-marketing-studio-higgsfield" } ] }, { "ID": "158", "Software": "Higgsfield", "Operation": "Marketing Studio reference-ad remake", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2f6d64b35f26\ncollected-8d7483a934b4\ncollected-1c684698859f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Marketing Studio remakes a reference advertisement with newly generated content.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=5; conditional=0; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.6; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not stated for this mode. The Marketing Studio video model is stated as Seedance 2.0 (ByteDance) on the AI Ad Generator page; see IRM_OP_157. Generates: Whole new clip (UGC video category, up to 15 seconds) that follows the structure and format of a supplied reference ad with the user's product in place of the original product. Kept from supplied media: From the reference ad: its structure and format (sequence and ad type), per the documentation; no frames or pixels of the reference ad are documented as kept. From the product image: the product as subject. Not established: The mechanism: script extraction by a language model, storyboard extraction, or direct video-reference conditioning. higgsfield.ai, 2026-08-18 • higgsfield.ai • github.com\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Marketing Studio UGC route). Designer controls: reference ad; product images or URL. Returns count not documented; a flat video. “Marketing Studio reproduces its structure and format with your product in place of the original.” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads\nhttps://higgsfield.ai/blog/new-marketing-studio-higgsfield\nhttps://github.com/higgsfield-ai/cli/blob/main/MODELS.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2f6d64b35f26", "Operation IDs": "156, 157, 158, 159", "Citation": "Higgsfield marketing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads" }, { "Source ID": "collected-8d7483a934b4", "Operation IDs": "156, 157, 158", "Citation": "higgsfield.ai, 2026-08-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/new-marketing-studio-higgsfield" }, { "Source ID": "collected-1c684698859f", "Operation IDs": "139, 158", "Citation": "github.com, fetched 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/higgsfield-ai/cli/blob/main/MODELS.md" } ] }, { "ID": "159", "Software": "Higgsfield", "Operation": "Marketing Studio editable poster workflow", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2f6d64b35f26\ncollected-e777837e3172\ncollected-2203989c3a2c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Marketing Studio generates a poster and exposes it through the Layers workflow.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=4; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Poster generation: the model is not named for the Posters category (the Marketing Studio image model documented elsewhere is MS Image / Marketing Studio Image; see IRM_OP_156). Editing: Higgsfield Layers, described in its page metadata as 'powered by Seedream 5.0' (ByteDance). Generates: First pass: a whole banner or poster image, including text and layout. Layers then decomposes the flat image into separate elements (subject, background, text lines, other objects). Kept from supplied media: Within Layers, the documentation states one object can be moved, replaced or edited without affecting the rest of the image, so unedited layers are kept. Edit Text keeps the original formatting and placement of text. Not established: Which model generates the poster in the first pass. higgsfield.ai • higgsfield.ai • higgsfield.ai\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Marketing Studio poster route). Designer controls: product images or URL; templates; Brand Kit; post-generation Layers editing (layout, text, elements). Returns count not documented; editable structure: layers in Higgsfield Layers: layout, text, elements. “Banners and posters you can edit after generation: layout, text, and elements, through Higgsfield Layers” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads\nhttps://higgsfield.ai/higgsfield-layers\nhttps://higgsfield.ai/marketing-studio-intro", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2f6d64b35f26", "Operation IDs": "156, 157, 158, 159", "Citation": "Higgsfield marketing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-marketing-studio-to-create-video-ads" }, { "Source ID": "collected-e777837e3172", "Operation IDs": "136, 137, 138, 139, 159", "Citation": "Higgsfield layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/higgsfield-layers" }, { "Source ID": "collected-2203989c3a2c", "Operation IDs": "159", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/marketing-studio-intro" } ] }, { "ID": "160", "Software": "Higgsfield", "Operation": "Popcorn storyboard frame generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b557d7029b2b\ncollected-fdbd3ce2a51b\ncollected-9f15e9755ac5", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Popcorn generates storyboard-frame imagery before board arrangement.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Popcorn, described by Higgsfield as its native model for storyboards and multi-frame work, made by Higgsfield. No base model or architecture name is given. Generates: A sequence of up to 8 still frames (a storyboard). Every frame is generated. Kept from supplied media: Character identity, lighting and atmosphere from up to four input reference images (for example character, setting, object, mood), used as conditioning and carried across the frames (facial details, clothing, proportions per the blog). No document states that input pixels are kept unchanged. Not established: Documentation does not establish: the model architecture; whether the Nebius-described model is the Popcorn model; whether frames are generated in one pass or one after another; whether a language model writes the per-frame prompts in Auto mode. higgsfield.ai, modified 2026-08-28 • higgsfield.ai, modified 2026-09-12 • higgsfield.ai, 2025-10-23\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Popcorn; model not named). Designer controls: per-shot prompts; up to four reference images (character, setting, object, mood); shot count in Auto Mode; aspect ratio (3:4, 2:3, 3:2, 1:1, 9:16); camera perspective; emotional tone. Returns several to choose from (4, 6 or 8 shots in Auto Mode; frames form a sequence); a flat image: frames on a storyboard. “You can input up to four reference images, describe each shot individually, and control everything from camera perspective to emotional tone.” Source", "Collected source links": "https://higgsfield.ai/blog/How-to-Use-AI-for-Storyboards-Higgsfield-Popcorn\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/which-ai-model-should-i-use\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-popcorn", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b557d7029b2b", "Operation IDs": "160", "Citation": "Higgsfield story", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/How-to-Use-AI-for-Storyboards-Higgsfield-Popcorn" }, { "Source ID": "collected-fdbd3ce2a51b", "Operation IDs": "135, 160", "Citation": "higgsfield.ai, modified 2026-08-28", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/which-ai-model-should-i-use" }, { "Source ID": "collected-9f15e9755ac5", "Operation IDs": "160", "Citation": "higgsfield.ai, modified 2026-09-12", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-popcorn" } ] }, { "ID": "161", "Software": "Higgsfield", "Operation": "Canvas Soul 2.0 to Seedance 2.5 video workflow", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-528409e79c7f\ncollected-e168ea7157e1\ncollected-8c813aa3eb96\ncollected-62ab2322d061", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Canvas chains Soul 2.0 image generation into Seedance 2.5 video generation.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Two models chained by the user in a Canvas node graph: Soul 2.0 (Higgsfield's in-house image model) generates the still; Seedance 2.5 (ByteDance) generates the video. Generates: Step 1: Soul 2.0 generates a whole still image. Step 2: Seedance 2.5 generates a whole video clip (up to 30 seconds) that uses the Soul image. Kept from supplied media: Within the chain, the Soul 2.0 image passes subject appearance to Seedance 2.5. Canvas help states Seedance nodes treat connected images as references (appearance kept as a condition, frame layout free); the API image-to-video workflow instead states output framing follows the input image. Not established: Documentation does not establish: which Seedance 2.5 workflow the Canvas node calls (reference-to-video or image-to-video); whether the Soul image becomes the first frame of the clip; whether Seedance 2.5 (rather than 2.0) is selectable in Canvas at the catalogue date; any Soul 2.0 architecture or training details. higgsfield.ai, modified 2026-09-01 • higgsfield.ai, modified 2026-09-09 • higgsfield.ai\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (chained models; help page names none). Designer controls: nodes; connection from one node's output to the next node's input; prompt. Returns count not documented; a flat video: generated image and video sit as connected nodes on the board. “Drag the output of one generation node into the input of the next.” Source", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-canvas\nhttps://open.higgsfield.ai/explore\nhttps://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-seedance\nhttps://higgsfield.ai/canvas-intro", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-528409e79c7f", "Operation IDs": "161", "Citation": "Higgsfield canvas", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/tools/how-do-i-use-canvas" }, { "Source ID": "collected-e168ea7157e1", "Operation IDs": "161", "Citation": "Higgsfield catalogue", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://open.higgsfield.ai/explore" }, { "Source ID": "collected-8c813aa3eb96", "Operation IDs": "149, 161", "Citation": "higgsfield.ai, modified 2026-09-09", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/ai-models/how-do-i-use-seedance" }, { "Source ID": "collected-62ab2322d061", "Operation IDs": "161", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/canvas-intro" } ] }, { "ID": "162", "Software": "Higgsfield", "Operation": "3D Jutsu agent-authored editable scene", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-504fc4d7d001", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "3D Jutsu's scene-building agent constructs a native editable scene.", "Previous record: rationale": "Learned object and spatial decisions survive native scene creation, supporting the current 2–10 semantic-construction envelope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). Higgsfield's 3D Jutsu post (5 September 2026) states the agent assembles objects, layout, lighting, cameras and animation as editable scene data, with a per-chat choice among 30+ LLMs and renders to MP4, still frame or GLB; listed scene content includes primitives, curated GLB assets and Mixamo characters. The post names no text-to-3D or image-to-3D model and does not state how the agent produces geometry. Not established: Whether the agent places library assets and primitives or generates meshes. higgsfield.ai, 2026-09-05", "Collected source links": "https://higgsfield.ai/blog/higgsfield-3d-jutsu", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-504fc4d7d001", "Operation IDs": "135, 162, 163, 164", "Citation": "Higgsfield 3d", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-3d-jutsu" } ] }, { "ID": "163", "Software": "Higgsfield", "Operation": "3D Jutsu manual layout after AI scene generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-504fc4d7d001", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "3D Jutsu's manual layout tools transform existing scene objects after generation.", "Previous record: rationale": "The scoped native transformations contain no newly executed learned planner. Earlier scene generation belongs to another operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: move; rotate; scale; duplicate; light swap; camera reframe; undo/redo; revision history. Returns one result; editable structure: 3D scene objects; renders to MP4, still frame or GLB. “objects move, rotate, scale, and duplicate, lights swap, cameras reframe. Undo, redo, and revision history apply to every change” Source", "Collected source links": "https://higgsfield.ai/blog/higgsfield-3d-jutsu", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-504fc4d7d001", "Operation IDs": "135, 162, 163, 164", "Citation": "Higgsfield 3d", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-3d-jutsu" } ] }, { "ID": "164", "Software": "Higgsfield", "Operation": "3D Jutsu camera trajectory editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-504fc4d7d001", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "3D Jutsu's manual camera-trajectory tools revise the existing scene animation.", "Previous record: rationale": "Prescribed camera-path editing is conventional; final rendering and newly requested agent motion are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: camera presets; trajectory editing; timeline; free viewport (Esc); undo/redo; revision history. Returns one result; editable structure: 3D scene with camera path; renders to MP4, still frame or GLB. “Camera presets and trajectory editing frame the shot; Esc leaves a camera view back to the free viewport” Source", "Collected source links": "https://higgsfield.ai/blog/higgsfield-3d-jutsu", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-504fc4d7d001", "Operation IDs": "135, 162, 163, 164", "Citation": "Higgsfield 3d", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-3d-jutsu" } ] }, { "ID": "165", "Software": "Higgsfield", "Operation": "Blender plugin Scene Builder scene authorship", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec5339a015a8\ncollected-ea3f12dea919", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield's Blender Scene Builder constructs native scene content through its scene-authoring service.", "Previous record: rationale": "Learned scene choices remain in the native objects. This host relationship supports 2–10 and is not an independent sample of the browser agent.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). The Higgsfield Blender plugin page (requires Blender 5.1 or later, current October 2026) lists Scene Builder as a tab separate from 3D Model, which uses Meshy 5, and from Image, Video, Camera and Asset tabs. Scene Builder delivers objects, layout and light as editable geometry; the page does not name its model or state whether its objects are generated meshes or library assets. Not established: Scene Builder's model and whether it generates meshes or places assets. higgsfield.ai, undated", "Collected source links": "https://higgsfield.ai/plugins/blender\nhttps://higgsfield.ai/creator-hub/help-center/integrations/external-integrations-higgsfield", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec5339a015a8", "Operation IDs": "165, 166", "Citation": "Higgsfield blender", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/plugins/blender" }, { "Source ID": "collected-ea3f12dea919", "Operation IDs": "154, 165, 167", "Citation": "Higgsfield plugins", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/integrations/external-integrations-higgsfield" } ] }, { "ID": "166", "Software": "Higgsfield", "Operation": "Blender plugin 3D Model mesh generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec5339a015a8\ncollected-f8566f6579f3\ncollected-0a9266d3168d\ncollected-b8a3e2d549b0\ncollected-7c20f7187428", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield's Blender 3D Model service generates mesh and material content.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Meshy 5, made by Meshy, per Higgsfield's Blender plugin blog FAQ and plugin page. A third-party explainer (scriptable.com) states Higgsfield's 3D features use Meshy for image-to-3D and rigging, Meta's SAM 3 for single objects, and a Tripo model for text-to-3D; Generates: Mesh geometry (quad topology, unwrapped UVs) and textures with PBR materials, both generated by the model, placed at the Blender 3D cursor. Rigging and animation are separate plugin functions. Kept from supplied media: With a reference image, shape and appearance are conditioned on the image (Meshy 5 image alignment); documentation does not state whether texture pixels are projected from the image or generated. With a text prompt, nothing comes from supplied media. Not established: Documentation does not establish: which engine handles text-only prompts (Meshy text-to-3D or a Tripo model); whether SAM 3 runs in the Blender tool; whether texture is projected from the reference image; what replaces Meshy 5 after 2026-10-10. higgsfield.ai • digitalproduction.com, 2026-09-23 • scriptable.com\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Higgsfield 3D Model service; model not named). Designer controls: prompt or reference photo; variants counter. Returns several to choose from (variants counter; maximum not stated); editable structure: mesh with materials, placed at the 3D cursor as editable Blender objects. “Type it or drop a reference - either way, a mesh lands at the 3D cursor” Source • Screenshot", "Collected source links": "https://higgsfield.ai/plugins/blender\nhttps://higgsfield.ai/blog/higgsfield-blender-plugin\nhttps://digitalproduction.com/2026/09/23/higgsfield-brings-ai-generation-into-blender-oh-goodey/\nhttps://scriptable.com/posts/higgsfield/3d-models/\nhttps://static.higgsfield.ai/plugins/blender/feature-step-2-thumbnail.webp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec5339a015a8", "Operation IDs": "165, 166", "Citation": "Higgsfield blender", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/plugins/blender" }, { "Source ID": "collected-f8566f6579f3", "Operation IDs": "166, 167", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-blender-plugin" }, { "Source ID": "collected-0a9266d3168d", "Operation IDs": "166", "Citation": "digitalproduction.com, 2026-09-23", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://digitalproduction.com/2026/09/23/higgsfield-brings-ai-generation-into-blender-oh-goodey/" }, { "Source ID": "collected-b8a3e2d549b0", "Operation IDs": "166", "Citation": "scriptable.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://scriptable.com/posts/higgsfield/3d-models/" }, { "Source ID": "collected-7c20f7187428", "Operation IDs": "166", "Citation": "static.higgsfield.ai — feature-step-2-thumbnail.webp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://static.higgsfield.ai/plugins/blender/feature-step-2-thumbnail.webp" } ] }, { "ID": "167", "Software": "Higgsfield", "Operation": "Blender plugin Character Animation rig and motion", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ea3f12dea919\ncollected-f8566f6579f3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Higgsfield Character Animation provides rigging and timeline motion through its Blender integration.", "Previous record: rationale": "The recorded feature description does not separate learned motion generation from conventional rigging or retargeting. Establish the selected processing path before issuing X.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Higgsfield's Blender plugin post (20 Aug 2026) states Character Animation comes with an armature fitted and weighted without manual skinning, and that a plain-language description of the move places keyframes on the timeline; generation runs in Higgsfield's cloud. No model, motion library or retargeting method is named. Not established: Whether keyframes come from a trained text-to-motion generator or from retrieval and retargeting of library motion; whether rig fitting and skin weighting use a learned model. higgsfield.ai, 2026-08-20 • higgsfield.ai", "Collected source links": "https://higgsfield.ai/creator-hub/help-center/integrations/external-integrations-higgsfield\nhttps://higgsfield.ai/blog/higgsfield-blender-plugin", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ea3f12dea919", "Operation IDs": "154, 165, 167", "Citation": "Higgsfield plugins", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/creator-hub/help-center/integrations/external-integrations-higgsfield" }, { "Source ID": "collected-f8566f6579f3", "Operation IDs": "166, 167", "Citation": "higgsfield.ai", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://higgsfield.ai/blog/higgsfield-blender-plugin" } ] }, { "ID": "168", "Software": "Hypar", "Operation": "Chat Beta door placement in an editable plan", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b0fba8c4e748\ncollected-789a07ffbfdc\ncollected-20a727a8fdcb\ncollected-b65b08378eb3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Hypar Chat Beta selects and places doors in a native plan through its documented chat-assisted workflow.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The in-scope learned contribution chooses types and wall positions for catalog door elements; door geometry comes from the catalog and wall attachment. That places the comparison nearer region determination than synthesis, at 6–10 within the same envelope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. Hypar's Use AI features page (verified against Hypar v0.2.035, 27 August 2026) describes Chat (Beta) acting directly in the model and states that a prompt to place doors and windows in all rooms usually works. Doors are catalog elements of set types (single swing, double swing, sliding and others) that snap to walls and take width, height and type from menus. The language model is not named. Not established: Whether the assistant computes door positions itself or calls a placement routine; which language model runs the chat. Previous reason: Learned spatial decisions remain after native wall attachment. The task supports 2–10; the earlier text-to-BIM prototype is not the current implementation. docs.hypar.io, verified against Hypar v0.2.035, 2026-08-27 • docs.hypar.io", "Collected source links": "https://aecmag.com/ai/hypar-text-to-bim-and-beyond/\nhttps://docs.hypar.io/\nhttps://docs.hypar.io/guides-task-based-how-tos-for-everything-you-do-in-hypar./use-ai-features\nhttps://docs.hypar.io/guides-task-based-how-tos-for-everything-you-do-in-hypar./model-your-plan", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b0fba8c4e748", "Operation IDs": "168", "Citation": "AEC Magazine · Text-to-BIM", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://aecmag.com/ai/hypar-text-to-bim-and-beyond/" }, { "Source ID": "collected-789a07ffbfdc", "Operation IDs": "168", "Citation": "Hypar · Current 3D planning scope", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.hypar.io/" }, { "Source ID": "collected-20a727a8fdcb", "Operation IDs": "168", "Citation": "docs.hypar.io, verified against Hypar v0.2.035, 2026-08-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.hypar.io/guides-task-based-how-tos-for-everything-you-do-in-hypar./use-ai-features" }, { "Source ID": "collected-b65b08378eb3", "Operation IDs": "168", "Citation": "docs.hypar.io", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.hypar.io/guides-task-based-how-tos-for-everything-you-do-in-hypar./model-your-plan" } ] }, { "ID": "169", "Software": "Ideogram", "Operation": "Canvas Magic Fill and Extend", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4859a931c70c\ncollected-01bf3064065c\ncollected-1858956f8567\ncollected-8609f5c5464c\ncollected-764e851b6447", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Ideogram Canvas Magic Fill and Extend synthesize image content within or beyond a selected region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (inpainting and outpainting). Designer controls: masking tools; generation window (position, size); prompt; Magic Prompt; style; model choice; aspect ratio. Returns several to choose from (4); editable structure: chosen image placed as a separate canvas object with layer order. “These tools place one selected image on the canvas, maintaining the size and position of the generation window used.” Source • Screenshot", "Collected source links": "https://docs.ideogram.ai/canvas-and-editing/canvas/magic-fill\nhttps://ideogram.ai/\nhttps://docs.ideogram.ai/canvas-and-editing\nhttps://x.com/ideogram_ai/status/1917985285679530232\nhttps://docs.ideogram.ai/canvas-and-editing/canvas/canvas-overview", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4859a931c70c", "Operation IDs": "169", "Citation": "Ideogram Magic Fill", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.ideogram.ai/canvas-and-editing/canvas/magic-fill" }, { "Source ID": "collected-01bf3064065c", "Operation IDs": "169, 170", "Citation": "Ideogram current platform", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://ideogram.ai/" }, { "Source ID": "collected-1858956f8567", "Operation IDs": "169", "Citation": "Ideogram Canvas and editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.ideogram.ai/canvas-and-editing" }, { "Source ID": "collected-8609f5c5464c", "Operation IDs": "169", "Citation": "Ideogram, May 1, 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://x.com/ideogram_ai/status/1917985285679530232" }, { "Source ID": "collected-764e851b6447", "Operation IDs": "169", "Citation": "docs.ideogram.ai — canvas-overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.ideogram.ai/canvas-and-editing/canvas/canvas-overview" } ] }, { "ID": "170", "Software": "Ideogram", "Operation": "Ideogram 4.0 image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-356465629afd\ncollected-01bf3064065c\ncollected-f460c1b56fce", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Ideogram 4.0 generates image content through the named generation route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Ideogram 4.0, open-weight, trained with bounding-box layout descriptions). Designer controls: prompt (text or JSON with bounding boxes); aspect ratio; resolution (1K/2K); quality (Turbo/Default/Quality); Magic Prompt; seed; number of images. Returns several to choose from (Auto or 1-8); a flat image. “Choose Auto or 1 to 8 images per run.” Source", "Collected source links": "https://the-decoder.com/ideogram-3-0s-ai-image-generator-gets-a-realism-boost-and-new-editing-tools/\nhttps://ideogram.ai/\nhttps://docs.ideogram.ai/create/image-settings", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-356465629afd", "Operation IDs": "170", "Citation": "The Decoder · Prompt Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://the-decoder.com/ideogram-3-0s-ai-image-generator-gets-a-realism-boost-and-new-editing-tools/" }, { "Source ID": "collected-01bf3064065c", "Operation IDs": "169, 170", "Citation": "Ideogram current platform", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://ideogram.ai/" }, { "Source ID": "collected-f460c1b56fce", "Operation IDs": "170", "Citation": "docs.ideogram.ai — image-settings", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.ideogram.ai/create/image-settings" } ] }, { "ID": "171", "Software": "Adobe Illustrator", "Operation": "Generative Recolor", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0ad28164dde3\ncollected-1acf4d80920f\ncollected-ac6ea5fe967e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Illustrator Generative Recolor predicts a palette and its assignment to existing vector artwork.", "Previous record: rationale": "Learned appearance allocation crosses fixed mapping and perceptual reconstruction, supporting 5–12. Existing path geometry does not numerically dilute that responsibility.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Current Adobe help for Illustrator Generative Recolor (Firefly-powered) describes a prompt that names a palette or theme, multiple output variations on selected colored vector artwork, and refinement of a variation through Edit > Edit Colors > Recolor Artwork. Path geometry stays as drawn; no 2025–2026 change to the feature's model is documented. Not established: Adobe does not state whether the trained model outputs a palette and color-group assignment for the prescribed Recolor Artwork operation, or derives colors from a generated image (the 2023 Forbes account in the record), or which Firefly model version runs. Adobe helpx refused direct fetch; quotes come from search-result text. helpx.adobe.com, not shown", "Collected source links": "https://helpx.adobe.com/illustrator/desktop/use-generative-ai/recolor-artwork-with-generative-recolor.html\nhttps://www.adobe.com/learn/illustrator/web/generative-recolor-overview\nhttps://helpx.adobe.com/ca/illustrator/using/generative-recolor.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0ad28164dde3", "Operation IDs": "171", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/illustrator/desktop/use-generative-ai/recolor-artwork-with-generative-recolor.html" }, { "Source ID": "collected-1acf4d80920f", "Operation IDs": "171", "Citation": "www.adobe.com — generative-recolor-overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.adobe.com/learn/illustrator/web/generative-recolor-overview" }, { "Source ID": "collected-ac6ea5fe967e", "Operation IDs": "171", "Citation": "helpx.adobe.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/ca/illustrator/using/generative-recolor.html" } ] }, { "ID": "172", "Software": "Adobe Illustrator", "Operation": "Manual Path and Anchor Editing Comparison", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8bda1e041651\ncollected-eb0cb723d272", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Illustrator directly modifies selected paths and anchor points.", "Previous record: rationale": "Bezier geometry edits and specified appearance changes are conventional; the nearby Firefly operations are separately scoped.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Direct Selection tool (A); click or Shift-click anchor points and segments; drag anchor point, handle or segment. Returns one result; editable structure: vector paths. “Drag the anchor point, handle, or path segment to change the shape of your object.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/illustrator/using/tool-techniques/direct-selection-tool.html\nhttps://helpx.adobe.com/illustrator/using/adjust-path-segments.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8bda1e041651", "Operation IDs": "172", "Citation": "Adobe: Direct Selection", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/illustrator/using/tool-techniques/direct-selection-tool.html" }, { "Source ID": "collected-eb0cb723d272", "Operation IDs": "172", "Citation": "Adobe: Adjust path segments", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/illustrator/using/adjust-path-segments.html" } ] }, { "ID": "173", "Software": "Adobe Illustrator", "Operation": "Text to Vector Graphic with Firefly", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4533563b32cf\ncollected-398e6e5a44d1\ncollected-0de8ab64f386\ncollected-e09233b1e979\ncollected-41167a449959\ncollected-5b2cece56ddb\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Firefly Text to Vector Graphic synthesizes vector shapes and their visual relationships.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Adobe (Sep 10, 2026): Text to Vector Graphic also runs on partner models GPT Image, GPT Image 1.5, Gemini 3 (Nano Banana Pro), Gemini 3.1 (Nano Banana 2) and Imagen 4. Partner models generate raster images that Illustrator converts to vectors automatically; partner runs return one variation. Auto Select recommends models from the prompt and returns up to 4 variations. Under the series rule, learned determination survives the conventional vector conversion; X stays 2. Adobe, Partner models in Illustrator, Sep 10, 2026\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Firefly vector model, or a partner model). Designer controls: prompt; content type (Scene, Subject, Icon, Pattern); detail slider; style reference (Artwork or picker); model choice (Auto Select or partner model). Returns several to choose from (up to 4 with Auto Select; 1 with a partner model); editable structure: vector artwork placed as a Generative Object. “If you use Auto Select, you can generate up to four variations, whereas using a partner model generates only one.” Source Quote taken from a search result for the vendor page.\nEarlier X comparison finding, 4 October 2026. Both coding passes mark five visual aspects as mixed. Zero exclusively model-labelled aspects therefore does not mean zero model decisions: controls constrain the result while the trained model still resolves visual details. The reviewed 2023 source supports the Firefly vector-model lineage, not every later partner route. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://helpx.adobe.com/illustrator/desktop/use-generative-ai/generate-scenes-subjects-and-icons.html\nhttps://research.adobe.com/news/adobe-research-at-max-2023/\nhttps://helpx.adobe.com/illustrator/desktop/use-generative-ai/partner-models-overview.html\nhttps://web.archive.org/web/20260825113735id_/https://helpx.adobe.com/illustrator/using/saving-artwork.html\nhttps://web.archive.org/web/20260517174156id_/https://helpx.adobe.com/illustrator/desktop/use-generative-ai/generate-scenes-subjects-and-icons.html\nhttps://web.archive.org/web/20260905090713id_/https://helpx.adobe.com/illustrator/desktop/use-generative-ai/scenarios-with-linked-variations.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4533563b32cf", "Operation IDs": "173", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/illustrator/desktop/use-generative-ai/generate-scenes-subjects-and-icons.html" }, { "Source ID": "collected-398e6e5a44d1", "Operation IDs": "173", "Citation": "research.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.adobe.com/news/adobe-research-at-max-2023/" }, { "Source ID": "collected-0de8ab64f386", "Operation IDs": "173", "Citation": "Adobe, Partner models in Illustrator, Sep 10, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/illustrator/desktop/use-generative-ai/partner-models-overview.html" }, { "Source ID": "collected-e09233b1e979", "Operation IDs": "173", "Citation": "web.archive.org — saving-artwork.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260825113735id_/https://helpx.adobe.com/illustrator/using/saving-artwork.html" }, { "Source ID": "collected-41167a449959", "Operation IDs": "173", "Citation": "web.archive.org — generate-scenes-subjects-and-icons.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260517174156id_/https://helpx.adobe.com/illustrator/desktop/use-generative-ai/generate-scenes-subjects-and-icons.html" }, { "Source ID": "collected-5b2cece56ddb", "Operation IDs": "173", "Citation": "web.archive.org — scenarios-with-linked-variations.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260905090713id_/https://helpx.adobe.com/illustrator/desktop/use-generative-ai/scenarios-with-linked-variations.html" } ] }, { "ID": "174", "Software": "Instant-NGP", "Operation": "NeRF Field Training with Crop Controls", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ec1af95b75ff\ncollected-5947ab1f4d82\ncollected-f29757d651d7", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Instant-NGP fits a neural radiance field to posed images and retains the active scene representation.", "Previous record: rationale": "Scene-specific fitting needs a representation-consistent calibration against explicit Gaussian fitting. Neural parameterization alone must not force a different branch for an equivalent fitted scene function.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). The NVlabs instant-ngp repository describes per-scene training of an MLP with multiresolution hash input encoding from posed images, with Crop size and Crop aabb controls; the latest release is version 2.0 (8 July 2024). The mechanism is documented; the open item is the scale calibration of scene-specific fitting against explicit Gaussian fitting, which vendor or project documentation does not settle. Not established: Which reference comparison applies to a scene-specific fitted function with no pretrained prior; this is a calibration decision for the scale, with no missing documentation fact. github.com • github.com, 2024-07-08 (v2.0)", "Collected source links": "https://github.com/NVlabs/instant-ngp\nhttps://www.matthewtancik.com/nerf\nhttps://github.com/NVlabs/instant-ngp/releases", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ec1af95b75ff", "Operation IDs": "174", "Citation": "GitHub · instant-ngp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/NVlabs/instant-ngp" }, { "Source ID": "collected-5947ab1f4d82", "Operation IDs": "174", "Citation": "Original NeRF project and representation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.matthewtancik.com/nerf" }, { "Source ID": "collected-f29757d651d7", "Operation IDs": "174", "Citation": "github.com, 2024-07-08 (v2.0)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/NVlabs/instant-ngp/releases" } ] }, { "ID": "175", "Software": "Invoke", "Operation": "FLUX.2 Klein regional generation on Canvas", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.75–3.5", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-bfac0c7a7e8b\ncollected-92022b258211\ncollected-71de6401df9e\ncollected-906c7dd9db30", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Invoke Canvas invokes FLUX.2 Klein for regional image generation.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 3 model-decided, 1 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.75–3.5; no midpoint is used.", "Previous record: calculation": "K=1; M=3; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 1 ÷ (5 − 1) = 2.75. Upper: 2 + 3 × 2 ÷ (5 − 1) = 3.5. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (FLUX.2 Klein). Designer controls: inpaint mask (brush/rectangle); generation bounding box; prompt; reference images; model choice; steps; CFG. Returns several to choose from (staged results; count not documented); editable structure: accepted result becomes a raster layer on the canvas. “The generation bounding box is the rectangle that defines where a Canvas generation takes place.” Source • Screenshot", "Collected source links": "https://support.invoke.ai/support/solutions/articles/151000096682-control-canvas\nhttps://github.com/invoke-ai/InvokeAI\nhttps://invoke.ai/features/canvas/bounding-box/\nhttps://github.com/user-attachments/assets/54e23f2a-4d66-4db7-b0fd-f34eb3d107b0", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-bfac0c7a7e8b", "Operation IDs": "175", "Citation": "Invoke · Control Canvas", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.invoke.ai/support/solutions/articles/151000096682-control-canvas" }, { "Source ID": "collected-92022b258211", "Operation IDs": "175", "Citation": "Invoke current supported model families", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/invoke-ai/InvokeAI" }, { "Source ID": "collected-71de6401df9e", "Operation IDs": "175", "Citation": "invoke.ai — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://invoke.ai/features/canvas/bounding-box/" }, { "Source ID": "collected-906c7dd9db30", "Operation IDs": "175", "Citation": "github.com — 54e23f2a-4d66-4db7-b0fd-f34eb3d107b0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/user-attachments/assets/54e23f2a-4d66-4db7-b0fd-f34eb3d107b0" } ] }, { "ID": "176", "Software": "Jitter", "Operation": "Oli Brainstorm editable animation generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-cb97c7ea6f1c\ncollected-e83b24403ba4\ncollected-0175b3100edb", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Jitter Oli Brainstorm selects animation instructions and produces editable timeline operations.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The documented learned contribution selects effects and assigns order and timing for layers supplied by the static design; layer content comes from that design. That places the comparison nearer region determination than synthesis, at 6–10 within the same envelope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. Jitter's Brainstorm article (13 May 2026) states that the AI model assigns effects and sequencing order from layer titles and a chosen mood, generating timings and transitions for the layers of a static design, with existing animations disregarded. The Superagents article (25 September 2026) lists Brainstorm as a quick action. The model behind Oli and Superagents is not named. Not established: The effect catalogue Brainstorm draws on; whether layout positions also change; the model. Previous reason: Learned motion planning survives conventional animation execution. The source supports 2–10 while leaving the precise planning model undisclosed. help.jitter.video, 2026-05-13 • help.jitter.video, 2026-09-25", "Collected source links": "https://jitter.video/\nhttps://help.jitter.video/en/articles/12932981-animate-with-ai-brainstorm\nhttps://help.jitter.video/en/articles/15656969-animate-with-ai-superagents", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-cb97c7ea6f1c", "Operation IDs": "176", "Citation": "Jitter · Product overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://jitter.video/" }, { "Source ID": "collected-e83b24403ba4", "Operation IDs": "176", "Citation": "Jitter Brainstorm model and editable results", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.jitter.video/en/articles/12932981-animate-with-ai-brainstorm" }, { "Source ID": "collected-0175b3100edb", "Operation IDs": "176", "Citation": "help.jitter.video, 2026-09-25", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.jitter.video/en/articles/15656969-animate-with-ai-superagents" } ] }, { "ID": "177", "Software": "Jitter", "Operation": "AI image to video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1bd108f947f3\ncollected-fc596553eb37\ncollected-4ad96295b261\ncollected-e002a9d4f07e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Jitter image-to-video replaces an image layer with a generated clip.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 3 unknown and 1 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=3; conditional=1; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Jitter refers to it as \"our AI model\" and names no provider or version. Model choice: No. Not established: Which video model generates the clip (in-house or a third-party model such as Veo, Kling, Seedance or Hailuo). help.jitter.video, page date not shown • jitter.video, 2026-02-10\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (image-to-video through assistant Oli). Designer controls: selected image layer; prompt (auto-written, editable); generate/regenerate. Returns one result; a flat video: replaces the image layer in the file. “our AI assistant Oli will generate a video clip for your and replace the image with the newly created video” Source • Screenshot", "Collected source links": "https://help.jitter.video/en/articles/13671056-transform-images-into-video-with-ai\nhttps://jitter.video/changelog/2026-02-10-image-to-video/\nhttps://help.jitter.video/en/articles/13671056-ai-image-video\nhttps://downloads.intercomcdn.com/i/o/l8r66h6o/2048653872/1102a707421ebc82585e76d3a0bb/itv_16_9.gif", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1bd108f947f3", "Operation IDs": "177", "Citation": "Jitter image-to-video output limits", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.jitter.video/en/articles/13671056-transform-images-into-video-with-ai" }, { "Source ID": "collected-fc596553eb37", "Operation IDs": "177", "Citation": "Jitter: Image to Video launch, 02102026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://jitter.video/changelog/2026-02-10-image-to-video/" }, { "Source ID": "collected-4ad96295b261", "Operation IDs": "177", "Citation": "help.jitter.video, page date not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.jitter.video/en/articles/13671056-ai-image-video" }, { "Source ID": "collected-e002a9d4f07e", "Operation IDs": "177", "Citation": "downloads.intercomcdn.com — itv_16_9.gif", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://downloads.intercomcdn.com/i/o/l8r66h6o/2048653872/1102a707421ebc82585e76d3a0bb/itv_16_9.gif" } ] }, { "ID": "178", "Software": "Jitter", "Operation": "Manual timing and motion revision of generated visuals", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1059002330d0\ncollected-2d448117afaf\ncollected-5e92ca1be52a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Jitter's manual editor changes timing, easing and motion segments of existing visuals.", "Previous record: rationale": "The specified animation rules execute conventionally; upstream Brainstorm or image generation is outside this edit.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: drag animation segment to retime; drag segment edges for duration; arrow-key nudges; easing; group; property panel. Returns one result; editable structure: timeline animation segments on layers. “Drag the edge to shorten or extend the animation.” Source • Screenshot", "Collected source links": "https://help.jitter.video/en/articles/14111550-make-your-first-animation\nhttps://help.jitter.video/en/articles/14111802-work-with-the-timeline\nhttps://downloads.intercomcdn.com/i/o/l8r66h6o/2175697975/c48923dcd2ef55623832b01d6154/CleanShot+2026-03-11+at+14_44_51.gif", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1059002330d0", "Operation IDs": "178", "Citation": "Jitter custom animation controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.jitter.video/en/articles/14111550-make-your-first-animation" }, { "Source ID": "collected-2d448117afaf", "Operation IDs": "178", "Citation": "Jitter timeline sequencing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.jitter.video/en/articles/14111802-work-with-the-timeline" }, { "Source ID": "collected-5e92ca1be52a", "Operation IDs": "178", "Citation": "downloads.intercomcdn.com — CleanShot+2026-03-11+at+14_44_51.gif", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://downloads.intercomcdn.com/i/o/l8r66h6o/2175697975/c48923dcd2ef55623832b01d6154/CleanShot+2026-03-11+at+14_44_51.gif" } ] }, { "ID": "179", "Software": "Kaedim", "Operation": "Image Reconstruction and Artist Mesh Refinement", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-72a012470480\ncollected-7beb75653834\ncollected-90eec17a7908\ncollected-6ef5984aca15\ncollected-1dfd1c33403d\ncollected-9758fe7f2c79\ncollected-a1b822490e6c\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Kaedim's documented service generates mesh content and then applies artist refinement.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=2; conditional=1; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Kaedim's own machine learning for 2D-to-3D reconstruction, unnamed, followed by in-house artists who check and clean every output; artist edits feed back into model training. No model name, version or third-party base is published. Model choice: No model choice. Not established: Architecture and name of the reconstruction model; whether it uses an open image-to-3D model; how much of the delivered mesh is model output versus artist work per quality tier. docs.kaedim3d.com, page date not shown • docs.kaedim3d.com, page date not shown • docs.kaedim3d.com\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (2D-to-3D reconstruction), then artist cleanup). Designer controls: input images (up to 10); markups on input images; art and tech specifications; dimensions; LoD0 target tris; topology (quads/tris); texture style; texture resolution; rig options; revision requests. Returns count not documented; editable structure: mesh; OBJ, FBX, USD, GLB, GLTF. “Our art team verifies and cleans all outputs, making any necessary edits to ensure that you always receive a production-ready model.” Source • Screenshot", "Collected source links": "https://docs.kaedim3d.com/welcome/how-kaedim-works\nhttps://www.kaedim3d.com/faq\nhttps://docs.kaedim3d.com/welcome/faq\nhttps://docs.kaedim3d.com/dev-logs/releases.md\nhttps://docs.kaedim3d.com/reviewing-assets/request-revisions/edit-asset-settings\nhttps://docs.kaedim3d.com/reviewing-assets/request-revisions/request-revisions\nhttps://docs.kaedim3d.com/reviewing-assets/branching", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-72a012470480", "Operation IDs": "179", "Citation": "Kaedim · Image-to-3D", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.kaedim3d.com/welcome/how-kaedim-works" }, { "Source ID": "collected-7beb75653834", "Operation IDs": "179", "Citation": "Kaedim — current enterprise FAQ and artist assurance", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.kaedim3d.com/faq" }, { "Source ID": "collected-90eec17a7908", "Operation IDs": "179", "Citation": "docs.kaedim3d.com, page date not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.kaedim3d.com/welcome/faq" }, { "Source ID": "collected-6ef5984aca15", "Operation IDs": "179", "Citation": "docs.kaedim3d.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.kaedim3d.com/dev-logs/releases.md" }, { "Source ID": "collected-1dfd1c33403d", "Operation IDs": "179", "Citation": "docs.kaedim3d.com — edit-asset-settings", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.kaedim3d.com/reviewing-assets/request-revisions/edit-asset-settings" }, { "Source ID": "collected-9758fe7f2c79", "Operation IDs": "179", "Citation": "docs.kaedim3d.com — request-revisions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.kaedim3d.com/reviewing-assets/request-revisions/request-revisions" }, { "Source ID": "collected-a1b822490e6c", "Operation IDs": "179", "Citation": "docs.kaedim3d.com — branching", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.kaedim3d.com/reviewing-assets/branching" } ] }, { "ID": "180", "Software": "KeyShot Studio", "Operation": "KeyShot AI Denoise", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2b30ec2d3de4\ncollected-fb66f025f5ae\ncollected-304a86f54728", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KeyShot AI Denoise predicts clean image information from a partially converged render.", "Previous record: rationale": "Trained render restoration matches fidelity restoration. A blend control does not itself calculate a different X.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). KeyShot's 2026.1 AI Shots post states that denoising, in particular progressive denoising, relies on custom trained AI models. The KeyShot Studio 2025 manual states that from 2024.3 progressive denoising runs on NVIDIA RTX 20-series or newer GPUs and that older GPUs fall back to the earlier denoiser; AI Shots is a separate generative feature outside this scope. Not established: The network used on non-RTX systems, and training data and loss for the custom models, are not published. keyshot.com, KeyShot 2026.1 • manuals.keyshot.com, KeyShot Studio 2025 manual • manual.keyshot.com, accessed 2026-10-03", "Collected source links": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/\nhttps://manual.keyshot.com/manual/image/image-styles/basic-image-style/\nhttps://manuals.keyshot.com/kss2025/en-us/manual/photographic-image-style_denoise.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2b30ec2d3de4", "Operation IDs": "180, 181, 182, 184, 185", "Citation": "KeyShot · Learned denoising and AI modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/" }, { "Source ID": "collected-fb66f025f5ae", "Operation IDs": "180", "Citation": "KeyShot · Denoise toggle and blend", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manual.keyshot.com/manual/image/image-styles/basic-image-style/" }, { "Source ID": "collected-304a86f54728", "Operation IDs": "180", "Citation": "manuals.keyshot.com, KeyShot Studio 2025 manual", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manuals.keyshot.com/kss2025/en-us/manual/photographic-image-style_denoise.html" } ] }, { "ID": "181", "Software": "KeyShot Studio", "Operation": "KeyShot Background", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2b30ec2d3de4\ncollected-fbbe77d663f2\ncollected-dc005522c18f\ncollected-c36ce987313d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KeyShot AI Shots Background generates surroundings around a rendered product.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=5; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: KeyShot 2026.1 (March 2026): custom-trained Qwen Image Editing model (Qwen-Image-Edit family, Alibaba Qwen team), run locally. KeyShot 2025.2 (July 2025): FLUX.1 [schnell] (Black Forest Labs), run locally. Model choice: No. Not established: KeyShot does not state which Qwen-Image-Edit release (original August 2025, 2509 or 2511) the custom model is fine-tuned from. keyshot.com • keyshot.com • cgchannel.com, 2026-03-27\nChecked 3 Oct 2026 KeyShot (2026.1): Restyle, Background and Imagine are modes of AI Shots, which runs locally. KeyShot names its model: \"The core AI model in AI Shots is the Qwen Image Editing model, which has been further custom trained using in-house generated images.\" Number of Images is user-set. Restyle and Background keep the CAD geometry; Imagine does not. X stays 2. KeyShot, AI Shots in 2026.1 • KeyShot manual 2026, AI Shots\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (generates a setting around the KeyShot render). Designer controls: prompt; resolution; seed/randomize; number of images. Returns several to choose from (user sets count; maximum not documented); a flat image. “Background takes the rendered CAD model in KeyShot and generates a setting that integrates the CAD model as it looks in KeyShot.” Source\nEarlier X comparison finding, 4 October 2026. Background retains the rendered product’s geometry, material appearance and lighting while generating the setting. Restyle (185) also delegates product-material appearance. This component-specific distinction survives uncertainty about exact pixel preservation; an aggregate mixed-label count can reverse or hide it. No numerical distance is established. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/\nhttps://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/\nhttps://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/\nhttps://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2b30ec2d3de4", "Operation IDs": "180, 181, 182, 184, 185", "Citation": "KeyShot · Learned denoising and AI modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/" }, { "Source ID": "collected-fbbe77d663f2", "Operation IDs": "181, 182, 184, 185", "Citation": "www.keyshot.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/" }, { "Source ID": "collected-dc005522c18f", "Operation IDs": "181, 182, 184, 185", "Citation": "cgchannel.com, 2026-03-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/" }, { "Source ID": "collected-c36ce987313d", "Operation IDs": "181, 182, 184, 185", "Citation": "KeyShot manual 2026, AI Shots", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm" } ] }, { "ID": "182", "Software": "KeyShot Studio", "Operation": "KeyShot Imagine", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2b30ec2d3de4\ncollected-fbbe77d663f2\ncollected-dc005522c18f\ncollected-c36ce987313d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KeyShot AI Shots Imagine generates an image without CAD-driven scene construction.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: KeyShot 2026.1 (March 2026): custom-trained Qwen Image Editing model (Qwen-Image-Edit family, Alibaba Qwen team), run locally. KeyShot 2025.2 (July 2025): FLUX.1 [schnell] (Black Forest Labs), run locally. Model choice: No. Not established: KeyShot does not state which Qwen-Image-Edit release (original August 2025, 2509 or 2511) the custom model is fine-tuned from. keyshot.com • keyshot.com • cgchannel.com, 2026-03-27\nChecked 3 Oct 2026 KeyShot (2026.1): Restyle, Background and Imagine are modes of AI Shots, which runs locally. KeyShot names its model: \"The core AI model in AI Shots is the Qwen Image Editing model, which has been further custom trained using in-house generated images.\" Number of Images is user-set. Restyle and Background keep the CAD geometry; Imagine does not. X stays 2. KeyShot, AI Shots in 2026.1 • KeyShot manual 2026, AI Shots\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (pure generation, ignores CAD data). Designer controls: prompt; resolution; seed/randomize; number of images. Returns several to choose from (user sets count; maximum not documented); a flat image. “Imagine is a pure generative AI model that ignores the CAD data and simply creates an image based on the prompt.” Source\nEarlier X comparison finding, 4 October 2026. Imagine ignores the CAD scene, whereas Restyle (185) uses CAD geometry to guide the product. The documented modes differ in the scope of visual decisions delegated to the model. This is support for a comparison, not a measured distance or a claim about computational work. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/\nhttps://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/\nhttps://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/\nhttps://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2b30ec2d3de4", "Operation IDs": "180, 181, 182, 184, 185", "Citation": "KeyShot · Learned denoising and AI modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/" }, { "Source ID": "collected-fbbe77d663f2", "Operation IDs": "181, 182, 184, 185", "Citation": "www.keyshot.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/" }, { "Source ID": "collected-dc005522c18f", "Operation IDs": "181, 182, 184, 185", "Citation": "cgchannel.com, 2026-03-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/" }, { "Source ID": "collected-c36ce987313d", "Operation IDs": "181, 182, 184, 185", "Citation": "KeyShot manual 2026, AI Shots", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm" } ] }, { "ID": "183", "Software": "KeyShot Studio", "Operation": "KeyShot Real-Time Rendering", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-fbc87cee40a1\ncollected-7f9dda4a60ce", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KeyShot evaluates conventional ray tracing with learned denoising and generative modes disabled.", "Previous record: rationale": "Physical rendering and procedural material evaluation contain no trained predictor in this scope.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a solver or simulation. Designer controls: materials (drag-and-drop, material graph); physical lights; HDRI editor; light layers; cameras; CPU or GPU rendering; caustics; global illumination. Returns continuous live output; a flat image. “Every change updates instantly in the viewport. What you see is what you get in the final render.” Source", "Collected source links": "https://www.keyshot.com/keyshot-studio/\nhttps://manual.keyshot.com/manual/materials/material-graph/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-fbc87cee40a1", "Operation IDs": "183", "Citation": "KeyShot · Scene rendering and editing controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/keyshot-studio/" }, { "Source ID": "collected-7f9dda4a60ce", "Operation IDs": "183", "Citation": "KeyShot · Editable material graph", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manual.keyshot.com/manual/materials/material-graph/" } ] }, { "ID": "184", "Software": "KeyShot Studio", "Operation": "KeyShot Replace", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2b30ec2d3de4\ncollected-fbbe77d663f2\ncollected-dc005522c18f\ncollected-c36ce987313d\ncollected-0a0fe734f1ad\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KeyShot AI Shots Replace synthesizes content within a selected image region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 3 unknown and 1 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=3; conditional=1; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Custom-trained Qwen Image Editing model (Qwen-Image-Edit family, Alibaba Qwen team), run locally, introduced in KeyShot 2026.1 (March 2026). Replace is an editing mode new in 2026.1, so no FLUX period applies. Model choice: No. Not established: KeyShot does not state which Qwen-Image-Edit release (original August 2025, 2509 or 2511) the custom model is fine-tuned from. keyshot.com • keyshot.com • cgchannel.com, 2026-03-27\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (localized edit of a user-highlighted region). Designer controls: region highlight; prompt. Returns count not documented; a flat image. “Replace is similar to Transform except the change is made to a specific region highlighted by the user.” Source", "Collected source links": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/\nhttps://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/\nhttps://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/\nhttps://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm\nhttps://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2025-3/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2b30ec2d3de4", "Operation IDs": "180, 181, 182, 184, 185", "Citation": "KeyShot · Learned denoising and AI modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/" }, { "Source ID": "collected-fbbe77d663f2", "Operation IDs": "181, 182, 184, 185", "Citation": "www.keyshot.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/" }, { "Source ID": "collected-dc005522c18f", "Operation IDs": "181, 182, 184, 185", "Citation": "cgchannel.com, 2026-03-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/" }, { "Source ID": "collected-c36ce987313d", "Operation IDs": "181, 182, 184, 185", "Citation": "KeyShot manual 2026, AI Shots", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm" }, { "Source ID": "collected-0a0fe734f1ad", "Operation IDs": "184", "Citation": "www.keyshot.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2025-3/" } ] }, { "ID": "185", "Software": "KeyShot Studio", "Operation": "KeyShot Restyle", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "4.25–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2b30ec2d3de4\ncollected-fbbe77d663f2\ncollected-dc005522c18f\ncollected-c36ce987313d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KeyShot AI Shots Restyle uses its CAD-informed trained image model to realize materials, lighting and surroundings.", "Previous record: rationale": "Working v5 documentation coding: 3 kept, 1 model-decided, 1 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 4.25–5.0; no midpoint is used.", "Previous record: calculation": "K=3; M=1; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 3 ÷ (5 − 1) = 4.25. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: KeyShot 2026.1 (March 2026): custom-trained Qwen Image Editing model (Qwen-Image-Edit family, Alibaba Qwen team), run locally. KeyShot 2025.2 (July 2025): FLUX.1 [schnell] (Black Forest Labs), run locally. Model choice: No. Not established: KeyShot does not state which Qwen-Image-Edit release (original August 2025, 2509 or 2511) the custom model is fine-tuned from. keyshot.com • keyshot.com • cgchannel.com, 2026-03-27\nChecked 3 Oct 2026 KeyShot (2026.1): Restyle, Background and Imagine are modes of AI Shots, which runs locally. KeyShot names its model: \"The core AI model in AI Shots is the Qwen Image Editing model, which has been further custom trained using in-house generated images.\" Number of Images is user-set. Restyle and Background keep the CAD geometry; Imagine does not. X stays 2. KeyShot, AI Shots in 2026.1 • KeyShot manual 2026, AI Shots\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (fine-tuned to take CAD 3D data as input). Designer controls: prompt; resolution; seed/randomize; number of images. Returns several to choose from (user sets count; maximum not documented); a flat image. “Restyle ignores the existing materials on the CAD model and with a prompt the user can describe what the CAD model should look like.” Source\nEarlier X comparison finding, 4 October 2026. Restyle uses CAD geometry to guide the depicted product while resolving material appearance and the setting, including lighting in the current documentation. Background (181) retains the rendered product’s appearance; Imagine (182) ignores CAD. The mode distinctions support a proposed ordering by delegated visual decisions, but the aspect-count test can reverse Restyle versus Background. Stored X was unchanged at that earlier review. Full comparison and sources.", "Collected source links": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/\nhttps://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/\nhttps://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/\nhttps://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2b30ec2d3de4", "Operation IDs": "180, 181, 182, 184, 185", "Citation": "KeyShot · Learned denoising and AI modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/ai-shots-2026-1-introduction/" }, { "Source ID": "collected-fbbe77d663f2", "Operation IDs": "181, 182, 184, 185", "Citation": "www.keyshot.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.keyshot.com/blog/keyshot-studio-performance-updates-version-2026-1/" }, { "Source ID": "collected-dc005522c18f", "Operation IDs": "181, 182, 184, 185", "Citation": "cgchannel.com, 2026-03-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/03/keyshot-releases-keyshot-studio-2026-1/" }, { "Source ID": "collected-c36ce987313d", "Operation IDs": "181, 182, 184, 185", "Citation": "KeyShot manual 2026, AI Shots", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manuals.keyshot.com/kss2026/en-us/manual/AI-Shots.htm" } ] }, { "ID": "186", "Software": "KIRI Engine", "Operation": "3DGS to Mesh Neural Surface Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2b5e141f925f\ncollected-109fb1f28152\ncollected-c4390ffb0f96\ncollected-3f55297cb8e0", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KIRI converts captured 3DGS material to a mesh through its documented neural surface-reconstruction route.", "Previous record: rationale": "The retained surface fit is established, but its training/prior relationship and comparison with explicit scene fitting are insufficiently resolved for a defensible coordinate.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). KIRI's blog (19 May 2025) states 3DGS to Mesh uses the same underlying algorithm as Featureless Mode and combines Neural Surface Reconstruction with improved normal prediction and reflection handling; KIRI's explainer (21 Jan 2025) states NSR employs a neural network to create a volumetric representation that is then reconstructed into a 3D model. Neither page states whether the network is fitted per capture or carries pretrained weights, nor what performs normal prediction. Not established: Per-capture fitting versus pretrained prior in KIRI's NSR; whether normal prediction uses a pretrained estimator; the scale calibration of scene fitting against explicit Gaussian fitting. kiriengine.app, 2025-05-19 • kiriengine.app, 2025-01-21 • docs.kiriengine.app", "Collected source links": "https://www.kiriengine.app/tutorial/introduction-to-3dgs-scan\nhttps://www.kiriengine.app/blog/from-2d-video-to-3d-magic\nhttps://www.kiriengine.app/blog/explained/featureless-object-scan\nhttps://docs.kiriengine.app/3dgs-scan/image-upload/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2b5e141f925f", "Operation IDs": "186", "Citation": "KIRI Engine · KIRI 3DGS Scan", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.kiriengine.app/tutorial/introduction-to-3dgs-scan" }, { "Source ID": "collected-109fb1f28152", "Operation IDs": "186, 187", "Citation": "KIRI · Scan mechanisms and neural mesh conversion", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.kiriengine.app/blog/from-2d-video-to-3d-magic" }, { "Source ID": "collected-c4390ffb0f96", "Operation IDs": "186, 187", "Citation": "kiriengine.app, 2025-01-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.kiriengine.app/blog/explained/featureless-object-scan" }, { "Source ID": "collected-3f55297cb8e0", "Operation IDs": "186", "Citation": "docs.kiriengine.app", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.kiriengine.app/3dgs-scan/image-upload/" } ] }, { "ID": "187", "Software": "KIRI Engine", "Operation": "Featureless Neural Surface Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-109fb1f28152\ncollected-c4390ffb0f96", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "KIRI Featureless mode reconstructs a surface from capture through its neural surface-reconstruction method.", "Previous record: rationale": "This capture-fitting route must be distinguished from pretrained single-image asset generation. A consistent calibration of scene fitting is still required before a point is issued.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). KIRI's explainer (21 Jan 2025) states NSR employs a neural network to create a volumetric representation of the object and uses it to anticipate the object's appearance from various viewpoints before surface-mesh reconstruction; the May 2025 blog says Featureless mode uses shape cues and smart inference for smooth objects. No KIRI page states whether the network is optimized per capture or uses pretrained weights. Not established: Per-capture fitting versus pretrained prior in Featureless mode; the scale calibration of scene fitting against explicit Gaussian fitting and against pretrained single-image asset generation. kiriengine.app, 2025-01-21 • kiriengine.app, 2025-05-19", "Collected source links": "https://www.kiriengine.app/blog/from-2d-video-to-3d-magic\nhttps://www.kiriengine.app/blog/explained/featureless-object-scan", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-109fb1f28152", "Operation IDs": "186, 187", "Citation": "KIRI · Scan mechanisms and neural mesh conversion", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.kiriengine.app/blog/from-2d-video-to-3d-magic" }, { "Source ID": "collected-c4390ffb0f96", "Operation IDs": "186, 187", "Citation": "kiriengine.app, 2025-01-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.kiriengine.app/blog/explained/featureless-object-scan" } ] }, { "ID": "188", "Software": "Kling AI (Kuaishou; hosted routes recorded)", "Operation": "Image 3.0 text-to-image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-effdb7b60b67\ncollected-f09e219ae895\ncollected-057ccd50240b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Kling IMAGE 3.0 generates a still image through the selected text-to-image route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Kling IMAGE 3.0). Designer controls: prompt; model choice; reference images (up to 10); format and output settings; resolution (2K/4K). Returns count not documented; a flat image. “Natural-language image creation, reference images, style transfer, portrait and character reference, multi-image blending, and local re-editing in a flexible image workflow.” Source", "Collected source links": "https://ir.kuaishou.com/news-releases/news-release-details/kling-ai-launches-30-model-ushering-era-where-everyone-can-be\nhttps://kling.ai/quickstart/klingai-image-3-model-user-guide\nhttps://kling.ai/blog/kling-image-3-vs-omni", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-effdb7b60b67", "Operation IDs": "132, 188", "Citation": "Kuaishou Kling 3.0 model family", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://ir.kuaishou.com/news-releases/news-release-details/kling-ai-launches-30-model-ushering-era-where-everyone-can-be" }, { "Source ID": "collected-f09e219ae895", "Operation IDs": "188", "Citation": "Kling IMAGE 3.0 Model User Guide, 02042026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://kling.ai/quickstart/klingai-image-3-model-user-guide" }, { "Source ID": "collected-057ccd50240b", "Operation IDs": "188", "Citation": "kling.ai — kling-image-3-vs-omni", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://kling.ai/blog/kling-image-3-vs-omni" } ] }, { "ID": "189", "Software": "Krea", "Operation": "Realtime sketch-conditioned image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-56381971e5a2\ncollected-5307571f4250", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Krea Realtime Draw transforms the evolving canvas through learned image generation.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 1 unknown and 2 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=1; conditional=2; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (live image model applied to the canvas). Designer controls: brush (color, size, opacity); shapes; eraser; upload image; background color; prompt; AI strength; model choice (8 options); resolution; aspect ratio; seed; output type (image/video). Returns continuous live output; a flat image: download or upscale; video output option. “The left-hand canvas is your input surface. Everything you place or draw there is interpreted by the AI model.” Source • Screenshot", "Collected source links": "https://judyailab.com/en/posts/krea-ai-realtime-creative-studio/\nhttps://www.krea.ai/docs/user-guide/features/realtime", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-56381971e5a2", "Operation IDs": "189", "Citation": "Judy AI Lab · Realtime Canvas", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://judyailab.com/en/posts/krea-ai-realtime-creative-studio/" }, { "Source ID": "collected-5307571f4250", "Operation IDs": "189, 190", "Citation": "Krea Realtime modes and separate input/output panels", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/docs/user-guide/features/realtime" } ] }, { "ID": "190", "Software": "Krea", "Operation": "Realtime live-feed visual restyling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–4.4", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7da2e28f6d3f\ncollected-5307571f4250\ncollected-618413b77769\ncollected-fd61f04ab37e\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Krea Realtime Edit transforms a live visual feed into generated imagery.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–4.4; no midpoint is used.", "Previous record: calculation": "K=1; M=2; unknown=3; conditional=0; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 4 ÷ (6 − 1) = 4.4. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (live image model applied to incoming frames). Designer controls: live source (webcam or screen share); prompt; AI strength; model choice; resolution; aspect ratio; seed; output type (image/video). Returns continuous live output; a flat image: downloadable as image or video clip. “Web Camera replaces the canvas with a live feed from your webcam. The model interprets the camera stream in real time.” Source • Screenshot", "Collected source links": "https://www.krea.ai/blog/realtime-edit\nhttps://www.krea.ai/docs/user-guide/features/realtime\nhttps://www.krea.ai/docs/user-guide/features/realtime.md\nhttps://docs.krea.ai/user-guide/features/realtime", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7da2e28f6d3f", "Operation IDs": "190", "Citation": "Krea · Realtime Edit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/blog/realtime-edit" }, { "Source ID": "collected-5307571f4250", "Operation IDs": "189, 190", "Citation": "Krea Realtime modes and separate input/output panels", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/docs/user-guide/features/realtime" }, { "Source ID": "collected-618413b77769", "Operation IDs": "190", "Citation": "www.krea.ai — realtime.md", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/docs/user-guide/features/realtime.md" }, { "Source ID": "collected-fd61f04ab37e", "Operation IDs": "190", "Citation": "docs.krea.ai — realtime", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.krea.ai/user-guide/features/realtime" } ] }, { "ID": "191", "Software": "Krita", "Operation": "Manual paintover of generated imagery on separate layers", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b748d07921bb\ncollected-eec30b6339bc\ncollected-ad0f88a4b12b\ncollected-8323541de517\ncollected-00202a19a437\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Krita's conventional brushes paint on separate layers over an existing image.", "Previous record: rationale": "The scoped brush and layer operations contain no trained generator. Earlier image generation is not recounted in the paintover command.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: brush presets; brush engines and engine settings; brush stabilizer; layer stack (add, order, flatten); paint, vector, filter, group and file layers; drawing assistants. Returns one result; editable structure: layer. “The layer stack is a way for you to separate elements of a drawing and manipulate the drawing order” Source • Screenshot", "Collected source links": "https://krita.org/en/features/\nhttps://docs.krita.org/en/user_manual/getting_started/basic_concepts.html\nhttps://krita.org/images/pages/application-screenshot.webp\nhttps://docs.krita.org/en/reference_manual/layers_and_masks/paint_layers.html\nhttps://docs.krita.org/en/reference_manual/dockers/layers.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b748d07921bb", "Operation IDs": "191", "Citation": "Krita · Layers/Brush Engines", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://krita.org/en/features/" }, { "Source ID": "collected-eec30b6339bc", "Operation IDs": "191, 192", "Citation": "Krita painting and layer workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.krita.org/en/user_manual/getting_started/basic_concepts.html" }, { "Source ID": "collected-ad0f88a4b12b", "Operation IDs": "191", "Citation": "krita.org — application-screenshot.webp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://krita.org/images/pages/application-screenshot.webp" }, { "Source ID": "collected-8323541de517", "Operation IDs": "191", "Citation": "docs.krita.org — paint_layers.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.krita.org/en/reference_manual/layers_and_masks/paint_layers.html" }, { "Source ID": "collected-00202a19a437", "Operation IDs": "191", "Citation": "docs.krita.org — layers.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.krita.org/en/reference_manual/dockers/layers.html" } ] }, { "ID": "192", "Software": "Krita with krita-ai-diffusion", "Operation": "FLUX.2 Klein selected-region image editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b30a46853083\ncollected-e301b76f0ae8\ncollected-6467d37e027f\ncollected-6483363336b5\ncollected-bcc354e09f4b\ncollected-eec30b6339bc\ncollected-85668c526cd5\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Krita's AI Diffusion plugin invokes FLUX.2 Klein for selected-region image revision.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (FLUX.2 Klein). Designer controls: selection; prompt; strength; fill mode (None/Neutral/Blur/Border/Inpaint); context area; feather/padding; style (model choice); seed. Returns several to choose from (batch size set by user); editable structure: new layer on top of the layer stack. “Use selections for generative fill, expand, to add or remove objects” Source • Screenshot", "Collected source links": "https://github.com/Acly/krita-ai-diffusion\nhttps://docs.interstice.cloud/base-models/\nhttps://docs.interstice.cloud/selections/\nhttps://github.com/Acly/krita-ai-diffusion/releases/tag/v1.12.0\nhttps://docs.interstice.cloud/edit-models/\nhttps://docs.krita.org/en/user_manual/getting_started/basic_concepts.html\nhttps://docs.interstice.cloud/basics/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b30a46853083", "Operation IDs": "192, 193", "Citation": "GitHub · krita-ai-diffusion", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Acly/krita-ai-diffusion" }, { "Source ID": "collected-e301b76f0ae8", "Operation IDs": "192, 193, 194", "Citation": "Krita plugin supported current base models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/base-models/" }, { "Source ID": "collected-6467d37e027f", "Operation IDs": "192", "Citation": "docs.interstice.cloud — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/selections/" }, { "Source ID": "collected-6483363336b5", "Operation IDs": "192", "Citation": "github.com — v1.12.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Acly/krita-ai-diffusion/releases/tag/v1.12.0" }, { "Source ID": "collected-bcc354e09f4b", "Operation IDs": "192", "Citation": "docs.interstice.cloud — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/edit-models/" }, { "Source ID": "collected-eec30b6339bc", "Operation IDs": "191, 192", "Citation": "Krita painting and layer workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.krita.org/en/user_manual/getting_started/basic_concepts.html" }, { "Source ID": "collected-85668c526cd5", "Operation IDs": "192, 193, 194", "Citation": "docs.interstice.cloud — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/basics/" } ] }, { "ID": "193", "Software": "Krita with krita-ai-diffusion", "Operation": "Live painting generation preview", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b30a46853083\ncollected-e301b76f0ae8\ncollected-7dd8a43ced90\ncollected-e25cb9df8abd\ncollected-85668c526cd5", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Krita AI Diffusion repeatedly generates a preview from the live painted canvas.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=5; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: User-selected diffusion checkpoint (set by the active Style: Stable Diffusion 1.5, SDXL, Illustrious, Flux 2, Z-Image and others supported) run through local or cloud ComfyUI, with the Hyper-SD speed LoRA as the default for Live mode since plugin v1.19.0 (June 2024). LCM and Lightning LoRA presets remain available; Model choice: Yes. Not established: Which checkpoint the default Live style loads in the current release; whether Flux 2 and Z-Image run in Live mode with a speed LoRA. github.com, 2024-06-23 • github.com, page date not shown\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (live diffusion preview of the painted canvas). Designer controls: painted canvas; prompt; strength; seed; resolution; regions. Returns continuous live output; editable structure: layer, on Apply; live mode overwrites layer content by default. “Let AI interpret your canvas in real time for immediate feedback.” Source • Screenshot", "Collected source links": "https://github.com/Acly/krita-ai-diffusion\nhttps://docs.interstice.cloud/base-models/\nhttps://github.com/Acly/krita-ai-diffusion/releases/tag/v1.19.0\nhttps://github.com/Acly/krita-ai-diffusion/blob/main/README.md\nhttps://docs.interstice.cloud/basics/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b30a46853083", "Operation IDs": "192, 193", "Citation": "GitHub · krita-ai-diffusion", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Acly/krita-ai-diffusion" }, { "Source ID": "collected-e301b76f0ae8", "Operation IDs": "192, 193, 194", "Citation": "Krita plugin supported current base models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/base-models/" }, { "Source ID": "collected-7dd8a43ced90", "Operation IDs": "193", "Citation": "github.com, 2024-06-23", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Acly/krita-ai-diffusion/releases/tag/v1.19.0" }, { "Source ID": "collected-e25cb9df8abd", "Operation IDs": "193", "Citation": "github.com, page date not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Acly/krita-ai-diffusion/blob/main/README.md" }, { "Source ID": "collected-85668c526cd5", "Operation IDs": "192, 193, 194", "Citation": "docs.interstice.cloud — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/basics/" } ] }, { "ID": "194", "Software": "Krita with krita-ai-diffusion", "Operation": "FLUX.2 Klein text to image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e301b76f0ae8\ncollected-85668c526cd5", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Krita AI Diffusion invokes FLUX.2 Klein text-to-image generation.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (FLUX.2 Klein). Designer controls: prompt; style (model choice); batch size; resolution; seed. Returns several to choose from (batch size set by user); editable structure: new layer on top of the layer stack. “It is common to generate multiple images for the same input and select the one that fits best.” Source • Screenshot", "Collected source links": "https://docs.interstice.cloud/base-models/\nhttps://docs.interstice.cloud/basics/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e301b76f0ae8", "Operation IDs": "192, 193, 194", "Citation": "Krita plugin supported current base models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/base-models/" }, { "Source ID": "collected-85668c526cd5", "Operation IDs": "192, 193, 194", "Citation": "docs.interstice.cloud — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.interstice.cloud/basics/" } ] }, { "ID": "195", "Software": "Leonardo AI", "Operation": "Realtime Canvas generation preview", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7cfe970efa38\ncollected-75913062213e\ncollected-7d3ddf637750", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Leonardo Realtime Canvas converts a sketch and prompt into a generated image preview.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 2 unknown and 1 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=2; conditional=1; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (sketch-conditioned live generation). Designer controls: brush; eraser; color; uploaded image; prompt; creativity strength; guidance; style presets; fixed seed; High Quality; inpaint mode and inpaint strength; upscale. Returns continuous live output; a flat image: 512x512 standard; download, save to feed, upscale. “As you sketch, Realtime Canvas turns each brushstroke into detail in near real-time.” Source • Screenshot", "Collected source links": "https://knowara.com/ai-tools/image/leonardo-ai-review/\nhttps://intercom.help/leonardo-ai/en/articles/8658301-realtime-canvas\nhttps://downloads.intercomcdn.com/i/o/916146471/f2c2e0904d1d6f4585b91c05/RTC_Settings_.gif", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7cfe970efa38", "Operation IDs": "195", "Citation": "Knowara · Leonardo Realtime Canvas", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://knowara.com/ai-tools/image/leonardo-ai-review/" }, { "Source ID": "collected-75913062213e", "Operation IDs": "195", "Citation": "Leonardo current Realtime Canvas guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://intercom.help/leonardo-ai/en/articles/8658301-realtime-canvas" }, { "Source ID": "collected-7d3ddf637750", "Operation IDs": "195", "Citation": "downloads.intercomcdn.com — RTC_Settings_.gif", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://downloads.intercomcdn.com/i/o/916146471/f2c2e0904d1d6f4585b91c05/RTC_Settings_.gif" } ] }, { "ID": "196", "Software": "Adobe Lightroom Classic 15.5", "Operation": "Generative Remove", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3e357a8ccfe4\ncollected-67b33a468b37\ncollected-423fc948bf33\ncollected-9c33556e8e8c\ncollected-36ee55feb0aa\ncollected-d3fe9de86368\ncollected-9b99eaf95a5f\ncollected-12314321e0c3\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Lightroom Generative Remove uses Firefly to synthesize replacement content in a brushed removal region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Adobe (Oct 27, 2025): Generative Remove is powered by Adobe Firefly and generates three variations per Remove; Generate makes three more. Detect objects adds the object's shadow and reflections to the brushed selection. No generative credits are deducted. Adobe, Generative Remove, Oct 27, 2025\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Firefly). Designer controls: brush; Use Generative AI checkbox; Detect objects; Generate (new set). Returns several to choose from (3); form not documented. “Lightroom Classic will use Adobe Firefly to generate three different variations.” Source • Screenshot Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/lightroom-classic/desktop/introduction-to-lightroom-classic/release-notes.html\nhttps://helpx.adobe.com/ca/lightroom-classic/desktop/process-and-develop-photos/remove-tool.html\nhttps://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/remove-tool.html\nhttps://helpx.adobe.com/lightroom-classic/help/remove-tool.html\nhttps://blog.adobe.com/en/publish/2024/05/21/lightroom-introduces-power-adobe-firefly-with-generative-remove\nhttps://web.archive.org/web/20260520203829id_/https://helpx.adobe.com/lightroom-classic/help/develop-module-options.html\nhttps://web.archive.org/web/20260826045346id_/https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/remove-tool.html\nhttps://web.archive.org/web/20251128223715id_/https://helpx.adobe.com/camera-raw/using/ai-edit-status.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3e357a8ccfe4", "Operation IDs": "196, 197", "Citation": "Adobe: Classic release notes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/lightroom-classic/desktop/introduction-to-lightroom-classic/release-notes.html" }, { "Source ID": "collected-67b33a468b37", "Operation IDs": "196", "Citation": "Adobe: Generative Remove procedure", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/ca/lightroom-classic/desktop/process-and-develop-photos/remove-tool.html" }, { "Source ID": "collected-423fc948bf33", "Operation IDs": "196", "Citation": "Adobe, Generative Remove, Oct 27, 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/remove-tool.html" }, { "Source ID": "collected-9c33556e8e8c", "Operation IDs": "196", "Citation": "helpx.adobe.com — remove-tool.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/lightroom-classic/help/remove-tool.html" }, { "Source ID": "collected-36ee55feb0aa", "Operation IDs": "196", "Citation": "blog.adobe.com — lightroom-introduces-power-adobe-firefly-with-generative-remove", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2024/05/21/lightroom-introduces-power-adobe-firefly-with-generative-remove" }, { "Source ID": "collected-d3fe9de86368", "Operation IDs": "196", "Citation": "web.archive.org — develop-module-options.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260520203829id_/https://helpx.adobe.com/lightroom-classic/help/develop-module-options.html" }, { "Source ID": "collected-9b99eaf95a5f", "Operation IDs": "196", "Citation": "web.archive.org — remove-tool.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260826045346id_/https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/remove-tool.html" }, { "Source ID": "collected-12314321e0c3", "Operation IDs": "196, 198, 386", "Citation": "web.archive.org — ai-edit-status.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20251128223715id_/https://helpx.adobe.com/camera-raw/using/ai-edit-status.html" } ] }, { "ID": "197", "Software": "Adobe Lightroom Classic 15.5", "Operation": "Select Subject and manual mask refinement", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-81daf85ef1ac\ncollected-2e6b507e6c64\ncollected-ccbc0eb510b0\ncollected-54916c7d1b56\ncollected-3e357a8ccfe4", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Lightroom Select Subject predicts a mask followed by conventional manual brush refinement with inferred coverage retained.", "Previous record: rationale": "The retained learned membership supports a region-reference match. Brush coverage does not create a justified proportional reduction in X.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Lightroom Classic 15.4 (June 2026) updated subject detection for Select Subject masks, and 15.5 (4 August 2026) added Feather and Edge sliders that soften or shift the boundary of AI-generated masks. Both changes refine mask membership; no generative step is documented for Select Subject. Not established: The Select Subject model architecture and processing location are undisclosed. Adobe helpx pages refused direct fetch. community.adobe.com, 2026-08-04 • vernonchalmers.photography, 2026-08 (v15.5) • helpx.adobe.com, June 2026 (v15.4)", "Collected source links": "https://blog.adobe.com/en/publish/2021/09/28/from-the-acr-team-masking-reimagined\nhttps://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/masking.html\nhttps://community.adobe.com/announcements-673/lightroom-classic-v15-5-is-here-better-masking-controls-render-to-dng-and-more-1635193\nhttps://www.vernonchalmers.photography/2026/08/what-is-new-in-lightroom-classic-155.html\nhttps://helpx.adobe.com/lightroom-classic/desktop/introduction-to-lightroom-classic/release-notes.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-81daf85ef1ac", "Operation IDs": "197", "Citation": "Adobe ACR team: Masking Reimagined", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2021/09/28/from-the-acr-team-masking-reimagined" }, { "Source ID": "collected-2e6b507e6c64", "Operation IDs": "197", "Citation": "Adobe: Lightroom Classic masking", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/masking.html" }, { "Source ID": "collected-ccbc0eb510b0", "Operation IDs": "197", "Citation": "community.adobe.com, 2026-08-04", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.adobe.com/announcements-673/lightroom-classic-v15-5-is-here-better-masking-controls-render-to-dng-and-more-1635193" }, { "Source ID": "collected-54916c7d1b56", "Operation IDs": "197", "Citation": "vernonchalmers.photography, 2026-08 (v15.5)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.vernonchalmers.photography/2026/08/what-is-new-in-lightroom-classic-155.html" }, { "Source ID": "collected-3e357a8ccfe4", "Operation IDs": "196, 197", "Citation": "Adobe: Classic release notes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/lightroom-classic/desktop/introduction-to-lightroom-classic/release-notes.html" } ] }, { "ID": "198", "Software": "Adobe Lightroom Classic 15.5", "Operation": "Denoise in the active Detail panel", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8ad03486ea84\ncollected-9e802c6bd5af\ncollected-0720eb5e5ab0\ncollected-fe1be6905f50\ncollected-049faaee02ce\ncollected-12314321e0c3\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Lightroom Denoise uses its documented learned RAW-restoration operation in the active Detail panel.", "Previous record: rationale": "Learned photographic signal recovery matches fidelity restoration. Amount controls are recorded without being converted into t.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 Adobe (May 1, 2026): Denoise sits in the Detail panel with a slider for intensity; Adobe describes it as using artificial intelligence. Adobe recommends Denoise before AI masks and Remove, because Enhance can change their results. Adobe, Enhance, May 1, 2026\nChecked 3 Oct 2026 X unchanged (8.0). Adobe's Lightroom Classic help page (updated 1 May 2026) describes Denoise as artificial-intelligence noise reduction that keeps fine detail, and Adobe's April 2023 engineering post states the model was trained on millions of high-noise and low-noise patch pairs to demosaic and denoise in one step. Lightroom Classic 15.5 (August 2026) lists no change to the Denoise model; its Denoise items concern iPad availability, Render to DNG and a Paste-workflow fix. Not established: Loss terms and any retraining after 2023 are not published. helpx.adobe.com, Last updated 2026-05-01 • blog.adobe.com, 2023-04-18 • lightroomqueen.com, 2026-08-03, updated 2026-09-29", "Collected source links": "https://blog.adobe.com/en/publish/2023/04/18/denoise-demystified\nhttps://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/enhance-details.html\nhttps://www.lightroomqueen.com/whats-new-in-lightroom-2026-08/\nhttps://web.archive.org/web/20260422080740id_/https://helpx.adobe.com/lightroom-classic/help/photos.html\nhttps://web.archive.org/web/20260922110218id_/https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/enhance-details.html\nhttps://web.archive.org/web/20251128223715id_/https://helpx.adobe.com/camera-raw/using/ai-edit-status.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8ad03486ea84", "Operation IDs": "198", "Citation": "blog.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2023/04/18/denoise-demystified" }, { "Source ID": "collected-9e802c6bd5af", "Operation IDs": "198", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/enhance-details.html" }, { "Source ID": "collected-0720eb5e5ab0", "Operation IDs": "198", "Citation": "lightroomqueen.com, 2026-08-03, updated 2026-09-29", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.lightroomqueen.com/whats-new-in-lightroom-2026-08/" }, { "Source ID": "collected-fe1be6905f50", "Operation IDs": "198, 386", "Citation": "web.archive.org — photos.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260422080740id_/https://helpx.adobe.com/lightroom-classic/help/photos.html" }, { "Source ID": "collected-049faaee02ce", "Operation IDs": "198, 386", "Citation": "web.archive.org — enhance-details.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260922110218id_/https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/enhance-details.html" }, { "Source ID": "collected-12314321e0c3", "Operation IDs": "196, 198, 386", "Citation": "web.archive.org — ai-edit-status.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20251128223715id_/https://helpx.adobe.com/camera-raw/using/ai-edit-status.html" } ] }, { "ID": "199", "Software": "LookX AI Cloud", "Operation": "Image to Video Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-91160c299f9d\ncollected-1150ae44e74e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "LookX generates moving imagery through its image-to-video service.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=4; conditional=0; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. A directory listing describes LookX as powered by diffusion and multimodal models without names. Model choice: Not documented. Not established: Which video model generates the clip (in-house or third-party such as Kling, Hailuo, Vidu, Wan or Seedance); clip length; whether users choose a model. lookx.ai, page date not shown • aidirectory.aecmag.com, page date not shown\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (image-to-video). Designer controls: image; text prompt. Returns count not documented; a flat video. “Generate video based on image or text. Bring static images or textual prompt to life with animations.” Source", "Collected source links": "https://www.lookx.ai/\nhttps://aidirectory.aecmag.com/entry/lookx/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-91160c299f9d", "Operation IDs": "199, 200", "Citation": "LookX · Image and video generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.lookx.ai/" }, { "Source ID": "collected-1150ae44e74e", "Operation IDs": "199", "Citation": "aidirectory.aecmag.com, page date not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://aidirectory.aecmag.com/entry/lookx/" } ] }, { "ID": "200", "Software": "LookX AI Cloud", "Operation": "Sketch to Rendered Image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4e351cb6dc42\ncollected-f1a7593051f6\ncollected-91160c299f9d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "LookX generates an architectural visualization from its sketch-to-image route.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=5; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented by LookX. A third-party tutorial states that LookX offers the official Stable Diffusion 1.5 model (Stability AI / Runway) plus custom-trained models made by LookX; Model choice: Yes, per the third-party source: the official Stable Diffusion 1.5 model, LookX custom-trained base models, and user-trained style models. Not established: LookX does not name its base model on its site or help pages located. toolify.ai • parametric-architecture.com, 2023-06-29 • lookx.ai, accessed 2026-10-04\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (sketch-to-render image generation). Designer controls: sketch or 3D geometry screenshot; text prompt; labels; style reference image; model choice; control mode (by content / by composition); aspect ratio. Returns count not documented; a flat image. “Transform text and sketches into rendering images instantly.” Source Includes a third-party source.", "Collected source links": "https://parametric-architecture.com/lookx-launched-an-ai-generated-tool-that-can-transform-sketch-or-squashed-paper-into-a-model/\nhttps://www.toolify.ai/ai-news/transform-your-sketch-plan-into-3d-render-with-lookx-ai-1540026\nhttps://www.lookx.ai/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4e351cb6dc42", "Operation IDs": "200", "Citation": "Parametric Architecture · LookX Render Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://parametric-architecture.com/lookx-launched-an-ai-generated-tool-that-can-transform-sketch-or-squashed-paper-into-a-model/" }, { "Source ID": "collected-f1a7593051f6", "Operation IDs": "200", "Citation": "www.toolify.ai — transform-your-sketch-plan-into-3d-render-with-lookx-ai-1540026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.toolify.ai/ai-news/transform-your-sketch-plan-into-3d-render-with-lookx-ai-1540026" }, { "Source ID": "collected-91160c299f9d", "Operation IDs": "199, 200", "Citation": "LookX · Image and video generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.lookx.ai/" } ] }, { "ID": "201", "Software": "Luma Dream Machine", "Operation": "Ray3.2 Modify Video V2 full-clip restyling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-fb1fbaa213a0\ncollected-6c1192177f05", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Luma Ray3.2 Modify Video V2 generates a restyled version of the scoped clip.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 1 unknown and 4 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=1; conditional=4; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Ray3.2 video-to-video). Designer controls: source video; keyframes (up to 16); motion and structure settings; character tracking (Poses or Blocking); face tracking; resolution; duration; aspect ratio; HDR/EXR output. Returns count not documented; a flat video: MP4, HDR or 16-bit EXR. “Modify Video reshapes footage you already have. Swap the wall, the world, the wardrobe.” Source", "Collected source links": "https://lumalabs.ai/ray\nhttps://lumalabs.ai/news/introducing-ray-3-2", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-fb1fbaa213a0", "Operation IDs": "201, 202", "Citation": "Luma Ray3.2 generation and multi-keyframes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://lumalabs.ai/ray" }, { "Source ID": "collected-6c1192177f05", "Operation IDs": "201", "Citation": "Luma Ray3.2 release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://lumalabs.ai/news/introducing-ray-3-2" } ] }, { "ID": "202", "Software": "Luma Dream Machine", "Operation": "Ray3.2 image-to-video with supplied keyframes", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.2–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-fb1fbaa213a0", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Luma Ray3.2 generates video through its supplied-keyframe image-to-video route.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 1 model-decided, 3 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 3.2–5.0; no midpoint is used.", "Previous record: calculation": "K=2; M=1; unknown=3; conditional=0; A=6. Lower: 2 + 3 × 2 ÷ (6 − 1) = 3.2. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Ray3.2 image-to-video). Designer controls: start image; keyframes (up to 16); prompt; duration (5 or 10 s); resolution (360p-1080p); aspect ratio; draft mode; HDR/EXR output. Returns count not documented; a flat video. “You can add up to 16 keyframes for image2video” Source", "Collected source links": "https://lumalabs.ai/ray", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-fb1fbaa213a0", "Operation IDs": "201, 202", "Citation": "Luma Ray3.2 generation and multi-keyframes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://lumalabs.ai/ray" } ] }, { "ID": "203", "Software": "Luminar Neo", "Operation": "GenErase masked object removal", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-10553a28e182\ncollected-fbfc0289a336\ncollected-e196e8506ead\ncollected-fe478d43e213\ncollected-8d7259d553bd", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Luminar GenErase synthesizes content replacing a selected object.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Stable Diffusion (Stability AI), version not stated, combined by Skylum with other diffusion models and its own Upscale AI model in one cloud pipeline. Model choice: No. Not established: The Stable Diffusion version (1.5, 2.x, XL or later) is not documented. skylum.com, 2023-10-26 • thephoblographer.com • techcrunch.com, 2023-10-26\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (content-aware generative fill). Designer controls: select/deselect brush; brush size; reset selection; Erase button (repeat for a new outcome); preview; save. Returns one result (click Erase again to cycle outcomes); a flat image: new TIF file. “It not only removes undesirable visual distractions from your images but also applies content-aware AI techniques to replace these elements intelligently.” Source", "Collected source links": "https://support.skylum.com/extra-tools/generative-tools/generase\nhttps://support.skylum.com/luminar-neo-tips/generative-tools-in-luminar-neo-everything-you-need-to-know\nhttps://skylum.com/newsroom/luminar-neo-releases-generative-ai-technology-in-its-awardwinning-photo-editing-software\nhttps://www.thephoblographer.com/2024/01/18/luminar-neo-review/\nhttps://techcrunch.com/2023/10/26/luminar-neo-brings-generative-ai-tools-to-hobbyist-photographers", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-10553a28e182", "Operation IDs": "203", "Citation": "E1 Skylum GenErase user guide — consulted 09272026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.skylum.com/extra-tools/generative-tools/generase" }, { "Source ID": "collected-fbfc0289a336", "Operation IDs": "203, 204, 205", "Citation": "support.skylum.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.skylum.com/luminar-neo-tips/generative-tools-in-luminar-neo-everything-you-need-to-know" }, { "Source ID": "collected-e196e8506ead", "Operation IDs": "203, 204, 205", "Citation": "skylum.com, 2023-10-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://skylum.com/newsroom/luminar-neo-releases-generative-ai-technology-in-its-awardwinning-photo-editing-software" }, { "Source ID": "collected-fe478d43e213", "Operation IDs": "203", "Citation": "thephoblographer.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.thephoblographer.com/2024/01/18/luminar-neo-review/" }, { "Source ID": "collected-8d7259d553bd", "Operation IDs": "203, 204, 205", "Citation": "techcrunch.com, 2023-10-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://techcrunch.com/2023/10/26/luminar-neo-brings-generative-ai-tools-to-hobbyist-photographers" } ] }, { "ID": "204", "Software": "Luminar Neo", "Operation": "GenExpand canvas outpainting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–2.75", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-20b36c5e8fe9\ncollected-fbfc0289a336\ncollected-e196e8506ead\ncollected-8d7259d553bd\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Luminar GenExpand generates content beyond the image boundary.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 1 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–2.75; no midpoint is used.", "Previous record: calculation": "K=0; M=4; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 1 ÷ (5 − 1) = 2.75. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Stable Diffusion (Stability AI), version not stated, inside the same Skylum cloud pipeline as GenErase and GenSwap. Model choice: No. Not established: The Stable Diffusion version and any outpainting-specific checkpoint are not documented. skylum.com, 2023-10-26 • techcrunch.com\nDesigner side, 3 Oct 2026 Other. Decided by a trained model (outpainting). Designer controls: expansion of image bounds (method not described); prompt (What do you envision here?); Expand button (repeat to cycle); undo/redo; save. Returns one result (click Expand again to cycle outcomes); a flat image: new TIF file. “extend the boundaries of your photos, whether you're creating panoramic landscapes or giving your favorite images more space.” Source", "Collected source links": "https://support.skylum.com/extra-tools/generative-tools/genexpand\nhttps://support.skylum.com/luminar-neo-tips/generative-tools-in-luminar-neo-everything-you-need-to-know\nhttps://skylum.com/newsroom/luminar-neo-releases-generative-ai-technology-in-its-awardwinning-photo-editing-software\nhttps://techcrunch.com/2023/10/26/luminar-neo-brings-generative-ai-tools-to-hobbyist-photographers", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-20b36c5e8fe9", "Operation IDs": "204", "Citation": "X1 Skylum GenExpand user guide — consulted 09272026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.skylum.com/extra-tools/generative-tools/genexpand" }, { "Source ID": "collected-fbfc0289a336", "Operation IDs": "203, 204, 205", "Citation": "support.skylum.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.skylum.com/luminar-neo-tips/generative-tools-in-luminar-neo-everything-you-need-to-know" }, { "Source ID": "collected-e196e8506ead", "Operation IDs": "203, 204, 205", "Citation": "skylum.com, 2023-10-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://skylum.com/newsroom/luminar-neo-releases-generative-ai-technology-in-its-awardwinning-photo-editing-software" }, { "Source ID": "collected-8d7259d553bd", "Operation IDs": "203, 204, 205", "Citation": "techcrunch.com, 2023-10-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://techcrunch.com/2023/10/26/luminar-neo-brings-generative-ai-tools-to-hobbyist-photographers" } ] }, { "ID": "205", "Software": "Luminar Neo", "Operation": "GenSwap masked object generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–2.75", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-85764c5e5f8b\ncollected-fbfc0289a336\ncollected-e196e8506ead\ncollected-8d7259d553bd", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Luminar GenSwap synthesizes inserted or replacement image content.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 1 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–2.75; no midpoint is used.", "Previous record: calculation": "K=0; M=4; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 1 ÷ (5 − 1) = 2.75. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Stable Diffusion (Stability AI), version not stated, inside the same Skylum cloud pipeline as GenErase and GenExpand. Model choice: No. Not established: The Stable Diffusion version and inpainting checkpoint are not documented. skylum.com, 2023-10-26 • techcrunch.com\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (masked object generation). Designer controls: select/deselect brush; brush size; prompt (What do you envision here?); Swap button (repeat to cycle); preview; undo/redo; save. Returns one result (click Swap again to cycle outcomes); a flat image: new TIF file. “GenSwap is advanced Generative AI technology that empowers photographers to alter image compositions by substituting or exchanging objects.” Source", "Collected source links": "https://support.skylum.com/extra-tools/generative-tools/genswap\nhttps://support.skylum.com/luminar-neo-tips/generative-tools-in-luminar-neo-everything-you-need-to-know\nhttps://skylum.com/newsroom/luminar-neo-releases-generative-ai-technology-in-its-awardwinning-photo-editing-software\nhttps://techcrunch.com/2023/10/26/luminar-neo-brings-generative-ai-tools-to-hobbyist-photographers", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-85764c5e5f8b", "Operation IDs": "205", "Citation": "S1 Skylum GenSwap user guide — consulted 09272026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.skylum.com/extra-tools/generative-tools/genswap" }, { "Source ID": "collected-fbfc0289a336", "Operation IDs": "203, 204, 205", "Citation": "support.skylum.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.skylum.com/luminar-neo-tips/generative-tools-in-luminar-neo-everything-you-need-to-know" }, { "Source ID": "collected-e196e8506ead", "Operation IDs": "203, 204, 205", "Citation": "skylum.com, 2023-10-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://skylum.com/newsroom/luminar-neo-releases-generative-ai-technology-in-its-awardwinning-photo-editing-software" }, { "Source ID": "collected-8d7259d553bd", "Operation IDs": "203, 204, 205", "Citation": "techcrunch.com, 2023-10-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://techcrunch.com/2023/10/26/luminar-neo-brings-generative-ai-tools-to-hobbyist-photographers" } ] }, { "ID": "206", "Software": "Luminar Neo", "Operation": "Sky AI replacement using a generated sky asset", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8853df9d4b88\ncollected-ec524ff2aae6\ncollected-80a0a5d9def2", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Luminar Sky AI estimates sky membership and composites an already supplied replacement sky.", "Previous record: rationale": "Learned segmentation remains consequential through conventional asset placement, supporting region determination. Earlier generation of the sky asset is outside this operation.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Skylum's Sky AI help describes automatic detection of the sky boundary and water, compositing of a library or user sky, and relight and reflection controls; it does not describe generating sky content. The Luminar update of 5 August 2026 rebuilt Skylum's generative model for other tools, and the changelog lists no Sky AI change for 2025 to 2026. Not established: The Sky AI segmentation model is undisclosed, and the mechanism of the relight step is not documented. support.skylum.com, undated • petapixel.com, 2026-08-05", "Collected source links": "https://skylum.com/luminar\nhttps://support.skylum.com/editing-tools/landscape-tools/sky-ai\nhttps://petapixel.com/2026/08/05/now-that-generative-ai-is-the-villain-luminar-is-spinning-its-messaging/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8853df9d4b88", "Operation IDs": "206", "Citation": "Skylum · Luminar SkyAI", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://skylum.com/luminar" }, { "Source ID": "collected-ec524ff2aae6", "Operation IDs": "206", "Citation": "Skylum Sky AI controls and supplied sky", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.skylum.com/editing-tools/landscape-tools/sky-ai" }, { "Source ID": "collected-80a0a5d9def2", "Operation IDs": "206", "Citation": "petapixel.com, 2026-08-05", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://petapixel.com/2026/08/05/now-that-generative-ai-is-the-villain-luminar-is-spinning-its-messaging/" } ] }, { "ID": "207", "Software": "Magnific (direct / Krea Enhancer)", "Operation": "Creative image upscaling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-351af39f3d6a\ncollected-be267953af28\ncollected-d26f8a914c85\ncollected-2d2a6d225a74", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magnific Creative Upscale synthesizes plausible detail through its creative enhancement route.", "Previous record: rationale": "The trained plausible-detail responsibility matches perceptual reconstruction at 5. This is a task analogy, not a claim that Magnific uses ESRGAN architecture or losses.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0). Freepik adopted the Magnific name company-wide on 28 April 2026 (third-party report); Magnific's Creative mode is described as reconstructing the image and inventing detail, with Precision mode (Magnific v1/v2) a separate faithful route. Krea's current Enhancer docs table lists no Magnific model, while Krea's blog (updated September 2026) still points to Magnific under Generative Enhancers. Mechanism evidence is unchanged. Not established: Vendor-stated architecture and training objective; whether the Krea route to Magnific remains available, since Krea's docs and blog disagree. magnific.com docs refused fetch. oakgen.ai, article 2026-09-26; event 2026-04-28 • krea.ai, 2026-06-08; updated 2026-09 • krea.ai, not shown", "Collected source links": "https://www.magnific.com/blog/how-to-use-magnific-ai-image-upscaler-a-comprehensive-guide/\nhttps://www.krea.ai/blog/how-to-enhance-images-with-magnific\nhttps://oakgen.ai/blog/magnific-ai-freepik-rebrand-pricing\nhttps://www.krea.ai/docs/user-guide/features/enhancer", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-351af39f3d6a", "Operation IDs": "207", "Citation": "Magnific · Upscaler", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.magnific.com/blog/how-to-use-magnific-ai-image-upscaler-a-comprehensive-guide/" }, { "Source ID": "collected-be267953af28", "Operation IDs": "207", "Citation": "Krea Magnific creative enhancement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/blog/how-to-enhance-images-with-magnific" }, { "Source ID": "collected-d26f8a914c85", "Operation IDs": "207", "Citation": "oakgen.ai, article 2026-09-26; event 2026-04-28", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://oakgen.ai/blog/magnific-ai-freepik-rebrand-pricing" }, { "Source ID": "collected-2d2a6d225a74", "Operation IDs": "207", "Citation": "krea.ai, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/docs/user-guide/features/enhancer" } ] }, { "ID": "208", "Software": "Magnific (Krea Enhancer reference route)", "Operation": "Precise image upscaling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-81e51f064d4a\ncollected-9ffed6180555\ncollected-288ffc853704\ncollected-ffbb71483396", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magnific Precise reconstructs enlarged image detail through its fidelity-oriented enhancement route.", "Previous record: rationale": "The task supports 5–8 between perceptual reconstruction and fidelity restoration. The name Precise alone does not establish a particular loss or exact point.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Krea's Magnific Precise controls (Sharpen, Smart Grain, Ultra Detail) match the parameters that Magnific's API documentation lists as new in Upscaler Precision V2. That documentation describes V2 as faithful upscaling that preserves the original content, while its endpoint page says V2 adds new visual elements or details and may modify the original image content; no architecture or training objective is published. Not established: Whether Precision V2 runs a diffusion, GAN or fidelity-trained network, and its training objective. www.magnific.com/ai/docs/image-upscaler refused automated fetch (HTTP 403). docs.magnific.com, not shown • docs.magnific.com, not shown • krea.ai, not shown", "Collected source links": "https://www.krea.ai/blog/how-to-enhance-images-with-magnific-precise\nhttps://www.magnific.com/ai/docs/image-upscaler\nhttps://docs.magnific.com/api-reference/image-upscaler-precision-v2/overview\nhttps://docs.magnific.com/api-reference/image-upscaler-precision-v2/post-image-upscaler-precision-v2", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-81e51f064d4a", "Operation IDs": "208", "Citation": "Krea Magnific Precise controls and distinction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/blog/how-to-enhance-images-with-magnific-precise" }, { "Source ID": "collected-9ffed6180555", "Operation IDs": "208", "Citation": "www.magnific.com — image-upscaler", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.magnific.com/ai/docs/image-upscaler" }, { "Source ID": "collected-288ffc853704", "Operation IDs": "208", "Citation": "docs.magnific.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.magnific.com/api-reference/image-upscaler-precision-v2/overview" }, { "Source ID": "collected-ffbb71483396", "Operation IDs": "208", "Citation": "docs.magnific.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.magnific.com/api-reference/image-upscaler-precision-v2/post-image-upscaler-precision-v2" } ] }, { "ID": "209", "Software": "Maket", "Operation": "AI Floor Plan Generation to Editable Layout", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–6.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a498dfcb2d26\ncollected-111aeb371e53\ncollected-edfd261f633c\ncollected-103a863a8429", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Maket generates a native floor-plan layout through its documented AI plan-generation route.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The model determines room geometry and arrangement of the plan as new coordinates, which places the comparison nearer synthesis than region determination, at 2–6 within the same envelope. Doors, windows and furniture come from a post-processing layer of undisclosed mechanism.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–0.5; X = 2–6. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 2.0–6.0. Maket's generator guide (13 April 2026, updated 12 August 2026) states that a frontier AI model receives user inputs with predefined architectural rules and returns room layouts as vector coordinates; a proprietary post-processing layer then fixes circulation and adds doors, windows and furniture. A third-party report dates Maket 2.0 to 2 October 2026 and describes a floor plan generator trained on plans. Not established: Which frontier model; whether post-processing is learned or rule-based; whether Maket 2.0 replaces the frontier-model route with its own trained generator. Previous reason: Learned room and spatial arrangement decisions support 2–10. Editable layout structure does not determine a separate X bonus. maket.ai, 2026-04-13, updated 2026-08-12 • progressiverobot.com, 2026-10-02", "Collected source links": "https://www.maket.ai/features\nhttps://www.maket.ai/blog/how-to-use-maket\nhttps://www.maket.ai/blog/ai-floor-plan-generator-guide\nhttps://www.progressiverobot.com/2026/10/02/maket-2-0-ai-home-design-floor-plans/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a498dfcb2d26", "Operation IDs": "209", "Citation": "Maket · Floor Plan Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.maket.ai/features" }, { "Source ID": "collected-111aeb371e53", "Operation IDs": "209", "Citation": "Maket · Floor plan workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.maket.ai/blog/how-to-use-maket" }, { "Source ID": "collected-edfd261f633c", "Operation IDs": "209", "Citation": "maket.ai, 2026-04-13, updated 2026-08-12", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.maket.ai/blog/ai-floor-plan-generator-guide" }, { "Source ID": "collected-103a863a8429", "Operation IDs": "209", "Citation": "progressiverobot.com, 2026-10-02", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.progressiverobot.com/2026/10/02/maket-2-0-ai-home-design-floor-plans/" } ] }, { "ID": "210", "Software": "Marble", "Operation": "Regional World Regeneration", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3325011d5982\ncollected-62c890a5a8e3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Marble regenerates a selected region of a learned three-dimensional world.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Marble world model, panorama edit). Designer controls: Select edit area (click-drag outline); edit prompt; add images; Preview draft; Apply edit; Create world. Returns count not documented; editable structure: new world version from the edited panorama; Gaussian splats. “Click and drag to outline the specific area you want to modify” Source", "Collected source links": "https://www.worldlabs.ai/blog/marble-world-model\nhttps://docs.worldlabs.ai/marble/edit/pano-edit", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3325011d5982", "Operation IDs": "210, 211", "Citation": "World Labs generation and output states", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.worldlabs.ai/blog/marble-world-model" }, { "Source ID": "collected-62c890a5a8e3", "Operation IDs": "210", "Citation": "docs.worldlabs.ai — pano-edit", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.worldlabs.ai/marble/edit/pano-edit" } ] }, { "ID": "211", "Software": "Marble", "Operation": "World Generation with Splat Export", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3325011d5982", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Marble's world model generates a scene that can be exported as Gaussian splats.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Marble multimodal world model). Designer controls: text prompt; single image; multiple images; video; Chisel 3D layout. Returns count not documented; editable structure: Gaussian splats; collider and high-quality triangle meshes; video export. “Gaussian splats are the highest-fidelity representation for Marble worlds.” Source", "Collected source links": "https://www.worldlabs.ai/blog/marble-world-model", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3325011d5982", "Operation IDs": "210, 211", "Citation": "World Labs generation and output states", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.worldlabs.ai/blog/marble-world-model" } ] }, { "ID": "212", "Software": "Max (Max/MSP/Jitter)", "Operation": "Jitter jit.gl.gridshape Working Patch", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-63d3d11ff9a7\ncollected-af58b89d542c\ncollected-e08e45d1ba73", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Max/Jitter evaluates the supplied jit.gl.gridshape patch and its explicit parameters.", "Previous record: rationale": "Conventional shape generation contains no trained predictor in the bounded visual patch.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: nodes/parameters (patch objects); shape; dim; gridmode; rad_minor; position; scale; rotate; color; lighting_enable. Returns continuous live output; editable structure: patch parameters; renders directly or outputs a matrix. “These shapes may be either rendered directly, or sent out the leftmost outlet as a matrix of values.” Source", "Collected source links": "https://cycling74.com/products/max\nhttps://cycling74.com/products/jitter\nhttps://docs.cycling74.com/reference/jit.gl.gridshape", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-63d3d11ff9a7", "Operation IDs": "212", "Citation": "Cycling 74 · Max/MSP Patching", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cycling74.com/products/max" }, { "Source ID": "collected-af58b89d542c", "Operation IDs": "212", "Citation": "Cycling 74 · Jitter video and graphics", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cycling74.com/products/jitter" }, { "Source ID": "collected-e08e45d1ba73", "Operation IDs": "212", "Citation": "docs.cycling74.com — jit.gl.gridshape", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.cycling74.com/reference/jit.gl.gridshape" } ] }, { "ID": "213", "Software": "Maxon Cinema 4D", "Operation": "MoGraph Cloner and Effector Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-483422cb1b05\ncollected-bd5586be372f\ncollected-0a104962e20c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Cinema 4D MoGraph evaluates Cloner and Effector rules for the specified procedural construction.", "Previous record: rationale": "Rule-driven distribution and animation are conventional. Procedural generation is not by itself learned synthesis.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: mode (Object, Linear, Radial, Grid, Honeycomb); count; per-clone move/rotate/scale offsets; child objects; effectors (Random, Shader, Target, Step, Volume). Returns one result; editable structure: procedural clones under the Cloner object. “Each clone will be moved, rotated or scaled by a certain value relative to the previous clone” Source • Screenshot", "Collected source links": "https://superrendersfarm.com/article/whats-new-cinema-4d-2026-features\nhttps://help.maxon.net/c4d/2026/en-us/Content/html/OMOGRAPH_CLONER.html\nhttps://help.maxon.net/c4d/r25/en-us/Content/html/OMOGRAPH_CLONER-ID_OBJECTPROPERTIES.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-483422cb1b05", "Operation IDs": "213", "Citation": "Super Renders Farm · Cinema 4D Parametric/MoGraph", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://superrendersfarm.com/article/whats-new-cinema-4d-2026-features" }, { "Source ID": "collected-bd5586be372f", "Operation IDs": "213", "Citation": "Maxon · MoGraph Cloner relationships", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/c4d/2026/en-us/Content/html/OMOGRAPH_CLONER.html" }, { "Source ID": "collected-0a104962e20c", "Operation IDs": "213", "Citation": "help.maxon.net — OMOGRAPH_CLONER-ID_OBJECTPROPERTIES.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/c4d/r25/en-us/Content/html/OMOGRAPH_CLONER-ID_OBJECTPROPERTIES.html" } ] }, { "ID": "214", "Software": "Maxon Cinema 4D iPad and Desktop", "Operation": "HY 3D Base Mesh Generation Announced", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-65f111af850f\ncollected-fa46f66561ca\ncollected-cc648601f7ec\ncollected-cca7d78197e8", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Maxon describes an announced HY 3D mesh-generation integration for Cinema 4D.", "Previous record: rationale": "The release status and selected implemented operation remain unverified. Preserve the announced case without issuing an installed-operation coordinate.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). No HY 3D generation feature has shipped as of Cinema 4D 2026.4 (released 30 September 2026); Maxon's 6 March 2026 announcement schedules it for Cinema 4D for iPad in late 2026, then desktop. The announcement names the 'Tencent HY 3D Global AI engine' without a model version; Tencent's Hunyuan3D 2.0 (January 2025) generates shape with a flow-based diffusion transformer conditioned on an image. Not established: Shipping date, the HY 3D model version in the installed route, and whether Maxon post-processes the generated mesh. If a Hunyuan3D 2.x diffusion model is confirmed in the shipped route, REF02 becomes the candidate lower reference. maxon.net, 2026-03-06 • cgchannel.com, 2026-09-30 (Cinema 4D 2026.4; article covers • cgchannel.com, 2026-04-16", "Collected source links": "https://www.therookies.co/blog/headlines/maxon-unveils-cinema-4d-ipad-edition\nhttps://www.maxon.net/en/article/maxon-and-tencent-cloud-partner-to-integrate-hy-3d-ai-engine-into-cinema-4d\nhttps://www.cgchannel.com/2026/09/maxon-releases-cinema-4d-2026-4-with-a-new-mcp-server/\nhttps://www.cgchannel.com/2026/04/maxon-ships-cinema-4d-2026-2-unveils-cinema-4d-for-ipad-beta/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-65f111af850f", "Operation IDs": "214", "Citation": "The Rookies · Cinema 4D iPad announcement (feature unconfirmed)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.therookies.co/blog/headlines/maxon-unveils-cinema-4d-ipad-edition" }, { "Source ID": "collected-fa46f66561ca", "Operation IDs": "214", "Citation": "Maxon · HY 3D integration announcement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.maxon.net/en/article/maxon-and-tencent-cloud-partner-to-integrate-hy-3d-ai-engine-into-cinema-4d" }, { "Source ID": "collected-cc648601f7ec", "Operation IDs": "214", "Citation": "cgchannel.com, 2026-09-30 (Cinema 4D 2026.4; article covers ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/09/maxon-releases-cinema-4d-2026-4-with-a-new-mcp-server/" }, { "Source ID": "collected-cca7d78197e8", "Operation IDs": "214", "Citation": "cgchannel.com, 2026-04-16", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/04/maxon-ships-cinema-4d-2026-2-unveils-cinema-4d-for-ipad-beta/" } ] }, { "ID": "215", "Software": "Maxon Red Giant Magic Bullet Looks", "Operation": "Conventional grading of generated video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b0927194ab26\ncollected-a66421a22c9c\ncollected-f036f1a35002\ncollected-b0b5a059780c\ncollected-3a82b82e92d1\ncollected-3201e449f87e\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Magic Bullet Looks evaluates the selected exposure, contrast and curve tools on existing footage.", "Previous record: rationale": "The conventional grading functions contain no trained predictor in this declared stack.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: presets; Tool Chain (add, reorder tools); Tools drawer (Selective, Camera, Color Correction, Film); per-tool controls/sliders; scopes. Returns one result; editable structure: parameters: tool chain applied to the clip in the host. “Every preset is comprised of its own tools, and each of those tools has its own controls.” Source • Screenshot", "Collected source links": "https://support.maxon.net/hc/en-us/articles/23368967860636-Red-Giant-2026-0-0-September-10-2025\nhttps://help.maxon.net/rg/en-us/Content/html/03-MBL-overview-of-looks-interface.html\nhttps://help.maxon.net/rg/en-us/Content/Resources/Images/MBL_Tools_Drawer.png\nhttps://web.archive.org/web/20260923032504id_/https://helpx.adobe.com/premiere/desktop/organize-media/create-projects/start-projects.html\nhttps://help.maxon.net/rg/en-us/Content/html/10-MBL-subject-tools.html\nhttps://help.maxon.net/rg/en-us/Content/html/14-MBL-tool-chain.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b0927194ab26", "Operation IDs": "215", "Citation": "Maxon · Magic Bullet Looks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.maxon.net/hc/en-us/articles/23368967860636-Red-Giant-2026-0-0-September-10-2025" }, { "Source ID": "collected-a66421a22c9c", "Operation IDs": "215", "Citation": "Maxon Looks tool chain", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/rg/en-us/Content/html/03-MBL-overview-of-looks-interface.html" }, { "Source ID": "collected-f036f1a35002", "Operation IDs": "215", "Citation": "help.maxon.net — MBL_Tools_Drawer.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/rg/en-us/Content/Resources/Images/MBL_Tools_Drawer.png" }, { "Source ID": "collected-b0b5a059780c", "Operation IDs": "215, 273", "Citation": "web.archive.org — start-projects.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260923032504id_/https://helpx.adobe.com/premiere/desktop/organize-media/create-projects/start-projects.html" }, { "Source ID": "collected-3a82b82e92d1", "Operation IDs": "215", "Citation": "help.maxon.net — 10-MBL-subject-tools.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/rg/en-us/Content/html/10-MBL-subject-tools.html" }, { "Source ID": "collected-3201e449f87e", "Operation IDs": "215", "Citation": "help.maxon.net — 14-MBL-tool-chain.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/rg/en-us/Content/html/14-MBL-tool-chain.html" } ] }, { "ID": "216", "Software": "Maxon ZBrush", "Operation": "ZBrush Sculpting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9ac71e662120\ncollected-19ae4167f703\ncollected-c6fc02de373e\ncollected-8e0f2b28a65e\ncollected-cb14256978df", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ZBrush's conventional sculpting tools modify mesh geometry.", "Previous record: rationale": "The selected brush and geometric operations are conventional; sculpt complexity does not establish learned inference.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: sculpting brushes; layers (record mode); layer intensity slider; polypaint; masking. Returns one result; editable structure: mesh with sculpt layers. “Layers allow for a non-linear workflow. You can work with a model at many different stages of development simultaneously.” Source • Screenshot", "Collected source links": "https://www.maxon.net/en/zbrush\nhttps://help.maxon.net/zbr/en-us/Content/html/reference-guide/tool/polymesh/layers/layers.html\nhttps://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/sculpting/sculpting-brushes/sculpting-brushes.html\nhttps://support.maxon.net/hc/en-us/sections/7456686788252-ZBrush\nhttps://help.maxon.net/zbr/en-us/Content/Resources/Images/4R5-Tool-Layer.jpg", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9ac71e662120", "Operation IDs": "216", "Citation": "Maxon · ZBrush Sculpting", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.maxon.net/en/zbrush" }, { "Source ID": "collected-19ae4167f703", "Operation IDs": "216", "Citation": "Maxon · Sculpt layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/html/reference-guide/tool/polymesh/layers/layers.html" }, { "Source ID": "collected-c6fc02de373e", "Operation IDs": "216", "Citation": "ZBrush sculpting brushes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/sculpting/sculpting-brushes/sculpting-brushes.html" }, { "Source ID": "collected-8e0f2b28a65e", "Operation IDs": "216, 217", "Citation": "ZBrush release notes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.maxon.net/hc/en-us/sections/7456686788252-ZBrush" }, { "Source ID": "collected-cb14256978df", "Operation IDs": "216", "Citation": "help.maxon.net — 4R5-Tool-Layer.jpg", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/Resources/Images/4R5-Tool-Layer.jpg" } ] }, { "ID": "217", "Software": "Maxon ZBrush", "Operation": "ZRemesher Retopology", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-da5ddaf474ed\ncollected-ed2e97f561ec\ncollected-8e0f2b28a65e\ncollected-f081c205ee0b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ZRemesher applies its documented topology reconstruction algorithms to a mesh.", "Previous record: rationale": "The current evidence supports conventional algorithmic retopology. Automated topology choice alone does not establish a trained predictor; this remains an implementation-qualified assessment.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Target Polygons Count; Half/Same/Double; Adaptive; curve guides; PolyPaint density; Keep Groups; Keep Creases; Detect Edge; symmetry. Returns one result; editable structure: new mesh topology. “ZRemesher is now able to better detect hard surface edges, producing fewer polygons and with better accuracy.” Source • Screenshot", "Collected source links": "https://help.maxon.net/zbr/en-us/Content/html/reference-guide/tool/polymesh/geometry/zremesher/zremesher.html\nhttps://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/topology/zremesher/zremesher-4-0/zremesher-4-0.html\nhttps://support.maxon.net/hc/en-us/sections/7456686788252-ZBrush\nhttps://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/topology/zremesher/zremesher.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-da5ddaf474ed", "Operation IDs": "217", "Citation": "Maxon ZRemesher reference, version-3 controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/html/reference-guide/tool/polymesh/geometry/zremesher/zremesher.html" }, { "Source ID": "collected-ed2e97f561ec", "Operation IDs": "217", "Citation": "Maxon ZRemesher 4.0 account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/topology/zremesher/zremesher-4-0/zremesher-4-0.html" }, { "Source ID": "collected-8e0f2b28a65e", "Operation IDs": "216, 217", "Citation": "ZBrush release notes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.maxon.net/hc/en-us/sections/7456686788252-ZBrush" }, { "Source ID": "collected-f081c205ee0b", "Operation IDs": "217", "Citation": "help.maxon.net — zremesher.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/topology/zremesher/zremesher.html" } ] }, { "ID": "218", "Software": "Autodesk Maya", "Operation": "Conventional Dependency Graph Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f51be5b28709\ncollected-80651fef2ee7\ncollected-9e7c1feeeb27", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Maya evaluates a NURBS revolve node from a profile and its geometric parameters.", "Previous record: rationale": "The resulting surface follows a prescribed construction and dependency graph, without a trained predictor.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: input curve; axis; pivot; startSweep; endSweep; degree; sections; useTolerance; tolerance; bridge; autoCorrectNormal. Returns one result; editable structure: NURBS surface with construction history node. “Given an input curve this node creates a revolved surface about a specified axis of revolution.” Source", "Collected source links": "https://help.autodesk.com/cloudhelp/2027/ENU/Maya-Basics/files/GUID-51096BC4-32B7-4391-BE39-21641B374745.htm\nhttps://help.autodesk.com/cloudhelp/2027/ENU/Maya-Modeling/files/GUID-DEDBFFEC-0AE6-4392-B5FD-F153B1D59FCF.htm\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Maya-Tech-Docs/Nodes/revolve.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f51be5b28709", "Operation IDs": "218", "Citation": "M1 Autodesk Maya 2027 Dependency graph", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Maya-Basics/files/GUID-51096BC4-32B7-4391-BE39-21641B374745.htm" }, { "Source ID": "collected-80651fef2ee7", "Operation IDs": "218", "Citation": "M2 Autodesk Maya 2027 Revolve Options", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Maya-Modeling/files/GUID-DEDBFFEC-0AE6-4392-B5FD-F153B1D59FCF.htm" }, { "Source ID": "collected-9e7c1feeeb27", "Operation IDs": "218", "Citation": "help.autodesk.com — revolve.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Maya-Tech-Docs/Nodes/revolve.html" } ] }, { "ID": "219", "Software": "Autodesk Maya", "Operation": "ML Deformer Delta Prediction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-cfc2687f0eb2\ncollected-470d05f3b5d6\ncollected-46dac08c4f4f\ncollected-1f3767693a94\ncollected-a6bb78aae91a\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Maya ML Deformer Delta mode predicts corrective deformation learned from a rig's training examples.", "Previous record: rationale": "The learned surrogate approximates a defined deformation relationship, matching fixed-function emulation at 12. Correction size and vertex count do not supply weights.", "Previous record: calculation": "X = 12 + 0 × (12 − 12) = 12.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (12.0). Maya 2026 help describes the ML Deformer as learning to approximate complex deformation systems from posed training samples, and Maya 2027 (25 March 2026) product material describes the same function. Coverage of Maya 2027 documents no change to Delta mode. Not established: Maya 2027 ML Deformer help pages were not checked directly. help.autodesk.com, Maya 2026 • autodesk.com, Maya 2027", "Collected source links": "https://help.autodesk.com/cloudhelp/2026/ENU/Maya-CharacterAnimation/files/GUID-F386DC20-6C66-40D7-AD40-2C1B66937A71.htm\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Maya-CharacterAnimation/files/GUID-2D7AF32F-814C-43FB-A93B-A8FBD95265A6.htm\nhttps://www.autodesk.com/products/maya/product-details\nhttps://help.autodesk.com/cloudhelp/2027/ENU/Maya-CharacterAnimation/files/GUID-2D7AF32F-814C-43FB-A93B-A8FBD95265A6.htm\nhttps://help.autodesk.com/cloudhelp/2027/ENU/Maya-CharacterAnimation/files/GUID-4098F004-0CB7-4B13-8F9C-146F747D7BAC.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-cfc2687f0eb2", "Operation IDs": "219", "Citation": "help.autodesk.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Maya-CharacterAnimation/files/GUID-F386DC20-6C66-40D7-AD40-2C1B66937A71.htm" }, { "Source ID": "collected-470d05f3b5d6", "Operation IDs": "219", "Citation": "help.autodesk.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Maya-CharacterAnimation/files/GUID-2D7AF32F-814C-43FB-A93B-A8FBD95265A6.htm" }, { "Source ID": "collected-46dac08c4f4f", "Operation IDs": "219", "Citation": "autodesk.com, Maya 2027", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.autodesk.com/products/maya/product-details" }, { "Source ID": "collected-1f3767693a94", "Operation IDs": "219", "Citation": "help.autodesk.com — GUID-2D7AF32F-814C-43FB-A93B-A8FBD95265A6.htm", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Maya-CharacterAnimation/files/GUID-2D7AF32F-814C-43FB-A93B-A8FBD95265A6.htm" }, { "Source ID": "collected-a6bb78aae91a", "Operation IDs": "219", "Citation": "help.autodesk.com — GUID-4098F004-0CB7-4B13-8F9C-146F747D7BAC.htm", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Maya-CharacterAnimation/files/GUID-4098F004-0CB7-4B13-8F9C-146F747D7BAC.htm" } ] }, { "ID": "220", "Software": "Autodesk Maya", "Operation": "MotionMaker Guided Motion Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-adfcbb891cec\ncollected-66dfb64fbe76\ncollected-62cd2a4811a5\ncollected-c9c4f1cedce0\ncollected-bb3bacc1e74e\ncollected-fa5ef075afe4\ncollected-10024112f718\ncollected-6445e472e625\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Maya MotionMaker generates guided character motion through its learned motion model.", "Previous record: rationale": "Current placement unresolved. Only motion and timing applies (A=1). The denominator A−1 is zero at every setting.", "Previous record: calculation": "No current numeric placement; the latest record states the remaining requirement. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Autodesk's own MotionMaker model (Autodesk Research): an autoregressive motion generator of multiple neural networks that predicts the next pose frame by frame, trained on Autodesk motion-capture data of human performers and dogs; horse motion added in Maya 2027. Output is character animation data, with no image or video model involved. Model choice: Yes, by character model. Not established: Network sizes, architecture details beyond autoregressive pose prediction, and the full training set size for later releases. research.autodesk.com • digitalproduction.com, 2026-07-24\nChecked 3 Oct 2026 Maya 2027 (Mar 25, 2026) added horse motion to MotionMaker. Maya 2027.2 (Jul 24, 2026) added Bring Your Own Data: teams train custom MotionMaker models locally from their own motion-capture or keyframe data and use them like the supplied models. With a custom model, the trained knowledge comes from the designer's own animation. X stays 2. CG Channel, Maya 2027, Mar 25, 2026 • Digital Production, Maya 2027.2, Jul 24, 2026\nDesigner side, 3 Oct 2026 Builds structure from the designer’s input. Decided by a trained model (learned motion model). Designer controls: path keyframes (path locator); action tags (e.g. jump); Auto Speed Ramp; playback range; prior one second of animation. Returns one result; editable structure: keyframes on the character rig; can sit on a separate animation layer. “The generated motion is all baked into keyframes, so any of the usual tools can be used to edit the resulting animation.” Source • Screenshot", "Collected source links": "https://blogs.autodesk.com/media-and-entertainment/2025/06/04/meet-motionmaker/\nhttps://blogs.autodesk.com/media-and-entertainment/2025/10/15/ai-powered-motionmaker-for-animation-your-top-questions-answered/\nhttps://www.research.autodesk.com/blog/motionmaker-mayas-latest-ai-tool-by-research/\nhttps://digitalproduction.com/2026/07/24/maya-2027-2-trains-motionmaker/\nhttps://www.cgchannel.com/2026/03/autodesk-releases-maya-2027-and-maya-creative-2027/\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Maya-MotionMaker/files/GUID-473B9D09-124C-4A5C-94F8-F3B4B89FAB18.html\nhttps://help.autodesk.com/cloudhelp/2027/ENU/Maya-MotionMaker/files/GUID-1b2094be-f4c9-4b64-8ca2-30a13f3d6a53.html\nhttps://help.autodesk.com/cloudhelp/2027/ENU/Maya-MotionMaker/files/GUID-F4822BDE-3235-4BBD-A311-100802FC2118.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-adfcbb891cec", "Operation IDs": "220", "Citation": "blogs.autodesk.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blogs.autodesk.com/media-and-entertainment/2025/06/04/meet-motionmaker/" }, { "Source ID": "collected-66dfb64fbe76", "Operation IDs": "220", "Citation": "blogs.autodesk.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blogs.autodesk.com/media-and-entertainment/2025/10/15/ai-powered-motionmaker-for-animation-your-top-questions-answered/" }, { "Source ID": "collected-62cd2a4811a5", "Operation IDs": "220", "Citation": "research.autodesk.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.research.autodesk.com/blog/motionmaker-mayas-latest-ai-tool-by-research/" }, { "Source ID": "collected-c9c4f1cedce0", "Operation IDs": "220", "Citation": "digitalproduction.com, 2026-07-24", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://digitalproduction.com/2026/07/24/maya-2027-2-trains-motionmaker/" }, { "Source ID": "collected-bb3bacc1e74e", "Operation IDs": "220", "Citation": "CG Channel, Maya 2027, Mar 25, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/03/autodesk-releases-maya-2027-and-maya-creative-2027/" }, { "Source ID": "collected-fa5ef075afe4", "Operation IDs": "220", "Citation": "help.autodesk.com — GUID-473B9D09-124C-4A5C-94F8-F3B4B89FAB18.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Maya-MotionMaker/files/GUID-473B9D09-124C-4A5C-94F8-F3B4B89FAB18.html" }, { "Source ID": "collected-10024112f718", "Operation IDs": "220", "Citation": "help.autodesk.com — GUID-1b2094be-f4c9-4b64-8ca2-30a13f3d6a53.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Maya-MotionMaker/files/GUID-1b2094be-f4c9-4b64-8ca2-30a13f3d6a53.html" }, { "Source ID": "collected-6445e472e625", "Operation IDs": "220", "Citation": "help.autodesk.com — GUID-F4822BDE-3235-4BBD-A311-100802FC2118.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Maya-MotionMaker/files/GUID-F4822BDE-3235-4BBD-A311-100802FC2118.html" } ] }, { "ID": "221", "Software": "Meshroom (AliceVision)", "Operation": "AliceVision Photogrammetry Graph", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c26b1d09cac3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Meshroom evaluates the classical AliceVision photogrammetry graph with learned extensions excluded.", "Previous record: rationale": "The defined geometric reconstruction and optimization pathway is conventional. Optional MeshroomHub models are not attributed to the baseline graph.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by reconstruction from captured images. Designer controls: nodes/parameters (photogrammetry graph); input photographs. Returns one result; editable structure: point cloud, mesh, textured mesh. “Standard computer vision pipelines such as photogrammetry, camera tracking, HDR panorama, Lidar Meshing, Raw image files conversion and color calibration are now unified” Source", "Collected source links": "https://github.com/alicevision/Meshroom/releases/tag/v2025.1.0", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c26b1d09cac3", "Operation IDs": "221", "Citation": "GitHub · Meshroom v2025.1.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/alicevision/Meshroom/releases/tag/v2025.1.0" } ] }, { "ID": "222", "Software": "Meshy", "Operation": "Meshy 7 Texture Base Color Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-be0923079dd1\ncollected-e1de69cc4205\ncollected-46b3ef8bd941\ncollected-abd38dc6fb14", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Meshy 7 Texture generates base-color appearance for a supplied mesh.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 2 unknown and 0 setting-dependent categories among 4 applicable. The recorded uncertainty gives 3.0–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=2; conditional=0; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. Upper: 2 + 3 × 3 ÷ (4 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Meshy 7 texturing). Designer controls: supplied mesh; text prompt; image reference; multi-view images (1-4); model choice; keep original UV; PBR maps toggle; texture resolution (2k/4k/8k); remove lighting; art style. Returns count not documented; editable structure: textured mesh with base-color map; glb, fbx, obj, usdz. “Generate PBR Maps (metallic, roughness, normal) in addition to the base color.” Source • Screenshot", "Collected source links": "https://help.meshy.ai/en/articles/9996850-how-to-use-the-text-to-texture-feature\nhttps://docs.meshy.ai/en/webapp/changelog\nhttps://docs.meshy.ai/en/api/retexture\nhttps://cdn.meshy.ai/webapp-assets/help/59d9ee9adc516ee0-how-to-use-the-text-to-texture-feature.webp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-be0923079dd1", "Operation IDs": "222", "Citation": "Meshy · Prompted texture generation on an existing mesh", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.meshy.ai/en/articles/9996850-how-to-use-the-text-to-texture-feature" }, { "Source ID": "collected-e1de69cc4205", "Operation IDs": "222", "Citation": "Meshy current texture release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/webapp/changelog" }, { "Source ID": "collected-46b3ef8bd941", "Operation IDs": "222, 223", "Citation": "docs.meshy.ai — retexture", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/retexture" }, { "Source ID": "collected-abd38dc6fb14", "Operation IDs": "222", "Citation": "cdn.meshy.ai — 59d9ee9adc516ee0-how-to-use-the-text-to-texture-feature.webp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.meshy.ai/webapp-assets/help/59d9ee9adc516ee0-how-to-use-the-text-to-texture-feature.webp" } ] }, { "ID": "223", "Software": "Meshy", "Operation": "Meshy 7.1 Single Image to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-aa9f1d23ecfb\ncollected-f27c1d9b1570\ncollected-c044c6a87c6e\ncollected-7fe7157d6385\ncollected-82720073e36a\ncollected-46b3ef8bd941\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Meshy 7.1 predicts a mesh through the single-image geometry-generation route.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=1; conditional=2; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Meshy 7.1 image-to-3D). Designer controls: image upload; multi-view (up to 4 images); model choice; geometry resolution (standard/2k/4k); topology (triangle/quad); remesh polycount; pose (A-pose/T-pose); texture toggle; PBR; image enhancement; batch mode (up to 10 images). Returns count not documented; editable structure: mesh; FBX, OBJ, GLB, USDZ, STL, BLEND, 3MF, DXF. “Upload a single image and Meshy rebuilds it as a fully textured 3D model in about a minute.” Source • Screenshot", "Collected source links": "https://www.meshy.ai/features/image-to-3d\nhttps://docs.meshy.ai/en/api/changelog\nhttps://cdn.meshy.ai/ti_w:3840,q:75,f:webp/landing-assets/home/bg_landing_home_image3d_UI.png\nhttps://docs.meshy.ai/en/api/asset-retention\nhttps://docs.meshy.ai/en/api/image-to-3d\nhttps://docs.meshy.ai/en/api/retexture", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-aa9f1d23ecfb", "Operation IDs": "223", "Citation": "Meshy · Image-to-3D", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.meshy.ai/features/image-to-3d" }, { "Source ID": "collected-f27c1d9b1570", "Operation IDs": "223, 224", "Citation": "Meshy API changelog", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/changelog" }, { "Source ID": "collected-c044c6a87c6e", "Operation IDs": "223", "Citation": "cdn.meshy.ai — bg_landing_home_image3d_UI.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.meshy.ai/ti_w:3840,q:75,f:webp/landing-assets/home/bg_landing_home_image3d_UI.png" }, { "Source ID": "collected-7fe7157d6385", "Operation IDs": "223, 224", "Citation": "docs.meshy.ai — asset-retention", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/asset-retention" }, { "Source ID": "collected-82720073e36a", "Operation IDs": "223", "Citation": "docs.meshy.ai — image-to-3d", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/image-to-3d" }, { "Source ID": "collected-46b3ef8bd941", "Operation IDs": "222, 223", "Citation": "docs.meshy.ai — retexture", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/retexture" } ] }, { "ID": "224", "Software": "Meshy", "Operation": "Meshy 7.1 Text to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d5f62cd6fdda\ncollected-f27c1d9b1570\ncollected-697b3b4ecfc4\ncollected-7fe7157d6385\ncollected-06a2bb12e6d7\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Meshy 7.1 generates a mesh through its text-to-geometry route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 3 model-decided, 0 unknown and 0 setting-dependent categories among 3 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=3; unknown=0; conditional=0; A=3. Lower: 2 + 3 × 0 ÷ (3 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Meshy 7.1 text-to-3D). Designer controls: prompt (up to 800 characters); model choice; model type (High Detail/Smart Topology); pose; number of generations; geometry resolution (API). Returns several to choose from (Number of Generations; count not stated); editable structure: textured mesh; 8 export formats. “Set the Number of Generations to produce multiple variations at once. Each generation costs 20 credits.” Source • Screenshot", "Collected source links": "https://www.meshy.ai/features/text-to-3d\nhttps://docs.meshy.ai/en/api/changelog\nhttps://cdn.meshy.ai/ti_w:3840,q:75,f:webp/landing-assets/home/bg_landing_home_text3d_UI.png\nhttps://docs.meshy.ai/en/api/asset-retention\nhttps://docs.meshy.ai/en/api/text-to-3d", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d5f62cd6fdda", "Operation IDs": "224", "Citation": "Meshy · Text-to-3D", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.meshy.ai/features/text-to-3d" }, { "Source ID": "collected-f27c1d9b1570", "Operation IDs": "223, 224", "Citation": "Meshy API changelog", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/changelog" }, { "Source ID": "collected-697b3b4ecfc4", "Operation IDs": "224", "Citation": "cdn.meshy.ai — bg_landing_home_text3d_UI.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.meshy.ai/ti_w:3840,q:75,f:webp/landing-assets/home/bg_landing_home_text3d_UI.png" }, { "Source ID": "collected-7fe7157d6385", "Operation IDs": "223, 224", "Citation": "docs.meshy.ai — asset-retention", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/asset-retention" }, { "Source ID": "collected-06a2bb12e6d7", "Operation IDs": "224", "Citation": "docs.meshy.ai — text-to-3d", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.meshy.ai/en/api/text-to-3d" } ] }, { "ID": "225", "Software": "Microsoft PowerPoint", "Operation": "Copilot agent arrangement of generated visual assets", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-aa3405d631c0\ncollected-b6865e4fbbd3\ncollected-f1167ea1542d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "PowerPoint Copilot's selected Claude route decides and authors the arrangement of supplied visual assets.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. Within the scoped arrangement, the model determines placement, size and grouping of supplied assets on slides, and asset content comes from earlier, separately scoped generation. That places the comparison nearer region determination than synthesis, at 6–10; the narrowing follows from the scoped result together with the documented agentic edit route.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. Microsoft support pages (2026) describe editing with Copilot, formerly marketed as Agent Mode, with a PowerPoint model switcher offering Claude or ChatGPT, and state that editing with Copilot uses Anthropic models in a custom multi-agent approach. The 22 April 2026 Microsoft 365 blog describes app-native actions in presentations. None of these pages describe how image positions and sizes are computed. Not established: Whether positions come from template placeholders or model-computed coordinates; which model the Auto setting selects. Microsoft Tech Community pages on Edit with Copilot did not render for fetching. Previous reason: Learned composition survives native slide commands, supporting 2–10. This excludes separately invoked image generation and does not equate all Copilot actions. support.microsoft.com • support.microsoft.com • microsoft.com, 2026-04-22", "Collected source links": "https://www.microsoft.com/en-us/microsoft-365/blog/2026/04/22/copilots-agentic-capabilities-in-word-excel-and-powerpoint-are-generally-available/\nhttps://support.microsoft.com/en-us/office/choose-your-model-when-editing-with-copilot-in-powerpoint\nhttps://support.microsoft.com/en-us/office/copilot-frontier/get-started-editing-with-copilot-in-office", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-aa3405d631c0", "Operation IDs": "225", "Citation": "Microsoft · Copilot Agent Mode release", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.microsoft.com/en-us/microsoft-365/blog/2026/04/22/copilots-agentic-capabilities-in-word-excel-and-powerpoint-are-generally-available/" }, { "Source ID": "collected-b6865e4fbbd3", "Operation IDs": "225", "Citation": "Microsoft Copilot model choice", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.microsoft.com/en-us/office/choose-your-model-when-editing-with-copilot-in-powerpoint" }, { "Source ID": "collected-f1167ea1542d", "Operation IDs": "225", "Citation": "support.microsoft.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.microsoft.com/en-us/office/copilot-frontier/get-started-editing-with-copilot-in-office" } ] }, { "ID": "226", "Software": "Microsoft PowerPoint", "Operation": "Designer layout suggestions for generated imagery", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-390f210145ea\ncollected-18477d274760\ncollected-e9b9f66896e0\ncollected-4e2413035924", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "PowerPoint Designer recommends and applies slide layouts.", "Previous record: rationale": "Designer-specific documentation now separates the route: trained models rank and select among prescribed layouts and set the image crop, while placement geometry comes from the layouts themselves. Learned determination confined to a crop region and a choice among prescribed arrangements sits between learned region determination (REF10) and a learned model setting what a fixed operator applies (REF12).", "Previous record: calculation": "X = 10 + t × (12 − 10); t = 0–1; X = 10–12. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X Unresolved → 10.0–12.0. Microsoft's Azure blog (26 March 2020) states Designer runs machine learning models for image categorization, slide structure analysis and suggestion ranking, and formats content from professionally designed templates; Microsoft's 18 June 2019 post states Designer's AI chooses layouts and intelligently crops images, with Design Ideas built from Slide Master layouts. Since May 2025 the feature appears as Design Suggestions for Copilot plans; the current support page still describes matching content to professionally designed layouts. Not established: The model disclosures date from 2019 and 2020; documentation does not state whether the 2025 Design Suggestions rename changed the ranking or cropping models, or how much ranking outweighs fixed template rules within the 10 to 12 interval. Previous reason: The source does not resolve learned ranking, template matching and prescribed placement for the selected route. Copilot's separate model evidence cannot establish Designer's branch. azure.microsoft.com, 2020-03-26 • microsoft.com, 2019-06-18 • support.microsoft.com, current page, undated", "Collected source links": "https://support.microsoft.com/en-us/office/create-professional-slide-layouts-with-designer-53c77d7b-dc40-45c2-b684-81415eac0617\nhttps://support.microsoft.com/en-us/powerpoint/create-professional-slide-layouts-with-designer\nhttps://azure.microsoft.com/en-us/blog/how-azure-machine-learning-enables-powerpoint-designer/\nhttps://www.microsoft.com/en-us/microsoft-365/blog/2019/06/18/powerpoint-ai-upgrade-designer-major-milestone-1-billion-slides/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-390f210145ea", "Operation IDs": "226", "Citation": "Microsoft · PowerPoint Designer (Design Suggestions)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.microsoft.com/en-us/office/create-professional-slide-layouts-with-designer-53c77d7b-dc40-45c2-b684-81415eac0617" }, { "Source ID": "collected-18477d274760", "Operation IDs": "226", "Citation": "Microsoft Designer visual arrangement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.microsoft.com/en-us/powerpoint/create-professional-slide-layouts-with-designer" }, { "Source ID": "collected-e9b9f66896e0", "Operation IDs": "226", "Citation": "azure.microsoft.com, 2020-03-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://azure.microsoft.com/en-us/blog/how-azure-machine-learning-enables-powerpoint-designer/" }, { "Source ID": "collected-4e2413035924", "Operation IDs": "226", "Citation": "microsoft.com, 2019-06-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.microsoft.com/en-us/microsoft-365/blog/2019/06/18/powerpoint-ai-upgrade-designer-major-milestone-1-billion-slides/" } ] }, { "ID": "227", "Software": "Microsoft PowerPoint", "Operation": "Master-layout revision of slides containing generated visuals", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b5faf8c7a343\ncollected-66a2a0d1ee4c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "PowerPoint propagates a manually revised Slide Master layout and visual attributes.", "Previous record: rationale": "Native layout inheritance is prescribed computation without a trained planner in this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: Slide Master view (View tab); layouts; placeholders (move, hide, show); theme colors; fonts; effects; background; Close Master View. Returns one result; editable structure: slide master, layouts, placeholders. “When you edit the slide master, all slides that are based on that master will contain those changes.” Source • Screenshot", "Collected source links": "https://support.microsoft.com/en-us/office/what-is-a-slide-master-b9abb2a0-7aef-4257-a14e-4329c904da54\nhttps://support.microsoft.com/en-us/powerpoint/training/customize-a-slide-master", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b5faf8c7a343", "Operation IDs": "227", "Citation": "Microsoft · PowerPoint Slide Master", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.microsoft.com/en-us/office/what-is-a-slide-master-b9abb2a0-7aef-4257-a14e-4329c904da54" }, { "Source ID": "collected-66a2a0d1ee4c", "Operation IDs": "227", "Citation": "Microsoft master and dependent layout edits", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.microsoft.com/en-us/powerpoint/training/customize-a-slide-master" } ] }, { "ID": "228", "Software": "Midjourney", "Operation": "V8.2 Draft Image Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1c89dd45d821\ncollected-fcca2cab092b\ncollected-969be1372c36\ncollected-830ffb4382c5", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Midjourney V8.2 Draft generates draft images through its selected generation mode.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Midjourney V8.2, draft quality). Designer controls: prompt; Draft mode toggle (lightning button or --draft); Vary; Remix; Enhance. Returns several to choose from (24); a flat image: low-resolution drafts. “Each draft mode generation creates 24 images at a lower resolution and quality.” Source • Screenshot Quote taken from a search result for the vendor page.", "Collected source links": "https://updates.midjourney.com/\nhttps://docs.midjourney.com/hc/en-us/articles/35577175650957-Draft-Conversational-Modes\nhttps://updates.midjourney.com/draft-mode-for-v8-1-and-new-feature-previews/\nhttps://storage.ghost.io/c/95/61/95616ddb-b827-4269-ad22-ff4a6b368a1f/content/images/2026/06/Screenshot-2026-06-16-at-2.53.46---PM.jpeg", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1c89dd45d821", "Operation IDs": "228", "Citation": "Midjourney · Draft Mode", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://updates.midjourney.com/" }, { "Source ID": "collected-fcca2cab092b", "Operation IDs": "228", "Citation": "Midjourney · Draft modes and V7 Enhance", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/35577175650957-Draft-Conversational-Modes" }, { "Source ID": "collected-969be1372c36", "Operation IDs": "228", "Citation": "updates.midjourney.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://updates.midjourney.com/draft-mode-for-v8-1-and-new-feature-previews/" }, { "Source ID": "collected-830ffb4382c5", "Operation IDs": "228", "Citation": "storage.ghost.io — Screenshot-2026-06-16-at-2.53.46---PM.jpeg", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://storage.ghost.io/c/95/61/95616ddb-b827-4269-ad22-ff4a6b368a1f/content/images/2026/06/Screenshot-2026-06-16-at-2.53.46---PM.jpeg" } ] }, { "ID": "229", "Software": "Midjourney", "Operation": "V8.2 Edit Model Masked Image Revision", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-11cbb91d4cc9\ncollected-25f3f3ee9a9f\ncollected-63dc50eacb33", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Midjourney V8.2 Edit Model generates revised content for a masked image region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=4; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Midjourney (Aug 27, 2026): the same V8.2 edit model performs inpainting and outpainting and edits from written instructions without hand masking. X stays 2. Midjourney, Edit model for V8, Aug 27, 2026\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Midjourney Edit Model, inpainting). Designer controls: Erase brush; Restore brush; layer selection; prompt; Submit Edit; canvas expansion (outpainting). Returns count not documented; a flat image. “Changing specific areas of your images (inpainting) or expand your canvas (outpainting)” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://docs.midjourney.com/hc/en-us/articles/32764383466893-Editor\nhttps://docs.midjourney.com/hc/en-us/articles/48495453462797-Edit-Model\nhttps://updates.midjourney.com/edit-model-for-v8/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-11cbb91d4cc9", "Operation IDs": "229", "Citation": "Midjourney · Editor guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/32764383466893-Editor" }, { "Source ID": "collected-25f3f3ee9a9f", "Operation IDs": "229, 232", "Citation": "Midjourney Edit Model compatibility", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/48495453462797-Edit-Model" }, { "Source ID": "collected-63dc50eacb33", "Operation IDs": "229, 230", "Citation": "Midjourney, Edit model for V8, Aug 27, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://updates.midjourney.com/edit-model-for-v8/" } ] }, { "ID": "230", "Software": "Midjourney", "Operation": "V8.2 Edit Model Reference Conditioned Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a6825464ca84\ncollected-a399e6415793\ncollected-63dc50eacb33", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Midjourney V8.2 Edit Model generates an image through its reference-conditioned route.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=2; conditional=2; A=5. Lower: 2 + 3 × 1 ÷ (5 − 1) = 2.75. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.75; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Midjourney (Aug 27, 2026): the V8.2 edit model generates images from up to 4 image references at once and replaces omni-reference; it also edits from written instructions, inpaints and outpaints. Number of results per job not stated. X stays 2. Midjourney, Edit model for V8, Aug 27, 2026\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Midjourney Edit Model, reference-conditioned). Designer controls: reference images (up to 4); prompt; written instructions; personalization; moodboards; style references; --hd. Returns count not documented; a flat image. “Generating images with other images (replacing omni-reference) with up to 4 image references at once” Source", "Collected source links": "https://aiwiki.ai/wiki/midjourney_v7\nhttps://docs.midjourney.com/hc/en-us/articles/36285124473997-Omni-Reference\nhttps://updates.midjourney.com/edit-model-for-v8/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a6825464ca84", "Operation IDs": "230", "Citation": "AI Wiki · Omni Reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://aiwiki.ai/wiki/midjourney_v7" }, { "Source ID": "collected-a399e6415793", "Operation IDs": "230", "Citation": "Midjourney official Omni Reference documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/36285124473997-Omni-Reference" }, { "Source ID": "collected-63dc50eacb33", "Operation IDs": "229, 230", "Citation": "Midjourney, Edit model for V8, Aug 27, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://updates.midjourney.com/edit-model-for-v8/" } ] }, { "ID": "231", "Software": "Midjourney V8.2", "Operation": "Text-to-image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-665902754958\ncollected-398b73946b26\ncollected-1516ca5b880b\ncollected-470550ff8811\ncollected-96cdf8ddb5fc\ncollected-eefc4dd9ae21\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Midjourney V8.2 generates an image through its text-to-image route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Midjourney V8.2). Designer controls: prompt; aspect ratio; stylize; chaos/variety; raw; seed; personalization; moodboards; style references; HD mode. Returns several to choose from (4); a flat image. “By default chaos is set to 0, and Midjourney creates four images based on your prompt.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://updates.midjourney.com/v8-1-is-now-the-default-model/\nhttps://docs.midjourney.com/hc/en-us/articles/32199405667853-Version\nhttps://docs.midjourney.com/hc/en-us/articles/32099348346765-Chaos-Variety\nhttps://web.archive.org/web/20260909060210id_/https://docs.midjourney.com/hc/en-us/articles/33329329805581-Modifying-Your-Creations\nhttps://web.archive.org/web/20260731111101id_/https://docs.midjourney.com/hc/en-us/articles/32799074515213-Remix\nhttps://docs.midjourney.com/hc/en-us/articles/33390732264589-Creating-on-Web", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-665902754958", "Operation IDs": "231", "Citation": "Midjourney · Prompt Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://updates.midjourney.com/v8-1-is-now-the-default-model/" }, { "Source ID": "collected-398b73946b26", "Operation IDs": "231", "Citation": "Midjourney current and historical version definitions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/32199405667853-Version" }, { "Source ID": "collected-1516ca5b880b", "Operation IDs": "231", "Citation": "docs.midjourney.com — 32099348346765-Chaos-Variety", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/32099348346765-Chaos-Variety" }, { "Source ID": "collected-470550ff8811", "Operation IDs": "231", "Citation": "web.archive.org — 33329329805581-Modifying-Your-Creations", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260909060210id_/https://docs.midjourney.com/hc/en-us/articles/33329329805581-Modifying-Your-Creations" }, { "Source ID": "collected-96cdf8ddb5fc", "Operation IDs": "231", "Citation": "web.archive.org — 32799074515213-Remix", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260731111101id_/https://docs.midjourney.com/hc/en-us/articles/32799074515213-Remix" }, { "Source ID": "collected-eefc4dd9ae21", "Operation IDs": "231", "Citation": "docs.midjourney.com — 33390732264589-Creating-on-Web", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/33390732264589-Creating-on-Web" } ] }, { "ID": "232", "Software": "Midjourney", "Operation": "V8.2 Edit Model Whole Image Restyling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.5–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-30d4f289d791\ncollected-25f3f3ee9a9f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Midjourney V8.2 Edit Model performs whole-image restyling.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 2 model-decided, 1 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 3.5–4.25; no midpoint is used.", "Previous record: calculation": "K=2; M=2; unknown=1; conditional=0; A=5. Lower: 2 + 3 × 2 ÷ (5 − 1) = 3.5. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Midjourney Edit Model, restyle). Designer controls: source image; prompt; moodboards; personalization; style references. Returns count not documented; a flat image. “The Edit model includes the ability to convert an image into a different style.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://docs.midjourney.com/hc/en-us/articles/33329788681101-Legacy-Features\nhttps://docs.midjourney.com/hc/en-us/articles/48495453462797-Edit-Model", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-30d4f289d791", "Operation IDs": "232", "Citation": "Midjourney legacy Retexture boundary", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/33329788681101-Legacy-Features" }, { "Source ID": "collected-25f3f3ee9a9f", "Operation IDs": "229, 232", "Citation": "Midjourney Edit Model compatibility", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.midjourney.com/hc/en-us/articles/48495453462797-Edit-Model" } ] }, { "ID": "233", "Software": "Miro", "Operation": "AI Flows Diagram Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–6.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-cc4b477fb44f\ncollected-2a9aa148fa63\ncollected-9ff9de785449\ncollected-9947caf295d3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Miro AI Flows uses a selected language model to construct diagram content and relationships.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The model writes node text and determines the relationships of the diagram from the prompt and board content, which places the comparison nearer synthesis than region determination, at 2–6 within the same envelope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–0.5; X = 2–6. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 2.0–6.0. Miro's AI reference (updated 21 September 2026) lists Create diagram – Flow chart as creating a flow chart from a user prompt and selected board content with GPT-5.4 and GPT-5.4 mini. The Flows public beta has ended (Flows overview, 8 September 2026), Diagrams appear among Flows-supported formats, and the model-selector list (21 September 2026) names Docs, Image formats, Sidekick logic and Instruction blocks, with diagrams absent from it. Not established: Whether node positions come from the model or from automatic layout; whether a Flow's diagram step uses a user-selected model or Miro's default pairing. Previous reason: Learned relational and semantic construction supports 2–10. Workflow routing and native diagram editability do not add scoring terms. help.miro.com, updated 2026-09-21 • help.miro.com, updated 2026-09-08 • help.miro.com, updated 2026-09-08", "Collected source links": "https://help.miro.com/hc/en-us/articles/29484232754578-Select-Your-Own-Model-Beta\nhttps://help.miro.com/hc/en-us/articles/20970362792210-Miro-AI-reference\nhttps://help.miro.com/hc/en-us/articles/29681832191378-Flows-overview\nhttps://help.miro.com/hc/en-us/articles/29687970855442-Flows", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-cc4b477fb44f", "Operation IDs": "233, 234", "Citation": "Miro model selector", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/hc/en-us/articles/29484232754578-Select-Your-Own-Model-Beta" }, { "Source ID": "collected-2a9aa148fa63", "Operation IDs": "233", "Citation": "Miro · Diagram generation operations", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/hc/en-us/articles/20970362792210-Miro-AI-reference" }, { "Source ID": "collected-9ff9de785449", "Operation IDs": "233", "Citation": "help.miro.com, updated 2026-09-08", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/hc/en-us/articles/29681832191378-Flows-overview" }, { "Source ID": "collected-9947caf295d3", "Operation IDs": "233", "Citation": "help.miro.com, updated 2026-09-08", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/hc/en-us/articles/29687970855442-Flows" } ] }, { "ID": "234", "Software": "Miro", "Operation": "AI Diagram Digitization with Current Models", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1436fe49d0e7\ncollected-cc4b477fb44f\ncollected-fd1db2976c96\ncollected-3d3f8fa70460\ncollected-bb3b71ed2539\ncollected-9410cf4d81e9\ncollected-9c61c6a33572\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Miro's documented diagram-digitization route interprets a diagram image and constructs native elements.", "Previous record: rationale": "Recognition and reconstruction of relationships support 8–10. The current digitization route is distinct from earlier unconstrained Sidekicks diagram generation.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). Miro help-centre pages, including the record sources, refused fetch (HTTP 403) on 3 October 2026. Miro's AI diagram page states that the generator proposes structured layouts from prompts or from board content, including imported reference material, and suggests diagram elements; it does not describe an image-constrained digitization mode. Not established: Whether the scoped digitization route limits output to elements present in the source image, or also generates elements and text, is not documented in fetchable pages. The models used for this route under the \"Select Your Own Model\" beta are not confirmed. miro.com, accessed 2026-10-03 • help.miro.com, 2026-10-03", "Collected source links": "https://help.miro.com/hc/en-us/articles/30139627329042-Sidekicks\nhttps://help.miro.com/hc/en-us/articles/29484232754578-Select-Your-Own-Model-Beta\nhttps://miro.com/ai/diagram-ai/\nhttps://help.miro.com/hc/en-us/articles/25275263961874-Miro-Diagrams\nhttps://help.miro.com/api/v2/help_center/en-us/articles/360017730953.json\nhttps://help.miro.com/api/v2/help_center/en-us/articles/28782102127890.json\nhttps://help.miro.com/api/v2/help_center/en-us/articles/360017730793.json", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1436fe49d0e7", "Operation IDs": "234", "Citation": "Miro Sidekicks visual formats", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/hc/en-us/articles/30139627329042-Sidekicks" }, { "Source ID": "collected-cc4b477fb44f", "Operation IDs": "233, 234", "Citation": "Miro model selector", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/hc/en-us/articles/29484232754578-Select-Your-Own-Model-Beta" }, { "Source ID": "collected-fd1db2976c96", "Operation IDs": "234", "Citation": "miro.com, accessed 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://miro.com/ai/diagram-ai/" }, { "Source ID": "collected-3d3f8fa70460", "Operation IDs": "234", "Citation": "help.miro.com, 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/hc/en-us/articles/25275263961874-Miro-Diagrams" }, { "Source ID": "collected-bb3b71ed2539", "Operation IDs": "234", "Citation": "help.miro.com — 360017730953.json", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/api/v2/help_center/en-us/articles/360017730953.json" }, { "Source ID": "collected-9410cf4d81e9", "Operation IDs": "234", "Citation": "help.miro.com — 28782102127890.json", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/api/v2/help_center/en-us/articles/28782102127890.json" }, { "Source ID": "collected-9c61c6a33572", "Operation IDs": "234", "Citation": "help.miro.com — 360017730793.json", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.miro.com/api/v2/help_center/en-us/articles/360017730793.json" } ] }, { "ID": "235", "Software": "MoGe by Microsoft", "Operation": "MoGe 3 CLI Depth Image Prediction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6e0bbff3809c\ncollected-e09a51e46365\ncollected-784711b0e453", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "MoGe 3 predicts the selected depth-image output through its documented CLI route.", "Previous record: rationale": "Learned spatial reconstruction crosses relational determination and fidelity estimation, supporting 8–10 without assuming current Meshroom-plugin compatibility.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). Microsoft released MoGe-3 on 18 August 2026 (arXiv 2607.17967, sparse volumetric refinement); the CLI writes metric point, depth and normal maps, requires an explicit checkpoint, and runs refinement steps on the base prediction. The repository training configuration supervises point maps with affine-invariant global and local regression losses, a normal loss and an L2 mask loss, including terms per refinement step, with no generative stage; the output is a dense geometric estimate rather than a discrete recognition result. Not established: The paper abstract gives no supervision detail beyond the repository configuration. macOS is unsupported for MoGe-3, and Meshroom plugin compatibility with v3 is not documented. Documentation does not place dense geometry estimation within 8–10. github.com, 2026-08-18 • github.com, main branch, accessed 2026-10-03", "Collected source links": "https://github.com/meshroomHub/mrDepthEstimation\nhttps://github.com/microsoft/MoGe\nhttps://github.com/microsoft/MoGe/blob/main/docs/train.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6e0bbff3809c", "Operation IDs": "235", "Citation": "Meshroom Hub · Depth estimation extension", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/meshroomHub/mrDepthEstimation" }, { "Source ID": "collected-e09a51e46365", "Operation IDs": "235", "Citation": "Microsoft · MoGe models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/microsoft/MoGe" }, { "Source ID": "collected-784711b0e453", "Operation IDs": "235", "Citation": "github.com, main branch, accessed 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/microsoft/MoGe/blob/main/docs/train.md" } ] }, { "ID": "236", "Software": "mnml.ai", "Operation": "Render Enhancer Image Output", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3dbc5756094d\ncollected-45876f4fac2e\ncollected-6832cc402bd4\ncollected-2cda80ee2fde\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "mnml.ai Render Enhancer transforms an existing rendering through its learned enhancement service.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–12.0 → 5.0. The current mnml.ai Render Enhancer page describes cleaning noise, rebuilding material detail and regrading light, with a Creativity control setting how much the AI may reinterpret the render and a Geometry control holding massing, openings and camera. mnml.ai docs (16 October 2025) separate it from 4K Upscale, which adds pixels with no change to detail or style. Not established: Model architecture and training objective; whether high Creativity settings extend to re-synthesis beyond detail, toward REF02. Previous reason: The named appearance-enhancement task supports 5–12 between perceptual reconstruction and fixed mapping. General platform generation and separate upscaling are excluded. mnml.ai, not shown • docs.mnml.ai, 2025-10-16", "Collected source links": "https://mnml.ai/app/render-enhancer\nhttps://docs.mnml.ai/docs/4k-upscale-vs-render-enhancer/\nhttps://mnml.ai/product\nhttps://mnml.ai/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3dbc5756094d", "Operation IDs": "236", "Citation": "mnml.ai · Render Generation and Enhancement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://mnml.ai/app/render-enhancer" }, { "Source ID": "collected-45876f4fac2e", "Operation IDs": "236", "Citation": "docs.mnml.ai, 2025-10-16", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.mnml.ai/docs/4k-upscale-vs-render-enhancer/" }, { "Source ID": "collected-6832cc402bd4", "Operation IDs": "236", "Citation": "mnml.ai — product", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://mnml.ai/product" }, { "Source ID": "collected-2cda80ee2fde", "Operation IDs": "236", "Citation": "mnml.ai — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://mnml.ai/" } ] }, { "ID": "237", "Software": "Nerfstudio", "Operation": "Nerfacto Neural Field Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-25ad337a1bb5\ncollected-00ebceb1a3fe\ncollected-c7e71405d2ba", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Nerfacto fits a neural field to posed capture images and retains that scene representation.", "Previous record: rationale": "The current calibration needs a consistent treatment of scene fitting across neural and explicit representations. A neural architecture alone cannot settle the distinction from Gaussian fitting.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Nerfstudio's Nerfacto page describes components trained on the input captures only: camera pose refinement by back-propagation, a proposal sampler using hash encoding with a small fused MLP, and per-image appearance conditioning; no pretrained weights are mentioned. The latest nerfstudio release is v1.1.5 (11 November 2024); v1.1.4 (28 August 2024) added bilateral-grid support to Nerfacto and Splatfacto. Not established: The remaining gap is a calibration decision: whether per-scene fitting of a scene representation counts as learned determination on this scale, applied consistently across neural fields and explicit Gaussians. Vendor or project documentation cannot settle that criterion. docs.nerf.studio, current docs, undated • github.com, v1.1.4, 2024-08-28; latest v1.1.5, 2024-11-11", "Collected source links": "https://radiancefields.com/nerfstudio-1-0-and-splatfacto-released\nhttps://docs.nerf.studio/nerfology/methods/nerfacto.html\nhttps://github.com/nerfstudio-project/nerfstudio/releases", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-25ad337a1bb5", "Operation IDs": "237", "Citation": "Radiance Fields · Nerfacto", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://radiancefields.com/nerfstudio-1-0-and-splatfacto-released" }, { "Source ID": "collected-00ebceb1a3fe", "Operation IDs": "237", "Citation": "Nerfstudio Nerfacto method", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.nerf.studio/nerfology/methods/nerfacto.html" }, { "Source ID": "collected-c7e71405d2ba", "Operation IDs": "237, 238", "Citation": "github.com, v1.1.4, 2024-08-28; latest v1.1.5, 2024-11-11", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/nerfstudio-project/nerfstudio/releases" } ] }, { "ID": "238", "Software": "Nerfstudio", "Operation": "Splatfacto Explicit Gaussian Optimization", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-86cdf9c119df\ncollected-c7e71405d2ba", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Splatfacto optimizes explicit Gaussian parameters from posed images and seed points, with learned extensions excluded.", "Previous record: rationale": "The retained fitted scene requires the same representation-consistent branch decision as gsplat and NeRF fitting. The historical non-neural label is not sufficient for 15.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Splatfacto documentation describes rasterized optimization of 3D Gaussian parameters, initialized from SfM points or randomly, on the gsplat backend, with no pretrained network. nerfstudio v1.1.4 (28 August 2024) added bilateral-grid appearance compensation, which falls within the learned extensions the record excludes; no later Splatfacto change is released (latest v1.1.5, 11 November 2024). Not established: The remaining gap is a calibration decision: whether per-scene fitting of a scene representation counts as learned determination on this scale, applied consistently across neural fields and explicit Gaussians. Vendor or project documentation cannot settle that criterion. docs.nerf.studio, current docs, undated • github.com, v1.0.0, 2024-01-20", "Collected source links": "https://docs.nerf.studio/nerfology/methods/splat.html\nhttps://github.com/nerfstudio-project/nerfstudio/releases", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-86cdf9c119df", "Operation IDs": "238", "Citation": "Nerfstudio Splatfacto representation and export limits", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.nerf.studio/nerfology/methods/splat.html" }, { "Source ID": "collected-c7e71405d2ba", "Operation IDs": "237, 238", "Citation": "github.com, v1.1.4, 2024-08-28; latest v1.1.5, 2024-11-11", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/nerfstudio-project/nerfstudio/releases" } ] }, { "ID": "239", "Software": "Notch", "Operation": "3D Object Material and Transform Graph", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-49a0300781a6\ncollected-a69ea64707d7\ncollected-beacf046abe6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Notch evaluates explicit object transforms and material nodes in the defined scene graph.", "Previous record: rationale": "The scoped graph excludes AI models, learned trackers and denoisers; its native functions are conventional.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: nodes/parameters; Object resource; Layer Index; Material node input; Transform Modifiers input; Local Transform Override; GPU Transform Only. Returns count not documented; editable structure: nodes. “A Material Node can also be input to control the objects material properties.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://manual.notch.one/2026.1/en/docs/reference/nodes/\nhttps://manual.notch.one/2026.1/en/docs/reference/user-interface/properties/\nhttps://manual.notch.one/2026.1/en/docs/reference/nodes/3d/3d-object/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-49a0300781a6", "Operation IDs": "239", "Citation": "Notch · Node Graph", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manual.notch.one/2026.1/en/docs/reference/nodes/" }, { "Source ID": "collected-a69ea64707d7", "Operation IDs": "239", "Citation": "Notch · Node properties", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manual.notch.one/2026.1/en/docs/reference/user-interface/properties/" }, { "Source ID": "collected-beacf046abe6", "Operation IDs": "239", "Citation": "manual.notch.one — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manual.notch.one/2026.1/en/docs/reference/nodes/3d/3d-object/" } ] }, { "ID": "240", "Software": "NVIDIA Audio2Face-3D in Autodesk Maya ACE", "Operation": "Audio2Face Diffusion to Maya Animation Curves", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5ef17e3a3661\ncollected-0267e24c24a1\ncollected-b3128aa6e943\ncollected-0a015f96ed18\ncollected-b4edbab97831", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Audio2Face Diffusion generates facial animation that is mapped to Maya animation curves.", "Previous record: rationale": "Current placement unresolved. Shape was disputed as not applicable versus model-decided. It cannot be treated as kept-versus-model unknown. With final fine detail not applicable: shape not applicable gives A=1/no value; shape model-decided gives A=2, K=0 and X=2.0.", "Previous record: calculation": "No current numeric placement; the latest record states the remaining requirement. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: NVIDIA Audio2Face-3D v3.0 (diffusion-based, multi-identity), released on Hugging Face as nvidia/Audio2Face-3D-v3.0 on 2025-09-24; supported in Maya ACE v2.0.0 (2025-09-24). Model choice: Yes. Not established: The plugin outputs blendshape weights, face and tongue geometry, jaw transform and eye rotation; the best-practices guide lists a Bake Animation section, but the page text describing how keys are written to Maya animation curves was not retrieved. github.com, 2025-09-24 • github.com • huggingface.co, 2025-09-24\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Audio2Face-3D diffusion, audio to facial motion). Designer controls: audio file; model choice (model card); output type (mesh deformation, object transforms or blendshape weights); interactive parameters (tuning guide); bake animation to keyframes. Returns continuous live output; editable structure: Maya keyframes on blendshape weights, mesh deformation or transforms; FBX export. “Export the final animation as native Maya keyframes or in the FBX format for use in other applications.” Source • Screenshot", "Collected source links": "https://github.com/NVIDIA/Audio2Face-3D\nhttps://github.com/NVIDIA/Maya-ACE/releases/tag/v2.0.0\nhttps://github.com/NVIDIA/Maya-ACE/blob/release/2.0/docs/concepts.md\nhttps://huggingface.co/nvidia/Audio2Face-3D-v3.0\nhttps://github.com/NVIDIA/Audio2Face-3D/raw/release/media/Mace_banner.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5ef17e3a3661", "Operation IDs": "240", "Citation": "GitHub · Audio2Face-3D", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/NVIDIA/Audio2Face-3D" }, { "Source ID": "collected-0267e24c24a1", "Operation IDs": "240", "Citation": "github.com, 2025-09-24", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/NVIDIA/Maya-ACE/releases/tag/v2.0.0" }, { "Source ID": "collected-b3128aa6e943", "Operation IDs": "240", "Citation": "github.com — concepts.md", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/NVIDIA/Maya-ACE/blob/release/2.0/docs/concepts.md" }, { "Source ID": "collected-0a015f96ed18", "Operation IDs": "240", "Citation": "huggingface.co, 2025-09-24", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://huggingface.co/nvidia/Audio2Face-3D-v3.0" }, { "Source ID": "collected-b4edbab97831", "Operation IDs": "240", "Citation": "github.com — Mace_banner.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/NVIDIA/Audio2Face-3D/raw/release/media/Mace_banner.png" } ] }, { "ID": "241", "Software": "NVIDIA DLSS in Real Time Engines", "Operation": "DLSS 4.5 Super Resolution Render Pass", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "7.3", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0edb8ca59fe8\ncollected-f52730689f39\ncollected-2107c5fe6690\ncollected-9c1daf23f79b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "DLSS 4.5 Super Resolution reconstructs the rendered image using learned spatial and temporal processing.", "Previous record: rationale": "Rule B (3 Oct 2026): vendor documentation states fidelity-oriented training or non-generative intent for an operation that creates missing samples, which follows the Lightroom Super Resolution comparison: t = 0.75 between 5 and 8, sensitivity t = 0.5–1 (6.5–8).", "Previous record: calculation": "X = 5 + 0.75 × (8 − 5) = 7.25. Sensitivity t = 0.5–1 gives 6.5–8.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 7.3. NVIDIA's DLSS 4.5 announcement (6 Jan 2026) states the second-generation transformer Super Resolution model was trained on an expanded high-fidelity dataset and trains and infers in linear space, which NVIDIA calls the game engine's native ground truth. NVIDIA's DLSS 4 technical page measures the model against ground-truth renders with hundreds of samples per pixel, and NVIDIA's Ada white paper (Oct 2022) describes the super-resolution model as trained to predict a 16K rendered reference image from the prediction-reference difference. Not established: Loss terms of the DLSS 4.5 transformer (pixel, perceptual or adversarial components). Previous reason: The restoration task spans fidelity and perceptual reconstruction, supporting 5–8. Frame Generation and Ray Reconstruction are separately scoped. nvidia.com, 6 Jan 2026 • research.nvidia.com, 2025 • images.nvidia.com, Oct 2022 (PDF metadata)", "Collected source links": "https://www.nvidia.com/en-us/geforce/technologies/dlss/\nhttps://research.nvidia.com/labs/adlr/DLSS4/\nhttps://www.nvidia.com/en-us/geforce/news/dlss-4-5-dynamic-multi-frame-gen-6x-2nd-gen-transformer-super-res/\nhttps://images.nvidia.com/aem-dam/Solutions/geforce/ada/ada-lovelace-architecture/nvidia-ada-gpu-science.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0edb8ca59fe8", "Operation IDs": "241", "Citation": "NVIDIA · NVIDIA DLSS 4 Render Pass", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.nvidia.com/en-us/geforce/technologies/dlss/" }, { "Source ID": "collected-f52730689f39", "Operation IDs": "241", "Citation": "NVIDIA Research · DLSS 4 mechanisms", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.nvidia.com/labs/adlr/DLSS4/" }, { "Source ID": "collected-2107c5fe6690", "Operation IDs": "241", "Citation": "nvidia.com, 6 Jan 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.nvidia.com/en-us/geforce/news/dlss-4-5-dynamic-multi-frame-gen-6x-2nd-gen-transformer-super-res/" }, { "Source ID": "collected-9c1daf23f79b", "Operation IDs": "241", "Citation": "images.nvidia.com, Oct 2022 (PDF metadata)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://images.nvidia.com/aem-dam/Solutions/geforce/ada/ada-lovelace-architecture/nvidia-ada-gpu-science.pdf" } ] }, { "ID": "242", "Software": "NVIDIA Omniverse (Kit SDK)", "Operation": "USD Asset Composition", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-26b486353e3e\ncollected-ed54c7279f4d\ncollected-19547633252d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Omniverse composes USD layers, references and variants through the native composition system.", "Previous record: rationale": "Prescribed scene composition contains no trained content generator within this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: sublayers; references; payloads; variant sets; inherits; specializes. Returns one result; editable structure: USD stage of layers. “OpenUSD's composition functionality incorporates a variety of composition arcs to support content reuse.” Source", "Collected source links": "https://docs.nvidia.com/omniverse/index.html\nhttps://docs.omniverse.nvidia.com/usd/latest/learn-openusd/independent/modularity-guide.html\nhttps://docs.omniverse.nvidia.com/usd/latest/learn-openusd/independent/modularity-guide/content-reuse.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-26b486353e3e", "Operation IDs": "242", "Citation": "NVIDIA · Omniverse USD Composition", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.nvidia.com/omniverse/index.html" }, { "Source ID": "collected-ed54c7279f4d", "Operation IDs": "242", "Citation": "NVIDIA · USD composition and asset hierarchy", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.omniverse.nvidia.com/usd/latest/learn-openusd/independent/modularity-guide.html" }, { "Source ID": "collected-19547633252d", "Operation IDs": "242", "Citation": "docs.omniverse.nvidia.com — content-reuse.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.omniverse.nvidia.com/usd/latest/learn-openusd/independent/modularity-guide/content-reuse.html" } ] }, { "ID": "243", "Software": "ON1 Photo RAW 2026", "Operation": "ON1 Generative Erase", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–4.25", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3128b5258c6c\ncollected-c3098fd02b0a\ncollected-3190431dff1b\ncollected-8e1c8ee36253\ncollected-c532c02c31b8", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ON1 Local Gen AI Erase synthesizes replacement content in the selected region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 2 model-decided, 3 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–4.25; no midpoint is used.", "Previous record: calculation": "K=0; M=2; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 3 ÷ (5 − 1) = 4.25. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Two options: an unnamed ON1 local model (Local Gen AI, scaled-down diffusion, runs on the user's GPU) or a Stability AI cloud model reached through the Stability AI API with the user's own key; the Stability model name or endpoint is not stated. Model choice: Yes. Not established: Neither the ON1 local model nor the Stability AI model is named or versioned. on1.com, 2024-09-17 • scottdavenportphoto.com • on1.com, 2025-12-16\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (generative fill of brushed area). Designer controls: brush; local or cloud rendering (cloud for subscribers and MAX owners). Returns count not documented; form not documented. “lets you remove unwanted objects, fill in missing details, and fix distractions in your photos in seconds—all with just a simple brushstroke” Source", "Collected source links": "https://on1.clickhelp.co/articles/#!photo-raw-2026-user-guide/generative-eraser\nhttps://www.on1.com/blog/how-generative-ai-is-revolutionizing-photography-new-tools-in-on1-photo-raw-2025/\nhttps://scottdavenportphoto.com/blog/how-to-setup-stability-ai-for-on1-photo-raw\nhttps://www.on1.com/blog/on1-photo-raw-2026-2-is-here-faster-smarter-and-even-more-powerful/\nhttps://www.on1.com/videos/select-and-remove-distractions-with-generative-erase/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3128b5258c6c", "Operation IDs": "243", "Citation": "ON1 · Generative Eraser modes and separate layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://on1.clickhelp.co/articles/#!photo-raw-2026-user-guide/generative-eraser" }, { "Source ID": "collected-c3098fd02b0a", "Operation IDs": "243", "Citation": "on1.com, 2024-09-17", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/blog/how-generative-ai-is-revolutionizing-photography-new-tools-in-on1-photo-raw-2025/" }, { "Source ID": "collected-3190431dff1b", "Operation IDs": "243", "Citation": "scottdavenportphoto.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://scottdavenportphoto.com/blog/how-to-setup-stability-ai-for-on1-photo-raw" }, { "Source ID": "collected-8e1c8ee36253", "Operation IDs": "243", "Citation": "on1.com, 2025-12-16", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/blog/on1-photo-raw-2026-2-is-here-faster-smarter-and-even-more-powerful/" }, { "Source ID": "collected-c532c02c31b8", "Operation IDs": "243", "Citation": "www.on1.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/videos/select-and-remove-distractions-with-generative-erase/" } ] }, { "ID": "244", "Software": "ON1 Photo RAW 2026", "Operation": "ON1 NoNoise AI", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f505eab0ed58\ncollected-d919551771bd\ncollected-c72c8499fc24", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ON1 NoNoise AI applies the learned denoising route, excluding the conventional Classic tab.", "Previous record: rationale": "The trained noise-removal responsibility matches fidelity restoration. Retained settings do not give an editability-based X adjustment.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). ON1 released NoNoise AI 2027 as a standalone app and plug-in on 1 October 2026 with rebuilt noise-reduction models and a selector of subject-tuned models (wildlife, portraits, astrophotography, scanned film); ON1 Photo RAW 2027, which carries it, is expected in early November 2026. The scoped ON1 Photo RAW 2026 route is unchanged, and ON1 publishes no training task for either model generation. Not established: Training data and objective for the 2026 and 2027 NoNoise models are not published, and whether the subject-tuned 2027 models add generative detail is not stated. The ON1 Photo RAW 2026 user-guide page returned no text for fetch. on1.com, 2026-10-02 • on1.com, 2026-09", "Collected source links": "https://on1.clickhelp.co/articles/#!photo-raw-2026-user-guide/noise-sharpening\nhttps://www.on1.com/blog/september-2026-on1-recap-nonoise-ai-photo-raw-2027/\nhttps://www.on1.com/blog/introducing-on1-photo-raw-2027-a-new-raw-processing-engine-improved-nonoise-ai-an-ai-culling-assistant-and-more/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f505eab0ed58", "Operation IDs": "244", "Citation": "ON1 · NoNoise neural model and Apply Adjust controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://on1.clickhelp.co/articles/#!photo-raw-2026-user-guide/noise-sharpening" }, { "Source ID": "collected-d919551771bd", "Operation IDs": "244", "Citation": "on1.com, 2026-10-02", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/blog/september-2026-on1-recap-nonoise-ai-photo-raw-2027/" }, { "Source ID": "collected-c72c8499fc24", "Operation IDs": "244", "Citation": "on1.com, 2026-09", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/blog/introducing-on1-photo-raw-2027-a-new-raw-processing-engine-improved-nonoise-ai-an-ai-culling-assistant-and-more/" } ] }, { "ID": "245", "Software": "ON1 Photo RAW 2026", "Operation": "ON1 Resize AI", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9c137ff2ff74\ncollected-9b7a9e08f71a\ncollected-5589b1b2712c\ncollected-a88c7287bcc6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ON1 Resize AI Highest Quality reconstructs enlarged imagery through its generative model.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. ON1's Resize AI product page states that the Highest Quality model uses diffusion to rebuild lost detail, and the Photo RAW 2026 press release (2 Sep 2025) describes it as a generative AI model that recreates fine details, restores sharpness and removes noise. ON1's Photo RAW page now announces Photo RAW 2027 for fall 2026, with new diffusion-based models in the integrated Resize AI. Not established: Identity and training objective of the diffusion model; whether Photo RAW 2027 replaces the 2026 Highest Quality model. Previous reason: Detail reconstruction spans the perceptual and fidelity references, supporting 5–8. The standard faster mode is outside this record. on1.com, Resize AI 2026 • on1.com, 2 Sep 2025 • on1.com, Photo RAW 2027 page, coming fall 2026", "Collected source links": "https://www.on1.com/press/on1-announces-photo-raw-2026-with-advanced-ai-tools-masking-layers-and-new-creative-filters/\nhttps://on1.clickhelp.co/articles/#!photo-raw-2026-user-guide/basic-resize-ai-workflow\nhttps://www.on1.com/products/resize-ai/\nhttps://www.on1.com/products/photo-raw/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9c137ff2ff74", "Operation IDs": "245", "Citation": "ON1 · Resize AI 2026 model choices", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/press/on1-announces-photo-raw-2026-with-advanced-ai-tools-masking-layers-and-new-creative-filters/" }, { "Source ID": "collected-9b7a9e08f71a", "Operation IDs": "245", "Citation": "ON1 · Resize adjustment and saved output workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://on1.clickhelp.co/articles/#!photo-raw-2026-user-guide/basic-resize-ai-workflow" }, { "Source ID": "collected-5589b1b2712c", "Operation IDs": "245", "Citation": "on1.com, Resize AI 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/products/resize-ai/" }, { "Source ID": "collected-a88c7287bcc6", "Operation IDs": "245", "Citation": "on1.com, Photo RAW 2027 page, coming fall 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/products/photo-raw/" } ] }, { "ID": "246", "Software": "ON1 Photo RAW 2026", "Operation": "ON1 Super Select AI", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-58c704699f02\ncollected-70d069898088\ncollected-d31cb24c2632\ncollected-55238fce779e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ON1 Super Select predicts semantic membership and uses the mask for a conventional exposure adjustment.", "Previous record: rationale": "Learned region determination survives the applied adjustment, supporting a 10 match. The two stages are not averaged.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). ON1 Photo RAW 2026 (announced 2 September 2025) added Mask Layers and improved AI edge quality for selections; the 2026.5 maintenance update fixes loss of small Super Select AI selections. No generative step is documented for Super Select AI. Not established: ON1 does not disclose the Super Select AI model. The ON1 clickhelp user guide did not render through fetch. on1.com, 2025-09-02 • dpreview.com, 2026.5", "Collected source links": "https://www.on1.com/products/plugins/\nhttps://www.on1.com/videos/how-to-select-edit-anything-in-on1-photo-raw-super-select-ai-and-masking-in-action/\nhttps://www.on1.com/blog/announcing-on1-photo-raw-2026-with-advanced-ai-tools-masking-layers-and-creative-filters/\nhttps://www.dpreview.com/forums/threads/on1-photo-raw-update_-2026-5.4865958/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-58c704699f02", "Operation IDs": "246", "Citation": "ON1 · Machine-learning masking", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/products/plugins/" }, { "Source ID": "collected-70d069898088", "Operation IDs": "246", "Citation": "ON1 · Super Select and local adjustments", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/videos/how-to-select-edit-anything-in-on1-photo-raw-super-select-ai-and-masking-in-action/" }, { "Source ID": "collected-d31cb24c2632", "Operation IDs": "246", "Citation": "on1.com, 2025-09-02", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.on1.com/blog/announcing-on1-photo-raw-2026-with-advanced-ai-tools-masking-layers-and-creative-filters/" }, { "Source ID": "collected-55238fce779e", "Operation IDs": "246", "Citation": "dpreview.com, 2026.5", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.dpreview.com/forums/threads/on1-photo-raw-update_-2026-5.4865958/" } ] }, { "ID": "247", "Software": "Onshape", "Operation": "Parametric Part Studio Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-174c2efd3172\ncollected-be1031ce7d51\ncollected-2958843d62e4\ncollected-dbdb132452fc\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Onshape evaluates native Part Studio features and parametric relationships.", "Previous record: rationale": "The scoped geometric construction contains no trained predictor; its informational AI Advisor is excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: sketches; features; Feature list; rollback bar; suppression; pause regeneration. Returns one result; editable structure: feature list and parts. “The model displayed in the Graphics area is visualized up to the position of the Rollback bar in the Feature list.” Source • Screenshot", "Collected source links": "https://www.onshape.com/en/\nhttps://cad.onshape.com/help/Content/PartStudio/features_and_parts_lists.htm\nhttps://cad.onshape.com/help/Content/Resources/Images/concepts/featurelist-panel-01.png\nhttps://cad.onshape.com/help/Content/PartStudio/feature_basics.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-174c2efd3172", "Operation IDs": "247", "Citation": "Onshape · Parametric Modeling", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.onshape.com/en/" }, { "Source ID": "collected-be1031ce7d51", "Operation IDs": "247", "Citation": "Onshape · Feature editing and rollback", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cad.onshape.com/help/Content/PartStudio/features_and_parts_lists.htm" }, { "Source ID": "collected-2958843d62e4", "Operation IDs": "247", "Citation": "cad.onshape.com — featurelist-panel-01.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cad.onshape.com/help/Content/Resources/Images/concepts/featurelist-panel-01.png" }, { "Source ID": "collected-dbdb132452fc", "Operation IDs": "247", "Citation": "cad.onshape.com — feature_basics.htm", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cad.onshape.com/help/Content/PartStudio/feature_basics.htm" } ] }, { "ID": "248", "Software": "OpenAI API", "Operation": "GPT Image 2.5 Sunburst masked image editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d2d463f2226c\ncollected-1a9fcf8d6701\ncollected-121b461632ab", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "The OpenAI Images API invokes the recorded GPT Image 2.5 Sunburst masked-edit route.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=5; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (GPT Image 2.5 Sunburst edit; mask as guidance). Designer controls: prompt; image; mask (alpha PNG); size; quality (low to max); n; background; output_format; output_compression. Returns several to choose from (n = 1–10 per edit request); a flat image: base64 PNG, JPEG or WebP. “Masking with GPT Image is entirely prompt-based. The model uses the mask as guidance, but may not follow its exact shape with complete precision.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://developers.openai.com/api/docs/models/gpt-image-2\nhttps://developers.openai.com/api/docs/models/gpt-image-2.5-sunburst\nhttps://developers.openai.com/api/docs/guides/image-generation", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d2d463f2226c", "Operation IDs": "248", "Citation": "OpenAI · GPT Image Edit (image + mask)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developers.openai.com/api/docs/models/gpt-image-2" }, { "Source ID": "collected-1a9fcf8d6701", "Operation IDs": "248", "Citation": "OpenAI current GPT Image 2.5 Sunburst", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developers.openai.com/api/docs/models/gpt-image-2.5-sunburst" }, { "Source ID": "collected-121b461632ab", "Operation IDs": "56, 248", "Citation": "developers.openai.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developers.openai.com/api/docs/guides/image-generation" } ] }, { "ID": "249", "Software": "OTOY OctaneRender", "Operation": "AI Upsampling Image Pass", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a4465b8aba21", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Octane's AI upsampler reconstructs a higher-resolution render from a lower-resolution image.", "Previous record: rationale": "Learned enlargement supports the 5–8 reconstruction envelope. The preceding conventional renderer is outside the isolated pass.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). OTOY's Cinema 4D documentation states that the AI up-sampler works like the AI denoiser, with progressive and on-completion modes at 2x2 or 4x4, and publishes no model, training data or objective. OTOY's current OctaneRender product page describes the Spectral AI Denoiser and AI Light and does not describe the up-sampler. Not established: Network architecture, training data and training objective of the up-sampler. docs.otoy.com, not shown", "Collected source links": "https://docs.otoy.com/cinema4d/AIUp-sampler.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a4465b8aba21", "Operation IDs": "249", "Citation": "OTOY · Octane AI Up-Sampler", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.otoy.com/cinema4d/AIUp-sampler.html" } ] }, { "ID": "250", "Software": "OTOY OctaneRender", "Operation": "Spectral AI Denoised Beauty Pass", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0b5f40d2c03f\ncollected-ce9fb6e64737\ncollected-15c74837ede3\ncollected-e5e3083a94bd", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Octane's Spectral AI denoiser predicts the clean beauty image using render buffers.", "Previous record: rationale": "The trained restoration task matches fidelity restoration; auxiliary-buffer count is not an X weight.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). The record source docs.otoy.com/standaloneSE/Denoiser.html returns HTTP 404 on 3 October 2026; the Octane for Cinema 4D manual describes the Spectral AI Denoiser as trained per render kernel (excluding PMC) and producing denoised beauty and light passes. OctaneRender 2027.1 alpha (March 2026) adds a neural viewport mode combining a neural radiance cache with NVIDIA DLSS Ray Reconstruction, a separate interactive route outside this scope. Not established: Training data and loss for the Spectral AI Denoiser are not published. Whether final renders in 2027.1 keep the same denoiser is not stated. docs.otoy.com, 2026-10-03 • docs.otoy.com, accessed 2026-10-03 • digitalproduction.com, 2026-03-18", "Collected source links": "https://docs.otoy.com/standaloneSE/Denoiser.html\nhttps://docs.otoy.com/standaloneSE/DenoisedAOVs.html\nhttps://docs.otoy.com/cinema4d/Denoiser.html\nhttps://digitalproduction.com/2026/03/18/octanerender-2027-1-alpha-goes-neural/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0b5f40d2c03f", "Operation IDs": "250", "Citation": "OTOY · Octane Spectral AI Denoiser", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.otoy.com/standaloneSE/Denoiser.html" }, { "Source ID": "collected-ce9fb6e64737", "Operation IDs": "250", "Citation": "OTOY · Deep learning denoised passes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.otoy.com/standaloneSE/DenoisedAOVs.html" }, { "Source ID": "collected-15c74837ede3", "Operation IDs": "250", "Citation": "docs.otoy.com, accessed 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.otoy.com/cinema4d/Denoiser.html" }, { "Source ID": "collected-e5e3083a94bd", "Operation IDs": "250", "Citation": "digitalproduction.com, 2026-03-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://digitalproduction.com/2026/03/18/octanerender-2027-1-alpha-goes-neural/" } ] }, { "ID": "251", "Software": "Penpot", "Operation": "Manual masking of an imported generated image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5c1849993c1c\ncollected-dab7dbe510ad", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Penpot applies manually specified masks to imported image layers.", "Previous record: rationale": "The scoped mask geometry and clipping are conventional. The imported image's earlier production process does not change this operation's X.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: layer selection; Mask (Ctrl/Cmd+M or layers menu); Unmask (Shift+Ctrl/Cmd+M); mask shape (lowest selected layer). Returns one result; editable structure: layer group with mask. “A mask is a layer that does a clipping and only shows parts of a layer or multiple layers that fall within its shape.” Source", "Collected source links": "https://penpot.app/\nhttps://help.penpot.app/user-guide/designing/layers/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5c1849993c1c", "Operation IDs": "251", "Citation": "Penpot · Penpot overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://penpot.app/" }, { "Source ID": "collected-dab7dbe510ad", "Operation IDs": "251", "Citation": "Penpot paths images and masking", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.penpot.app/user-guide/designing/layers/" } ] }, { "ID": "252", "Software": "Adobe Photoshop", "Operation": "Current Beta AI Assistant Whole Image Variation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-82dd5221e682\ncollected-a1c4ed804eb6\ncollected-7a7f94063e04\ncollected-c6ceb1429990\ncollected-75fba7787a78\ncollected-53ad144e76bd\ncollected-6edf3f300cfa", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop's recorded AI Assistant image-generation route produces a whole-image variation.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=5; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented for the AI Assistant itself. The assistant calls Photoshop generative features such as Generative Fill and Generate Image. Model choice: Partly documented. Not established: Which model generates a whole-image variation when the AI Assistant handles the request: Adobe does not say whether it routes to Generative Fill, Generate Image or Prompt to edit, or which Firefly or partner model it picks. community.adobe.com, 2026-06-19 • helpx.adobe.com • blog.adobe.com, 2025-09-25\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (generator invoked by AI Assistant). Designer controls: prompt (conversation). Returns several to choose from (count not stated); form not documented. “enter simple prompts to make quick edits, generate variations of your image, and refine specific areas without manual selections” Source • Screenshot Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/whats-new/ai-assistant-overview.html\nhttps://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/edit-images-with-ai-assistant.html\nhttps://community.adobe.com/announcements-698/photoshop-desktop-ai-assistant-early-beta-1628556\nhttps://blog.adobe.com/en/publish/2025/09/25/photoshop-beta-expands-generative-fillmore-ai-models-more-possibilities\nhttps://blog.adobe.com/en/publish/2026/06/18/adobe-firefly-introduces-new-agentic-capabilities-and-an-upgraded-creative-ai-studio-built-for-the-way-you-work\nhttps://helpx.adobe.com/firefly/web/firefly-ai-assistant/ai-assistant-faq.html\nhttps://adoberesearch.ctlprojects.com/wp-content/uploads/2026/06/AI-asst-in-Ps-684-x-480-px-1.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-82dd5221e682", "Operation IDs": "252", "Citation": "Adobe: AI Assistant overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/whats-new/ai-assistant-overview.html" }, { "Source ID": "collected-a1c4ed804eb6", "Operation IDs": "252", "Citation": "Adobe: editing workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/edit-images-with-ai-assistant.html" }, { "Source ID": "collected-7a7f94063e04", "Operation IDs": "252", "Citation": "community.adobe.com, 2026-06-19", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.adobe.com/announcements-698/photoshop-desktop-ai-assistant-early-beta-1628556" }, { "Source ID": "collected-c6ceb1429990", "Operation IDs": "252", "Citation": "blog.adobe.com, 2025-09-25", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2025/09/25/photoshop-beta-expands-generative-fillmore-ai-models-more-possibilities" }, { "Source ID": "collected-75fba7787a78", "Operation IDs": "252", "Citation": "Adobe, Jun 18, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2026/06/18/adobe-firefly-introduces-new-agentic-capabilities-and-an-upgraded-creative-ai-studio-built-for-the-way-you-work" }, { "Source ID": "collected-53ad144e76bd", "Operation IDs": "90, 252", "Citation": "Firefly AI Assistant FAQ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/firefly/web/firefly-ai-assistant/ai-assistant-faq.html" }, { "Source ID": "collected-6edf3f300cfa", "Operation IDs": "252", "Citation": "adoberesearch.ctlprojects.com — AI-asst-in-Ps-684-x-480-px-1.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://adoberesearch.ctlprojects.com/wp-content/uploads/2026/06/AI-asst-in-Ps-684-x-480-px-1.png" } ] }, { "ID": "253", "Software": "Adobe Photoshop", "Operation": "Generative Fill with Firefly", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–3.5", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7d33071efc3d\ncollected-897a3b739bb0\ncollected-c4ab777d01d4\ncollected-16b23fbd6f4d\ncollected-09d36179860b\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop Generative Fill with Firefly Fill and Expand synthesizes replacement image content within the selected region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 3 model-decided, 2 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–3.5; no midpoint is used.", "Previous record: calculation": "K=0; M=3; unknown=2; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 2 ÷ (5 − 1) = 3.5. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Firefly Fill & Expand or a chosen partner model). Designer controls: selection; prompt (optional); model choice (Adobe and partner models); reference image; Generate Similar. Returns several to choose from (3); editable structure: Generative Layer, new per Generate. “Each time you select Generate, a new Generative Layer is created.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/create-open-import-images/create-images/edit-images-with-generative-fill.html\nhttps://research.adobe.com/news/photoshop-turns-35/\nhttps://web.archive.org/web/20260905002207id_/https://helpx.adobe.com/photoshop/desktop/generative-ai/frequently-asked-questions-about-generative-ai-features.html\nhttps://helpx.adobe.com/photoshop/desktop/create-open-import-images/create-images/generate-sharper-variations-with-enhance-detail.html\nhttps://web.archive.org/web/20260514081947id_/https://helpx.adobe.com/photoshop/desktop/create-manage-layers/color-adjustment-fill-layers/adjustment-and-fill-layers-overview.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7d33071efc3d", "Operation IDs": "253", "Citation": "Adobe Fill guide, updated September 15, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/create-open-import-images/create-images/edit-images-with-generative-fill.html" }, { "Source ID": "collected-897a3b739bb0", "Operation IDs": "253, 397", "Citation": "Adobe Research, September 15, 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.adobe.com/news/photoshop-turns-35/" }, { "Source ID": "collected-c4ab777d01d4", "Operation IDs": "253", "Citation": "web.archive.org — frequently-asked-questions-about-generative-ai-features.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260905002207id_/https://helpx.adobe.com/photoshop/desktop/generative-ai/frequently-asked-questions-about-generative-ai-features.html" }, { "Source ID": "collected-16b23fbd6f4d", "Operation IDs": "253", "Citation": "helpx.adobe.com — generate-sharper-variations-with-enhance-detail.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/create-open-import-images/create-images/generate-sharper-variations-with-enhance-detail.html" }, { "Source ID": "collected-09d36179860b", "Operation IDs": "253, 259", "Citation": "web.archive.org — adjustment-and-fill-layers-overview.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260514081947id_/https://helpx.adobe.com/photoshop/desktop/create-manage-layers/color-adjustment-fill-layers/adjustment-and-fill-layers-overview.html" } ] }, { "ID": "254", "Software": "Adobe Photoshop", "Operation": "Harmonize Subject Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e5a90e558679\ncollected-1204d15d0e22\ncollected-a485dd75b212", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop Harmonize uses a trained unified process for foreground relighting, boundary blending and generated shadows or reflections.", "Previous record: rationale": "The product-linked research establishes more than a fixed color transform. Appearance reconstruction and generated scene effects support a 2–5 synthesis-to-perceptual comparison; the evidence does not select an exact t.", "Previous record: calculation": "X = 2 + t × (5 − 2); t = 0–1; X = 2–5. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–5.0). Adobe Research's Harmonize article (31 October 2025) and the linked CVPR 2024 paper (v2, 7 April 2024) describe a lighting-aware diffusion model for foreground relighting, shadow and reflection generation and boundary blending; both are among the record's sources. No newer documentation states how much of the foreground is regenerated versus kept. Not established: Share of foreground pixels regenerated versus preserved, which would select between REF02 and REF05. research.adobe.com, 2025-10-31 • arxiv.org, v2, 2024-04-07", "Collected source links": "https://research.adobe.com/news/the-research-behind-harmonize/\nhttps://helpx.adobe.com/photoshop/desktop/repair-retouch/remove-objects-fill-space/blend-subjects-with-harmonize.html\nhttps://arxiv.org/abs/2312.06886", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e5a90e558679", "Operation IDs": "254", "Citation": "Adobe Harmonize development account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.adobe.com/news/the-research-behind-harmonize/" }, { "Source ID": "collected-1204d15d0e22", "Operation IDs": "254", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/repair-retouch/remove-objects-fill-space/blend-subjects-with-harmonize.html" }, { "Source ID": "collected-a485dd75b212", "Operation IDs": "254", "Citation": "arxiv.org, v2, 2024-04-07", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://arxiv.org/abs/2312.06886" } ] }, { "ID": "255", "Software": "Adobe Photoshop", "Operation": "Manual Layer and Mask Compositing Comparison", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-57a36df33987\ncollected-8c9daaef9e18\ncollected-c7e2b30dd3cf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop applies manually arranged layers, masks and blend settings in a conventional composite.", "Previous record: rationale": "No trained determination occurs within the scoped composition. An imported generated layer retains its provenance, but that earlier generation is outside this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: layer order; Opacity; Fill opacity; blending mode; layer mask (density, feather); visibility. Returns one result; editable structure: layers, masks, blend settings. “You use layer blending modes to determine how a layer blends with the pixels in layers beneath it.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/using/editing-layer-masks.html\nhttps://developer.adobe.com/photoshop/uxp/2022/ps-reference/media/imaging\nhttps://helpx.adobe.com/photoshop/desktop/create-manage-layers/get-started-layers/layers-overview.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-57a36df33987", "Operation IDs": "255, 389", "Citation": "Adobe Edit layer masks, July 18, 2024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/using/editing-layer-masks.html" }, { "Source ID": "collected-8c9daaef9e18", "Operation IDs": "255", "Citation": "Adobe UXP: Imaging API", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.adobe.com/photoshop/uxp/2022/ps-reference/media/imaging" }, { "Source ID": "collected-c7e2b30dd3cf", "Operation IDs": "255", "Citation": "helpx.adobe.com — layers-overview.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/create-manage-layers/get-started-layers/layers-overview.html" } ] }, { "ID": "256", "Software": "Adobe Photoshop", "Operation": "Neural Filters Colorize with Smart Filter Output", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-aaecb9349472\ncollected-57dd0579a2aa\ncollected-9eedd3f14733", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop Colorize predicts image color and returns it through the selected Smart Filter output route.", "Previous record: rationale": "Learned color estimation changes appearance beyond a fixed prescribed transform. The fixed-function-to-perceptual comparison supports 5–12 without a precise t; colorization is not equated with unrestricted scene synthesis.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Colorize remains a cloud-processed Neural Filter in Photoshop 2026 (v27.0, October 2025); current descriptions say Adobe-trained models predict plausible colors. No model replacement, version change or training objective is documented beyond the record's sources. Not established: Training objective (color classification, regression or adversarial) and model version of the shipped filter. Adobe helpx and adobe.com pages refused fetch; the quoted description is third-party. msyeditor.com, 2026", "Collected source links": "https://helpx.adobe.com/photoshop/using/colorize.html\nhttps://helpx.adobe.com/photoshop/desktop/effects-filters/neural-filters/use-neural-filters-to-enhance-images.html\nhttps://msyeditor.com/photoshop-neural-filters-2026-complete-tutorial-for-video-stills-and-photos/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-aaecb9349472", "Operation IDs": "256", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/using/colorize.html" }, { "Source ID": "collected-57dd0579a2aa", "Operation IDs": "256", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/effects-filters/neural-filters/use-neural-filters-to-enhance-images.html" }, { "Source ID": "collected-9eedd3f14733", "Operation IDs": "256", "Citation": "msyeditor.com, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://msyeditor.com/photoshop-neural-filters-2026-complete-tutorial-for-video-stills-and-photos/" } ] }, { "ID": "257", "Software": "Adobe Photoshop", "Operation": "Remove Tool Brushed Region with Generative AI On", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-19e8a423d7a6\ncollected-1630edf28cfe\ncollected-0a36b698f136\ncollected-b6afb5616a78\ncollected-f59b39b690af\ncollected-ef165773d5b7\ncollected-f7a8d3a75453\ncollected-aada92b2030e\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop Remove Tool with Generative AI On synthesizes replacement content for the brushed region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=4; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Adobe Firefly \"Remove Tool 3 model\" (Adobe), generally available per Adobe's January 2026 Photoshop announcement. Earlier: the Remove tool used \"the latest Adobe Firefly Image Model\" (July 2025) and Adobe Firefly when Generative AI is on (October 2024, Photoshop 2025). Model choice: No model picker documented for the Remove tool. Not established: Adobe gives no architecture or version history for the Remove Tool model line (what Remove Tool 1 and 2 were). blog.adobe.com, 2026-01-27 • blog.adobe.com, 2025-07-29 • photoshopessentials.com\nChecked 3 Oct 2026 Remove tool Mode has three settings: Auto (May use generative AI), Generative AI is on (Cloud or Device), Generative AI is off. This record covers Generative AI is on. Find distractions detects General distractions, People, Wires and cables. Create new layer applies removals to a new layer. Adobe, Remove objects, Aug 18, 2026\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Firefly, Mode set to Generative AI on). Designer controls: brush size; Mode (Auto, Generative AI on, Generative AI off); Remove after each stroke. Returns one result; form not documented. “Use the Mode option in the Options bar to control how generative AI is applied.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/repair-retouch/remove-objects-fill-space/remove-unwanted-objects-and-distractions.html\nhttps://helpx.adobe.com/photoshop/using/tool-techniques/remove-tool.html\nhttps://blog.adobe.com/en/publish/2026/01/27/new-photoshop-innovations-provide-creative-pros-more-control-realism-precision\nhttps://blog.adobe.com/en/publish/2025/07/29/powerful-new-photoshop-innovations-creators-creative-pros\nhttps://www.photoshopessentials.com/photo-editing/using-generative-ai-with-the-remove-tool-in-photoshop/\nhttps://web.archive.org/web/20260926081858id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/manage-image-states.html\nhttps://web.archive.org/web/20260913075516id_/https://helpx.adobe.com/photoshop/using/tool-techniques/remove-tool.html\nhttps://web.archive.org/web/20260925075845id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/history-panel-overview.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-19e8a423d7a6", "Operation IDs": "257", "Citation": "Adobe · Remove Tool / Distraction Removal", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/repair-retouch/remove-objects-fill-space/remove-unwanted-objects-and-distractions.html" }, { "Source ID": "collected-1630edf28cfe", "Operation IDs": "257", "Citation": "Adobe · Remove Tool modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/using/tool-techniques/remove-tool.html" }, { "Source ID": "collected-0a36b698f136", "Operation IDs": "257", "Citation": "blog.adobe.com, 2026-01-27", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2026/01/27/new-photoshop-innovations-provide-creative-pros-more-control-realism-precision" }, { "Source ID": "collected-b6afb5616a78", "Operation IDs": "257", "Citation": "blog.adobe.com, 2025-07-29", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2025/07/29/powerful-new-photoshop-innovations-creators-creative-pros" }, { "Source ID": "collected-f59b39b690af", "Operation IDs": "257", "Citation": "photoshopessentials.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.photoshopessentials.com/photo-editing/using-generative-ai-with-the-remove-tool-in-photoshop/" }, { "Source ID": "collected-ef165773d5b7", "Operation IDs": "257, 388, 389", "Citation": "web.archive.org — manage-image-states.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260926081858id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/manage-image-states.html" }, { "Source ID": "collected-f7a8d3a75453", "Operation IDs": "257", "Citation": "web.archive.org — remove-tool.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260913075516id_/https://helpx.adobe.com/photoshop/using/tool-techniques/remove-tool.html" }, { "Source ID": "collected-aada92b2030e", "Operation IDs": "257, 388, 389", "Citation": "web.archive.org — history-panel-overview.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260925075845id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/history-panel-overview.html" } ] }, { "ID": "259", "Software": "Adobe Photoshop", "Operation": "Sky Replacement with Editable Layer Output", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0bc94a3c1b13\ncollected-443449c68bd8\ncollected-e05d26cb6f9b\ncollected-f9592c3f87ef\ncollected-bb5438c6ff12\ncollected-09d36179860b\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop Sky Replacement predicts the sky boundary, then combines a supplied sky with the image.", "Previous record: rationale": "The documented learned region decision survives subsequent compositing, supporting the region reference at 10. This does not assign a generation score to the separately supplied sky asset.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). The Adobe Sky Replacement product page now attributes sky detection and replacement to Adobe Firefly generative AI, while the Photoshop help page still describes choosing a preset sky or adding a user image. No documentation states that the feature synthesizes sky or foreground pixels; the help page describes the lighting controls as a blending mode and a foreground contrast adjustment. Not established: Whether the Firefly label reflects a changed model or a relabelling of the earlier Sensei segmentation is not documented. The mechanism of the automatic lighting and shadow match is undisclosed. adobe.com, page undated, fetched 2026-10-03 • helpx.adobe.com, help page, 2026", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/effects-filters/artistic-stylize-filters/replace-the-sky-in-images.html\nhttps://www.adobe.com/learn/photoshop/web/sky-replacement\nhttps://www.adobe.com/products/photoshop/sky-replacement.html\nhttps://helpx.adobe.com/photoshop/desktop/save-and-export/save-files/save-your-work.html\nhttps://web.archive.org/web/20260928191723id_/https://helpx.adobe.com/photoshop/desktop/effects-filters/artistic-stylize-filters/replace-the-sky-in-images.html\nhttps://web.archive.org/web/20260514081947id_/https://helpx.adobe.com/photoshop/desktop/create-manage-layers/color-adjustment-fill-layers/adjustment-and-fill-layers-overview.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0bc94a3c1b13", "Operation IDs": "259", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/effects-filters/artistic-stylize-filters/replace-the-sky-in-images.html" }, { "Source ID": "collected-443449c68bd8", "Operation IDs": "259", "Citation": "www.adobe.com — sky-replacement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.adobe.com/learn/photoshop/web/sky-replacement" }, { "Source ID": "collected-e05d26cb6f9b", "Operation IDs": "259", "Citation": "adobe.com, page undated, fetched 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.adobe.com/products/photoshop/sky-replacement.html" }, { "Source ID": "collected-f9592c3f87ef", "Operation IDs": "259", "Citation": "helpx.adobe.com — save-your-work.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/save-and-export/save-files/save-your-work.html" }, { "Source ID": "collected-bb5438c6ff12", "Operation IDs": "259", "Citation": "web.archive.org — replace-the-sky-in-images.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260928191723id_/https://helpx.adobe.com/photoshop/desktop/effects-filters/artistic-stylize-filters/replace-the-sky-in-images.html" }, { "Source ID": "collected-09d36179860b", "Operation IDs": "253, 259", "Citation": "web.archive.org — adjustment-and-fill-layers-overview.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260514081947id_/https://helpx.adobe.com/photoshop/desktop/create-manage-layers/color-adjustment-fill-layers/adjustment-and-fill-layers-overview.html" } ] }, { "ID": "260", "Software": "Pika", "Operation": "Pika 2.5 prompt to video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1ab5308ce3f3\ncollected-49800485ce10", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Pika 2.5 generates a video through its recorded prompt-to-video route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Pika 2.5 video model). Designer controls: prompt; resolution (720p or 1080p). Returns count not documented; a flat video: 5 s, silent. “A prompt on its own is the whole brief: subject, camera move, lens language, grade.” Source", "Collected source links": "https://pika.art\nhttps://create.pika.art/apps/pika-2-5", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1ab5308ce3f3", "Operation IDs": "260", "Citation": "Pika · Pika overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://pika.art" }, { "Source ID": "collected-49800485ce10", "Operation IDs": "260", "Citation": "Pika current 2.5 generation route", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://create.pika.art/apps/pika-2-5" } ] }, { "ID": "261", "Software": "Plasticity", "Operation": "Direct Solid and Surface Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-cbfe176ced3f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Plasticity performs direct solid and surface editing through its geometric kernel.", "Previous record: rationale": "The specified construction uses conventional geometry operations with no documented trained prediction.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: push/pull; fillets; booleans; extrude; dimensions; construction planes; bridge curves. Returns one result; editable structure: solids and surfaces. “Push, pull, and adjust existing geometry without starting from scratch.” Source", "Collected source links": "https://www.plasticity.xyz/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-cbfe176ced3f", "Operation IDs": "261", "Citation": "Plasticity · Parasolid direct modeling", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.plasticity.xyz/" } ] }, { "ID": "262", "Software": "Playground", "Operation": "GPT Image 2.5 Flare design with Smart Layers", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4c72b55c99ff\ncollected-a4806d79e018\ncollected-07e7e1d2e061", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Playground combines its recorded GPT Image generation route with Smart Layers separation.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (GPT Image 2.5 Flare generation, then AI layer separation). Designer controls: prompt; model choice (GPT Image 2.5 Flare or Auto Mode); Smart Layers select, move, scale, rotate; edit text in place; delete and inpaint; batch apply. Returns count not documented; editable structure: layers: separate selectable objects and editable text after Smart Layers. “Smart Layers uses AI to break any design into individually selectable objects and text.” Source • Screenshot", "Collected source links": "https://playgroundai.com/blog/introducing-playground-v3\nhttps://playgroundai.com/design/whats-new\nhttps://cdn.playgroundai.com/whats-new/auto-mode.png?c=1", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4c72b55c99ff", "Operation IDs": "262", "Citation": "Playground · Design Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://playgroundai.com/blog/introducing-playground-v3" }, { "Source ID": "collected-a4806d79e018", "Operation IDs": "262", "Citation": "Playground current models and Smart Layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://playgroundai.com/design/whats-new" }, { "Source ID": "collected-07e7e1d2e061", "Operation IDs": "262", "Citation": "cdn.playgroundai.com — auto-mode.png?c=1", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.playgroundai.com/whats-new/auto-mode.png?c=1" } ] }, { "ID": "263", "Software": "Point-E", "Operation": "Text to Point Cloud Research Example", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-34c5ba6d04cf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Point-E's research example generates a point cloud through its learned diffusion pipeline.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (text-conditioned point-cloud diffusion). Designer controls: prompt; model choice (base40M-textvec plus upsampler); guidance_scale; num_points; batch_size. Returns one result (example notebook batch_size 1); editable structure: RGB point cloud; mesh via separate SDF model. “use our small, worse quality pure text-to-3D model to produce 3D point clouds directly from text descriptions” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://github.com/openai/point-e", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-34c5ba6d04cf", "Operation IDs": "263", "Citation": "OpenAI · Point-E point cloud diffusion", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/openai/point-e" } ] }, { "ID": "264", "Software": "Polycam", "Operation": "Object Mode Gaussian capture with crop preview", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1916959d8e89\ncollected-d38038a86f6d\ncollected-a548a41725d6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Polycam Gaussian Object Mode constructs a fitted scene representation from captures.", "Previous record: rationale": "The full route does not separate pose estimation, learned preprocessing and per-scene representation fitting. A coordinate requires that separation and the same fitting criterion used for the other radiance-field cases.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Polycam's Gaussian splatting page describes per-scene training: renders are compared with the user's images from the same pose and the error is back-propagated over thousands of iterations; no pretrained model or learned prior is mentioned. Pose estimation and any preprocessing in Polycam's cloud route are not described. Not established: Camera pose method and any learned preprocessing in Polycam's cloud pipeline. The remaining gap is a calibration decision: whether per-scene fitting of a scene representation counts as learned determination on this scale, applied consistently across neural fields and explicit Gaussians. Vendor or project documentation cannot settle that criterion. poly.cam, undated • learn.poly.cam, undated", "Collected source links": "https://poly.cam/object-capture\nhttps://learn.poly.cam/hc/en-us/articles/48565771018772-Which-Capture-Mode-Should-I-Use\nhttps://poly.cam/tools/gaussian-splatting", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1916959d8e89", "Operation IDs": "264, 265, 266", "Citation": "Polycam · Gaussian Splat Mode", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/object-capture" }, { "Source ID": "collected-d38038a86f6d", "Operation IDs": "264, 265, 266", "Citation": "Polycam · Separate capture modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.poly.cam/hc/en-us/articles/48565771018772-Which-Capture-Mode-Should-I-Use" }, { "Source ID": "collected-a548a41725d6", "Operation IDs": "264", "Citation": "poly.cam, undated", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/tools/gaussian-splatting" } ] }, { "ID": "265", "Software": "Polycam", "Operation": "LiDAR Depth to Mesh Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1916959d8e89\ncollected-d38038a86f6d\ncollected-ea4eb83f7b33\ncollected-51f3b62c11f0\ncollected-a1fb9579ec83", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Polycam LiDAR mode constructs a mesh from captured depth and imagery.", "Previous record: rationale": "A trained model (ARKit depth fusion) determines the dense depth values from which the LiDAR mesh is reconstructed, so the learned branch is documented. The rating input stays empty because the learned share within fused depth and Polycam's multi-frame reconstruction method are undocumented, so no reference pair can be selected.", "Previous record: calculation": "", "Previous record: notes": "Changed 3 Oct 2026 X Unresolved → Unresolved. Apple's WWDC20 session 'Explore ARKit 4' states that LiDAR depth and the RGB image are fused by machine learning algorithms into the dense depth map ARKit exposes, and that ARKit scene geometry aggregates that depth across frames into a mesh. Polycam's polyform README states its LiDAR-mode depth maps are generated by ARKit and that Polycam globally optimizes ARKit poses before reconstructing the mesh; Polycam's 19 September 2026 blog places LiDAR capture under Space Mode with on-device or cloud processing. Not established: How much of ARKit's fused depth the trained model determines relative to the LiDAR measurement; whether Polycam's on-device and cloud reconstructions use ARKit depth maps, the ARKit scene mesh, or added image-based processing. Previous reason: Depth sensing does not by itself establish a conventional-only pipeline. The retained contribution of trained image or depth processing is not established for the complete scoped service; no branch or coordinate is asserted. developer.apple.com, WWDC 2020 • github.com, undated README • poly.cam, 2026-09-19", "Collected source links": "https://poly.cam/object-capture\nhttps://learn.poly.cam/hc/en-us/articles/48565771018772-Which-Capture-Mode-Should-I-Use\nhttps://developer.apple.com/videos/play/wwdc2020/10611/\nhttps://github.com/PolyCam/polyform/blob/main/README.md\nhttps://poly.cam/blog/lidar-vs-photogrammetry-which-3d-scanning-method-is-right-for-you", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1916959d8e89", "Operation IDs": "264, 265, 266", "Citation": "Polycam · Gaussian Splat Mode", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/object-capture" }, { "Source ID": "collected-d38038a86f6d", "Operation IDs": "264, 265, 266", "Citation": "Polycam · Separate capture modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.poly.cam/hc/en-us/articles/48565771018772-Which-Capture-Mode-Should-I-Use" }, { "Source ID": "collected-ea4eb83f7b33", "Operation IDs": "265", "Citation": "developer.apple.com, WWDC 2020", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.apple.com/videos/play/wwdc2020/10611/" }, { "Source ID": "collected-51f3b62c11f0", "Operation IDs": "265", "Citation": "github.com, undated README", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/PolyCam/polyform/blob/main/README.md" }, { "Source ID": "collected-a1fb9579ec83", "Operation IDs": "265, 266", "Citation": "poly.cam, 2026-09-19", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/blog/lidar-vs-photogrammetry-which-3d-scanning-method-is-right-for-you" } ] }, { "ID": "266", "Software": "Polycam", "Operation": "Object Mode Photogrammetric Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1916959d8e89\ncollected-d38038a86f6d\ncollected-c49f8cf36adc\ncollected-a1fb9579ec83", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Polycam Object photogrammetry mode reconstructs a mesh from photographs.", "Previous record: rationale": "Classical photogrammetry is a plausible component, but the saved service documentation does not isolate trained feature or depth estimation. A conventional-only result cannot be assumed for the entire managed pipeline.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). A third-party review (November 2021, updated September 2022) states Polycam Photo Mode was powered by Apple's Object Capture with cloud processing; Polycam's own pages (blog 19 September 2026) describe Object Mode photogrammetry as analysis of overlapping photos processed in the cloud, without naming the engine or any trained component. Apple's Object Capture documentation (WWDC21 to WWDC24) does not identify trained components either. Not established: Whether current Object Mode photogrammetry still runs Apple Object Capture or another engine, and whether learned feature matching, depth or automatic masking determine part of the mesh. 3dwithus.com, 2021-11-25, updated 2022-09-26 • poly.cam, 2026-09-19 • poly.cam, undated", "Collected source links": "https://poly.cam/object-capture\nhttps://learn.poly.cam/hc/en-us/articles/48565771018772-Which-Capture-Mode-Should-I-Use\nhttps://3dwithus.com/polycam-3d-scanning-app-review-lidar-vs-photo-mode\nhttps://poly.cam/blog/lidar-vs-photogrammetry-which-3d-scanning-method-is-right-for-you", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1916959d8e89", "Operation IDs": "264, 265, 266", "Citation": "Polycam · Gaussian Splat Mode", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/object-capture" }, { "Source ID": "collected-d38038a86f6d", "Operation IDs": "264, 265, 266", "Citation": "Polycam · Separate capture modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.poly.cam/hc/en-us/articles/48565771018772-Which-Capture-Mode-Should-I-Use" }, { "Source ID": "collected-c49f8cf36adc", "Operation IDs": "266", "Citation": "3dwithus.com, 2021-11-25, updated 2022-09-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://3dwithus.com/polycam-3d-scanning-app-review-lidar-vs-photo-mode" }, { "Source ID": "collected-a1fb9579ec83", "Operation IDs": "265, 266", "Citation": "poly.cam, 2026-09-19", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/blog/lidar-vs-photogrammetry-which-3d-scanning-method-is-right-for-you" } ] }, { "ID": "267", "Software": "Polycam", "Operation": "Web Single Image to Model", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c485fe11f6bb\ncollected-69b93a66688b\ncollected-0f913ce9af80", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Polycam's web 3D Generator predicts a mesh from a single image.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=3; conditional=0; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Model choice: No model choice is documented. Not established: Polycam does not name the image-to-3D model on its product page, tool page or help article located. poly.cam, accessed 2026-10-04 • learn.poly.cam\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (single image to 3D). Designer controls: image upload (one JPEG or PNG). Returns count not documented; editable structure: 3D model exportable for further editing; file format not stated on fetched text. “You can only use one image to generate a 3D model.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://poly.cam/tools/3d-model-generator\nhttps://learn.poly.cam/hc/en-us/articles/35116894902292-How-to-Create-a-Model-from-the-3D-Generator\nhttps://poly.cam/ai-capture", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c485fe11f6bb", "Operation IDs": "267", "Citation": "Polycam · 3D Generator", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/tools/3d-model-generator" }, { "Source ID": "collected-69b93a66688b", "Operation IDs": "267", "Citation": "Polycam · Image to editable model", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://learn.poly.cam/hc/en-us/articles/35116894902292-How-to-Create-a-Model-from-the-3D-Generator" }, { "Source ID": "collected-0f913ce9af80", "Operation IDs": "267", "Citation": "poly.cam, accessed 2026-10-04", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://poly.cam/ai-capture" } ] }, { "ID": "268", "Software": "Postshot", "Operation": "Explicit Splat Optimization and Cleanup", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8490751c6e09\ncollected-d6ee8856b643\ncollected-bfe8766e3992\ncollected-ebc3a4d778ca\ncollected-8bdba9e2d0e5\ncollected-7661341e0719\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Postshot's Splat mode fits an explicit Gaussian scene with supplied camera poses.", "Previous record: rationale": "This excludes camera tracking and NeRF mode. The per-scene fitting criterion remains unresolved across explicit splats and neural fields; representing the fit without neural layers does not alone justify 15.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Postshot v1.1 (13 May 2026) adds Photometric Compensation (auto exposure, white point, vignetting) and fixes MCMC-profile artefacts and camera-tracking failures. Current splat profiles are Splat3 (recommended), Splat MCMC and legacy Splat ADC; no pretrained network or learned prior is documented for them. Not established: Whether Photometric Compensation fits per-scene parameters or applies a trained model. The remaining gap is a calibration decision: whether per-scene fitting of a scene representation counts as learned determination on this scale, applied consistently across neural fields and explicit Gaussians. Vendor or project documentation cannot settle that criterion. jawset.com, v1.1, 2026-05-13 • jawset.com, current docs (Splat3 profile)", "Collected source links": "https://radiancefields.com/postshot-launches-v1-0-out-of-beta\nhttps://www.jawset.com/docs/d/Postshot%2BUser%2BGuide/Release%2BNotes/v0.3\nhttps://www.jawset.com/docs/d/Postshot+User+Guide/Release+Notes/v1.1\nhttps://www.jawset.com/docs/d/Postshot+User+Guide/Interface/Training+Configuration\nhttps://www.jawset.com/docs/d/Postshot+User+Guide/Interface/Scene+Tree/Rdnc+Field\nhttps://www.jawset.com/docs/d/Postshot%20User%20Guide/Interface/Viewport%20Tools", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8490751c6e09", "Operation IDs": "268", "Citation": "Radiance Fields · Splat Training", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://radiancefields.com/postshot-launches-v1-0-out-of-beta" }, { "Source ID": "collected-d6ee8856b643", "Operation IDs": "268", "Citation": "Jawset · Splat selection deletion and export", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.jawset.com/docs/d/Postshot%2BUser%2BGuide/Release%2BNotes/v0.3" }, { "Source ID": "collected-bfe8766e3992", "Operation IDs": "268", "Citation": "jawset.com, v1.1, 2026-05-13", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.jawset.com/docs/d/Postshot+User+Guide/Release+Notes/v1.1" }, { "Source ID": "collected-ebc3a4d778ca", "Operation IDs": "268", "Citation": "jawset.com, current docs (Splat3 profile)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.jawset.com/docs/d/Postshot+User+Guide/Interface/Training+Configuration" }, { "Source ID": "collected-8bdba9e2d0e5", "Operation IDs": "268", "Citation": "www.jawset.com — Rdnc+Field", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.jawset.com/docs/d/Postshot+User+Guide/Interface/Scene+Tree/Rdnc+Field" }, { "Source ID": "collected-7661341e0719", "Operation IDs": "268", "Citation": "www.jawset.com — Viewport Tools", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.jawset.com/docs/d/Postshot%20User%20Guide/Interface/Viewport%20Tools" } ] }, { "ID": "269", "Software": "Adobe Premiere", "Operation": "Auto Reframe", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-771156988d6f\ncollected-645bd1d9756f\ncollected-543cd7cda932\ncollected-01a81b25fbe6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere Auto Reframe uses learned subject and action analysis to determine a moving crop.", "Previous record: rationale": "The retained responsibility is identifying and following the relevant region, supporting the region-determination reference at 10. Conventional crop execution adds no second score.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Premiere 26.2.2 (May 2026) added a target-resolution field to the Auto Reframe dialog. Help pages continue to describe automatic subject-following reframing across aspect ratios; no generative step or model change is documented. Not established: The subject-tracking model remains undisclosed. Adobe helpx pages refused direct fetch; search-result text was used. helpx.adobe.com, May 2026 (v26.2.2) • helpx.adobe.com, help page, 2026", "Collected source links": "https://blog.adobe.com/en/publish/2018/11/14/projectsmoothoperator-brings-intelligent-automated-cropping-to-video-publishing\nhttps://blog.adobe.com/en/publish/2019/11/04/streamline-video-editing-and-deliver-better-stories-faster\nhttps://helpx.adobe.com/premiere/desktop/whats-new/release-notes.html\nhttps://helpx.adobe.com/premiere/desktop/add-video-effects/commonly-used-effects/auto-reframe-overview.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-771156988d6f", "Operation IDs": "269", "Citation": "Adobe · Subject saliency and crop-path computation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2018/11/14/projectsmoothoperator-brings-intelligent-automated-cropping-to-video-publishing" }, { "Source ID": "collected-645bd1d9756f", "Operation IDs": "269", "Citation": "Adobe · Released Auto Reframe workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2019/11/04/streamline-video-editing-and-deliver-better-stories-faster" }, { "Source ID": "collected-543cd7cda932", "Operation IDs": "269", "Citation": "helpx.adobe.com, May 2026 (v26.2.2)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/whats-new/release-notes.html" }, { "Source ID": "collected-01a81b25fbe6", "Operation IDs": "269", "Citation": "helpx.adobe.com, help page, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/add-video-effects/commonly-used-effects/auto-reframe-overview.html" } ] }, { "ID": "270", "Software": "Adobe Premiere", "Operation": "Generative Extend", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.2–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ccaec2225bb6\ncollected-ef1b5cf80c96\ncollected-7a0ac0c15eef\ncollected-faac2470b86a\ncollected-94248dad41d0\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere Generative Extend uses Firefly Video to synthesize additional frames.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 1 model-decided, 3 unknown and 0 setting-dependent categories among 6 applicable. The recorded uncertainty gives 3.2–5.0; no midpoint is used.", "Previous record: calculation": "K=2; M=1; unknown=3; conditional=0; A=6. Lower: 2 + 3 × 2 ÷ (6 − 1) = 3.2. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Adobe (Sep 9, 2026): Generative Extend uses a cloud Firefly model; source clips 360p to 4K UHD, any aspect ratio; extensions above 30 fps are generated at 30 fps, in 8-bit and SDR; dialog is not created or extended and clips with music are not eligible. Adobe, Generative Extend overview, Sep 9, 2026\nDesigner side, 3 Oct 2026 Other. Decided by a trained model (Firefly Video). Designer controls: Generative Extend tool; drag clip edge to target length. Returns one result; a flat video: generated extension; original media unchanged. “select and drag the edge of a video or audio clip to the desired length” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://research.adobe.com/news/adobe-research-and-firefly-team-pioneer-new-video-generation-tech/\nhttps://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/generative-extend-overview.html\nhttps://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/add-frames-using-generative-extend.html\nhttps://web.archive.org/web/20260822065133id_/https://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/generative-extend-faq.html\nhttps://web.archive.org/web/20260418145938id_/https://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/add-frames-using-generative-extend.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ccaec2225bb6", "Operation IDs": "270", "Citation": "research.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.adobe.com/news/adobe-research-and-firefly-team-pioneer-new-video-generation-tech/" }, { "Source ID": "collected-ef1b5cf80c96", "Operation IDs": "270", "Citation": "helpx.adobe.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/generative-extend-overview.html" }, { "Source ID": "collected-7a0ac0c15eef", "Operation IDs": "270", "Citation": "helpx.adobe.com — add-frames-using-generative-extend.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/add-frames-using-generative-extend.html" }, { "Source ID": "collected-faac2470b86a", "Operation IDs": "270", "Citation": "web.archive.org — generative-extend-faq.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260822065133id_/https://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/generative-extend-faq.html" }, { "Source ID": "collected-94248dad41d0", "Operation IDs": "270", "Citation": "web.archive.org — add-frames-using-generative-extend.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260418145938id_/https://helpx.adobe.com/premiere/desktop/edit-projects/edit-with-generative-ai/add-frames-using-generative-extend.html" } ] }, { "ID": "271", "Software": "Adobe Premiere", "Operation": "Text-Based Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-dbf5000ddc7b\ncollected-fae9ff261453\ncollected-fdc2aa2ec65e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere Text Based Editing includes learned transcription and time alignment, with those alignments retained in edit boundaries.", "Previous record: rationale": "The fidelity-to-region envelope 8–10 records constrained recognition and localization. It differs from a separately scoped edit of an already aligned transcript, which would exclude the learned stage.", "Previous record: calculation": "X = 8 + t × (10 − 8); t = 0–1; X = 8–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0–10.0). Adobe's Text-Based Editing overview (updated 7 January 2026) states that transcripts align dialogue with timecodes and that editing the text trims and places matching clips in the timeline, with filler-word detection; Speechmatics reported on 23 June 2026 that it remains Premiere's speech-to-text engine, now run on-device. The documented learned output is a time-aligned transcript, the edited sequence is built from trimmed source clips, and no generative stage is documented in this route. Not established: The speech model's architecture and training are not published. Documentation does not decide whether transcript-driven cut boundaries compare nearer region determination (10) or fidelity reconstruction (8). helpx.adobe.com, Last updated 2026-01-07 • speechmatics.com, 2026-06-23", "Collected source links": "https://www.speechmatics.com/company/articles-and-news/speechmatics-versus-whisper-how-adobe-premieres-on-device-speech-engine-got-rebuilt\nhttps://helpx.adobe.com/premiere/desktop/edit-projects/edit-video-using-text-based-editing/edit-sequences-using-text-based-editing.html\nhttps://helpx.adobe.com/premiere/desktop/edit-projects/edit-video-using-text-based-editing/overview-of-text-based-editing.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-dbf5000ddc7b", "Operation IDs": "271", "Citation": "Speechmatics · Neural transcription engine supplied to Premiere", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.speechmatics.com/company/articles-and-news/speechmatics-versus-whisper-how-adobe-premieres-on-device-speech-engine-got-rebuilt" }, { "Source ID": "collected-fae9ff261453", "Operation IDs": "271", "Citation": "Adobe · Transcript edits and editable video timeline", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/edit-projects/edit-video-using-text-based-editing/edit-sequences-using-text-based-editing.html" }, { "Source ID": "collected-fdc2aa2ec65e", "Operation IDs": "271", "Citation": "helpx.adobe.com, Last updated 2026-01-07", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/edit-projects/edit-video-using-text-based-editing/overview-of-text-based-editing.html" } ] }, { "ID": "273", "Software": "Adobe Premiere / adb-mcp", "Operation": "LLM Timeline Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2b330cc004ee\ncollected-21e81d38dc4b\ncollected-2999958aade6\ncollected-b0b5a059780c\ncollected-0fd36e49e2d8\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere's recorded adb-mcp workflow uses an LLM to construct a timeline through native commands.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The model determines which existing project items appear, their order, trim points and effect settings, while frame content comes unchanged from the supplied media; this is the which-and-when kind of determination REF10 makes and places the operation nearer REF10 than REF02. The interval stays open below REF10 because the LLM also authors the sequence order and effect settings.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. adb-mcp (described by its author as a proof of concept; Premiere Pro Beta 25.3 build 46 or later) exposes 26 Premiere tools that import media, build sequences from existing project items in a given order, trim clips, add transitions and preset effects, and set clip opacity and blend mode. No Premiere tool calls a generative media model. Not established: The LLM driving the client is chosen by the user and is outside the bridge documentation. Previous reason: Learned structural choices survive conventional command execution. The synthesis-to-region comparison is an explicit cross-domain envelope; evidence does not support one exact t for timeline authorship. raw.githubusercontent.com, main branch, October 2026 • github.com, 214 commits", "Collected source links": "https://github.com/mikechambers/adb-mcp\nhttps://github.com/mikechambers/adb-mcp/blob/main/mcp/pr-mcp.py\nhttps://raw.githubusercontent.com/mikechambers/adb-mcp/main/mcp/pr-mcp.py\nhttps://web.archive.org/web/20260923032504id_/https://helpx.adobe.com/premiere/desktop/organize-media/create-projects/start-projects.html\nhttps://web.archive.org/web/20260929204046id_/https://helpx.adobe.com/premiere/desktop/edit-projects/change-clip-sequence/rearrange-clips-on-the-timeline.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2b330cc004ee", "Operation IDs": "273", "Citation": "Project developer · Adobe MCP supported apps and operations", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/mikechambers/adb-mcp" }, { "Source ID": "collected-21e81d38dc4b", "Operation IDs": "273", "Citation": "B34-P2 Developer source: Premiere MCP tool definitions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/mikechambers/adb-mcp/blob/main/mcp/pr-mcp.py" }, { "Source ID": "collected-2999958aade6", "Operation IDs": "273", "Citation": "raw.githubusercontent.com, main branch, October 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/mikechambers/adb-mcp/main/mcp/pr-mcp.py" }, { "Source ID": "collected-b0b5a059780c", "Operation IDs": "215, 273", "Citation": "web.archive.org — start-projects.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260923032504id_/https://helpx.adobe.com/premiere/desktop/organize-media/create-projects/start-projects.html" }, { "Source ID": "collected-0fd36e49e2d8", "Operation IDs": "273, 395", "Citation": "web.archive.org — rearrange-clips-on-the-timeline.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260929204046id_/https://helpx.adobe.com/premiere/desktop/edit-projects/change-clip-sequence/rearrange-clips-on-the-timeline.html" } ] }, { "ID": "274", "Software": "Prezi", "Operation": "AI visual presentation generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-73e2a75a050d\ncollected-756b1fd45436", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Prezi AI constructs the initial visual presentation; unspecified later revisions are excluded.", "Previous record: rationale": "Learned organization and content construction remain in the presentation. The task crosses the synthesis and region references, so 2–10 is retained as a comparison envelope rather than an invented point.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). Prezi's help article (updated 29 September 2026) states that Prezi AI creates the outline and design from a prompt, that file-based creation adds visuals selected from Prezi's image gallery, and that color palettes and designs come as options that can be regenerated. Generated text and gallery-selected visuals sit together in the result, consistent with the mixed-responsibility envelope kept for Gamma (IRM_OP_116). Not established: The model; whether prompt-based creation also draws on gallery visuals or on generated images; how slide layout is chosen. support.prezi.com, updated 2026-09-29", "Collected source links": "https://prezi.com/features/ai/\nhttps://support.prezi.com/hc/en-us/articles/22436099469847-How-to-use-Prezi-AI-to-design-your-presentation", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-73e2a75a050d", "Operation IDs": "274", "Citation": "Prezi · Prezi overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://prezi.com/features/ai/" }, { "Source ID": "collected-756b1fd45436", "Operation IDs": "274", "Citation": "Prezi AI native presentation workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.prezi.com/hc/en-us/articles/22436099469847-How-to-use-Prezi-AI-to-design-your-presentation" } ] }, { "ID": "275", "Software": "Procreate (iPad)", "Operation": "Manual Painting on Separate Layers Comparison", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e0f00326256b\ncollected-9d67774d60d7\ncollected-4dba2315f4b4\ncollected-b5dbe7371b26\ncollected-d422f9a56759\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Procreate performs manually directed painting in the native layered document.", "Previous record: rationale": "Conventional brush and layer operations contain no documented trained determination in this scope. Imported generated references do not change the branch.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Paint, Smudge and Erase tools; brush library; layers. Returns one result; editable structure: layer. “You can use any brush with Paint, Smudge and Erase.” Source", "Collected source links": "https://procreate.com/ai\nhttps://help.procreate.com/procreate/handbook/brushes/paint-smudge-erase\nhttps://help.procreate.com/procreate/handbook/layers\nhttps://help.procreate.com/procreate/handbook/layers/layers-blend\nhttps://help.procreate.com/procreate/handbook/layers/layers-interface", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e0f00326256b", "Operation IDs": "275", "Citation": "Procreate · Painting", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://procreate.com/ai" }, { "Source ID": "collected-9d67774d60d7", "Operation IDs": "275", "Citation": "Procreate · Paint Smudge and Erase", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.procreate.com/procreate/handbook/brushes/paint-smudge-erase" }, { "Source ID": "collected-4dba2315f4b4", "Operation IDs": "275", "Citation": "help.procreate.com — layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.procreate.com/procreate/handbook/layers" }, { "Source ID": "collected-b5dbe7371b26", "Operation IDs": "275", "Citation": "help.procreate.com — layers-blend", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.procreate.com/procreate/handbook/layers/layers-blend" }, { "Source ID": "collected-d422f9a56759", "Operation IDs": "275", "Citation": "help.procreate.com — layers-interface", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.procreate.com/procreate/handbook/layers/layers-interface" } ] }, { "ID": "276", "Software": "Procreate Dreams", "Operation": "Move and Scale Keyframe Interpolation in Dreams 2.0", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-83ad81912eb2\ncollected-b210fa60a8bf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Procreate Dreams evaluates a specified Move and Scale animation with conventional interpolation.", "Previous record: rationale": "Keyframe interpolation is prescribed computation. Automatic playback and intermediate-frame evaluation do not establish learned inference.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: playhead position; Action menu; Move or Filter keyframe type; keyframe parameters panel; keyframe track; easing. Returns one result; editable structure: keyframe track on timeline. “A keyframe track will appear under the content with a highlighted keyframe and a panel with that keyframe's parameters.” Source • Screenshot", "Collected source links": "https://procreate.com/dreams\nhttps://help.procreate.com/dreams/handbook/2.0/keyframes-and-performing/keyframes", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-83ad81912eb2", "Operation IDs": "276", "Citation": "Procreate · Keyframing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://procreate.com/dreams" }, { "Source ID": "collected-b210fa60a8bf", "Operation IDs": "276", "Citation": "Procreate Dreams 2.0 · Keyframe interpolation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.procreate.com/dreams/handbook/2.0/keyframes-and-performing/keyframes" } ] }, { "ID": "277", "Software": "PromeAI", "Operation": "Sketch to Rendered Image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-58061bc719b4\ncollected-ac7f84d9804b\ncollected-d537035e40a4\ncollected-ce67d2334d76", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "PromeAI Sketch Rendering synthesizes a rendered appearance from the supplied sketch.", "Previous record: rationale": "Current placement unresolved. K=2; M=0; unknown=3; conditional=0; A=5. Lower: 2 + 3 × 2 ÷ (5 − 1) = 3.5. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.5; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: PromeAI's developer LibAI published sdxl-controlnet-lineart-promeai, a lineart ControlNet trained on an SDXL base (Stability AI), and states it is used by promeai.pro. LibAI also published FLUX.1-controlnet-lineart-promeai on FLUX.1-dev (Black Forest Labs), without a stated link to the product. Model choice: Not documented. Not established: The model card names the base as 'sdxl Realistic_Vision_V2.0', a checkpoint name otherwise associated with SD 1.5, so the exact base checkpoint is uncertain. huggingface.co • huggingface.co\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (sketch-conditioned image synthesis). Designer controls: sketch upload; prompt; style and scene templates; advanced settings (perspective, light); reference image; creativity/fidelity slider; 3D model upload (.obj, .fbx, .stl, .3ds). Returns count not documented; a flat image. “Describe the result you want to generate, or directly select the high-quality style and scene templates provided” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.promeai.pro/architecture-sketch-transformation\nhttps://huggingface.co/promeai/sdxl-controlnet-lineart-promeai\nhttps://huggingface.co/promeai/FLUX.1-controlnet-lineart-promeai\nhttps://www.promeai.pro/blender", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-58061bc719b4", "Operation IDs": "277", "Citation": "PromeAI · Sketch Rendering", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.promeai.pro/architecture-sketch-transformation" }, { "Source ID": "collected-ac7f84d9804b", "Operation IDs": "277", "Citation": "huggingface.co", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://huggingface.co/promeai/sdxl-controlnet-lineart-promeai" }, { "Source ID": "collected-d537035e40a4", "Operation IDs": "277", "Citation": "huggingface.co", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://huggingface.co/promeai/FLUX.1-controlnet-lineart-promeai" }, { "Source ID": "collected-ce67d2334d76", "Operation IDs": "277", "Citation": "www.promeai.pro — blender", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.promeai.pro/blender" } ] }, { "ID": "278", "Software": "Radiant Photo 2", "Operation": "Automatic scene-based photo correction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-514fd28ebc4c\ncollected-0b2640a92c80\ncollected-1fb3db2d04bc\ncollected-48481d1403c1\ncollected-6f3b056324c7", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Radiant Photo 2 performs scene analysis followed by automatic image correction.", "Previous record: rationale": "The available account does not separate learned scene decisions from fixed correction rules or establish their retained contribution. Product claims about intelligent enhancement cannot determine a defensible t.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Radiant Photo 2.2 is the current release; its notes extend the Enhance slider to 200% and describe scene-aware enhancement without naming models. Vendor and press pages (November to December 2024) state that AI recognizes subject, lighting and environment and that edits are non-generative and pixel-level, without separating learned scene decisions from fixed correction algorithms. Not established: Whether trained models compute the correction parameters or only classify the scene to select presets that fixed algorithms apply. manual.radiantimaginglabs.com, Radiant Photo 2.2 (date not s • radiantimaginglabs.com, 2024-12-06 • digitalcameraworld.com, 2024-11-25", "Collected source links": "https://radiantimaginglabs.com/radiant-photo-2-workflows/\nhttps://www.digitalcameraworld.com/cameras/radiant-photo-2-review\nhttps://manual.radiantimaginglabs.com/1/en/topic/radiant-photo-2-2\nhttps://radiantimaginglabs.com/radiant-updates/understanding-workflows/\nhttps://www.digitalcameraworld.com/news/radiant-photo-2-unleashes-improved-ai-image-editing-power-and-introduces-specialized-workflows", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-514fd28ebc4c", "Operation IDs": "278", "Citation": "Radiant Photo current workflow description", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://radiantimaginglabs.com/radiant-photo-2-workflows/" }, { "Source ID": "collected-0b2640a92c80", "Operation IDs": "278", "Citation": "Digital Camera World · Radiant Photo Auto Corrections", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.digitalcameraworld.com/cameras/radiant-photo-2-review" }, { "Source ID": "collected-1fb3db2d04bc", "Operation IDs": "278", "Citation": "manual.radiantimaginglabs.com, Radiant Photo 2.2 (date not s", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://manual.radiantimaginglabs.com/1/en/topic/radiant-photo-2-2" }, { "Source ID": "collected-48481d1403c1", "Operation IDs": "278", "Citation": "radiantimaginglabs.com, 2024-12-06", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://radiantimaginglabs.com/radiant-updates/understanding-workflows/" }, { "Source ID": "collected-6f3b056324c7", "Operation IDs": "278", "Citation": "digitalcameraworld.com, 2024-11-25", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.digitalcameraworld.com/news/radiant-photo-2-unleashes-improved-ai-image-editing-power-and-introduces-specialized-workflows" } ] }, { "ID": "279", "Software": "Real-ESRGAN", "Operation": "RealESR General x4v3 image upscaling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-79005cb42d3a\ncollected-664bcbe60da8", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Real-ESRGAN real-general-x4v3 performs native fourfold enlargement with face enhancement disabled.", "Previous record: rationale": "Its trained perceptual reconstruction matches the ESRGAN-family reference at 5. This is a task and objective comparison, not an assumption of identical network architecture or an upscaling-factor score.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0). The xinntao repository has no release after v0.3.0 (20 September 2022) and no push after 6 August 2024; realesr-general-x4v3 remains the tiny general-scene model trained under the synthetic-degradation GAN pipeline. Its -dn option (denoise strength) is a configuration setting the record does not name. Not established: The -dn value used in the recorded configuration. github.com, repository front page, 2024-08 • github.com, v0.2.5.0 model", "Collected source links": "https://github.com/xinntao/Real-ESRGAN\nhttps://github.com/xinntao/Real-ESRGAN/blob/master/docs/model_zoo.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-79005cb42d3a", "Operation IDs": "279", "Citation": "GitHub · Real-ESRGAN", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/xinntao/Real-ESRGAN" }, { "Source ID": "collected-664bcbe60da8", "Operation IDs": "279", "Citation": "github.com, v0.2.5.0 model", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/xinntao/Real-ESRGAN/blob/master/docs/model_zoo.md" } ] }, { "ID": "280", "Software": "RealityKit by Apple", "Operation": "Current Object Capture Photo to Model", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0cc40282cdab\ncollected-4695f4e7ac6c\ncollected-0c671821657e\ncollected-9db1a8e8120d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Apple RealityKit Object Capture reconstructs an object through a managed capture pipeline.", "Previous record: rationale": "The available documentation does not establish which trained feature or depth predictions survive in the reconstructed object. Neural Engine availability is insufficient evidence for either the learned branch or a coordinate.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Apple's WWDC23 session states Object Capture collects LiDAR points to reconstruct low-texture objects and that Mac reconstruction uses the LiDAR data saved in images; WWDC24 describes automatic segmentation with optional custom masks. WWDC21 mentions only the Neural Engine speeding computer vision algorithms, and the WWDC26 RealityKit session adds Gaussian splat rendering without changing Object Capture reconstruction. Not established: Whether automatic segmentation, feature matching or the saved LiDAR depth (ARKit depth is machine-learning fused per WWDC20) are trained models in this pipeline, and what share of the reconstructed geometry they determine. developer.apple.com, WWDC 2023 • developer.apple.com, WWDC 2024 • developer.apple.com, WWDC 2021", "Collected source links": "https://developer.apple.com/documentation/realitykit/realitykit-object-capture/\nhttps://developer.apple.com/videos/play/wwdc2021/10076/\nhttps://developer.apple.com/videos/play/wwdc2023/10191/\nhttps://developer.apple.com/videos/play/wwdc2024/10107/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0cc40282cdab", "Operation IDs": "280", "Citation": "Apple · Apple Object Capture", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.apple.com/documentation/realitykit/realitykit-object-capture/" }, { "Source ID": "collected-4695f4e7ac6c", "Operation IDs": "280", "Citation": "Apple WWDC21 · Object Capture", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.apple.com/videos/play/wwdc2021/10076/" }, { "Source ID": "collected-0c671821657e", "Operation IDs": "280", "Citation": "developer.apple.com, WWDC 2023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.apple.com/videos/play/wwdc2023/10191/" }, { "Source ID": "collected-9db1a8e8120d", "Operation IDs": "280", "Citation": "developer.apple.com, WWDC 2024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.apple.com/videos/play/wwdc2024/10107/" } ] }, { "ID": "281", "Software": "RealityScan by Epic Games", "Operation": "2.2 AI Masks and Photogrammetric Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-bd074c413f81\ncollected-2dee45bcda99\ncollected-a7c92e40850f\ncollected-27827f4baad4\ncollected-31eef5def4c5", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "RealityScan 2.2 reconstructs a mesh using enabled AI-generated masks.", "Previous record: rationale": "The documented learned region decisions constrain the reconstruction and remain consequential in it, supporting 10. Matching and mesh construction do not dilute that responsibility through an automatic hybrid adjustment.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). RealityScan 2.0 documents AI masking as background removal by a trained segmentation model; 2.1.1 (31 March 2026) added AI mask export to COLMAP, and 2.2 (24 June 2026) added AMD GPU support with no documented change to AI masks. Not established: The segmentation model architecture is undisclosed. realityscan.com, 2025-06 (2.0) • realityscan.com, 2026-06-24 (2.2) • radiancefields.com, 2026-03-31 (2.1.1)", "Collected source links": "https://www.realityscan.com/news/realityscan-20-new-release-brings-powerful-new-features-to-a-rebranded-realityscan\nhttps://www.realityscan.com/spotlights/how-cd-projekt-red-used-realityscan-to-create-the-witcher-4-unreal-engine-5-tech-demo\nhttps://www.realityscan.com/news/realityscan-20-new-release-brings-powerful-new-features-to-a-rebranded-realitycapture\nhttps://www.realityscan.com/news/realityscan-2-2-is-here-with-full-amd-gpu-support-download-today\nhttps://radiancefields.com/realityscan-releases-v2.1.1-with-colmap-updates", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-bd074c413f81", "Operation IDs": "281", "Citation": "RealityScan · Photogrammetry", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.realityscan.com/news/realityscan-20-new-release-brings-powerful-new-features-to-a-rebranded-realityscan" }, { "Source ID": "collected-2dee45bcda99", "Operation IDs": "281", "Citation": "RealityScan · AI masking in production", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.realityscan.com/spotlights/how-cd-projekt-red-used-realityscan-to-create-the-witcher-4-unreal-engine-5-tech-demo" }, { "Source ID": "collected-a7c92e40850f", "Operation IDs": "281", "Citation": "realityscan.com, 2025-06 (2.0)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.realityscan.com/news/realityscan-20-new-release-brings-powerful-new-features-to-a-rebranded-realitycapture" }, { "Source ID": "collected-27827f4baad4", "Operation IDs": "281", "Citation": "realityscan.com, 2026-06-24 (2.2)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.realityscan.com/news/realityscan-2-2-is-here-with-full-amd-gpu-support-download-today" }, { "Source ID": "collected-31eef5def4c5", "Operation IDs": "281", "Citation": "radiancefields.com, 2026-03-31 (2.1.1)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://radiancefields.com/realityscan-releases-v2.1.1-with-colmap-updates" } ] }, { "ID": "282", "Software": "Rebelle 8.3", "Operation": "Layered manual painting over a generated reference", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-42a582e3ab34\ncollected-59be6d348063\ncollected-4492e107c3da", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Rebelle 8.3 simulates manually directed painting over an imported reference, before NanoPixel export.", "Previous record: rationale": "Paint-flow and pigment simulation are conventional within this scope. The reference image's production history does not introduce an in-scope learned mechanism.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by a solver or simulation. Designer controls: brush presets; Layers panel (add layer over reference); Color panel; Properties panel; import (File > Import, drag and drop); tracing option; layer masks and clipping masks. Returns one result; editable structure: layer. “Paint pigments color mixing, oil thickness, watercolor diffusion, and NanoPixel technology convincingly mimic the way natural media interact with the canvas and itself.” Source • Screenshot", "Collected source links": "https://www.escapemotions.com/products/rebelle/about\nhttps://escapemotions.com/products/rebelle/manual/8/starting-painting/workflow-options/\nhttps://escapemotions1.b-cdn.net/images/rebelle7/about/gui/eng/r8_gui_dark_en.jpg", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-42a582e3ab34", "Operation IDs": "282", "Citation": "Escape Motions · Rebelle overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.escapemotions.com/products/rebelle/about" }, { "Source ID": "collected-59be6d348063", "Operation IDs": "282", "Citation": "Rebelle 8 — painting from a reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://escapemotions.com/products/rebelle/manual/8/starting-painting/workflow-options/" }, { "Source ID": "collected-4492e107c3da", "Operation IDs": "282", "Citation": "escapemotions1.b-cdn.net — r8_gui_dark_en.jpg", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://escapemotions1.b-cdn.net/images/rebelle7/about/gui/eng/r8_gui_dark_en.jpg" } ] }, { "ID": "283", "Software": "Rebelle 8.3 Pro", "Operation": "NanoPixel 2 enlarged raster export", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1facda7d6061\ncollected-742fdb2d72bf\ncollected-05cc492d057f\ncollected-9e34ecfbc90b\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Rebelle NanoPixel 2 enlarges the painted image for raster export.", "Previous record: rationale": "Earlier NanoPixel machine-learning claims do not establish the current NanoPixel 2 mechanism or its trained contribution. Neither the word fractal nor the absence of a published neural architecture resolves the branch.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). The current Rebelle 8 product page describes Fractal Image Processing as a fractal resizing algorithm based on machine learning, and the 8.2 manual (2025) says Liquify's fractal processing is a similar technology to NanoPixel image resizing. Neither source states the trained task or that NanoPixel 2 export runs a trained model; the manual describes NanoPixel 2 only as delivering sharper detail and exporting up to 20x20 larger. Not established: Whether NanoPixel 2 export runs a trained model and, if so, its objective (detail reconstruction or fidelity); neither the branch nor a reference pair is established. escapemotions.com, current product page (Rebelle 8), fetched • escapemotions.com, manual v8.2, 2025", "Collected source links": "https://escapemotions.com/products/rebelle/documents/Rebelle_Manual_v.8.2.pdf\nhttps://escapemotions.com/products/rebelle/manual/8/\nhttps://www.escapemotions.com/products/rebelle/\nhttps://escapemotions.com/products/rebelle/manual/latest/creating-opening-saving/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1facda7d6061", "Operation IDs": "283", "Citation": "Rebelle current mechanism documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://escapemotions.com/products/rebelle/documents/Rebelle_Manual_v.8.2.pdf" }, { "Source ID": "collected-742fdb2d72bf", "Operation IDs": "283", "Citation": "Rebelle 8 — NanoPixel 2", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://escapemotions.com/products/rebelle/manual/8/" }, { "Source ID": "collected-05cc492d057f", "Operation IDs": "283", "Citation": "escapemotions.com, current product page (Rebelle 8), fetched", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.escapemotions.com/products/rebelle/" }, { "Source ID": "collected-9e34ecfbc90b", "Operation IDs": "283", "Citation": "escapemotions.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://escapemotions.com/products/rebelle/manual/latest/creating-opening-saving/" } ] }, { "ID": "284", "Software": "Recraft Studio V4.1", "Operation": "Text to raster image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-df908c3d7711\ncollected-ae111cd442b9\ncollected-c7074d2ddb59", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Recraft V4.1 produces a raster image through the recorded generation route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Recraft V4.1 raster). Designer controls: prompt; model choice; style; aspect ratio (presets); color palette; number of outputs. Returns several to choose from (1–4 per prompt in Studio per help-page search excerpt; API 1–6); a flat image: PNG, JPG, TIFF, PDF export. “Generate multiple outputs at once.” Source", "Collected source links": "https://www.recraft.ai/blog/recraft-introduces-a-revolutionary-ai-model-that-thinks-in-design-language\nhttps://www.recraft.ai/blog/recraft-v4-1-more-beautiful-by-nature\nhttps://www.recraft.ai/docs/recraft-studio/image-generation/prompt-panel", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-df908c3d7711", "Operation IDs": "284", "Citation": "Recraft · Prompt Generation (raster + native vector)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.recraft.ai/blog/recraft-introduces-a-revolutionary-ai-model-that-thinks-in-design-language" }, { "Source ID": "collected-ae111cd442b9", "Operation IDs": "284, 285", "Citation": "Recraft — V4.1 release and branches", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.recraft.ai/blog/recraft-v4-1-more-beautiful-by-nature" }, { "Source ID": "collected-c7074d2ddb59", "Operation IDs": "284", "Citation": "www.recraft.ai — prompt-panel", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.recraft.ai/docs/recraft-studio/image-generation/prompt-panel" } ] }, { "ID": "285", "Software": "Recraft Studio V4.1 Vector", "Operation": "Text to editable SVG", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-61d019e7de08\ncollected-ae111cd442b9", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Recraft V4.1 Vector generates editable SVG paths and layers.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Recraft V4.1 Vector). Designer controls: prompt; model choice (V4.1 Vector); palette; drop background elements; level of detail. Returns several to choose from (number not stated; 'pick from multiple generations'); editable structure: SVG vector paths, strokes, fills, layers. “Type a prompt and Recraft V4.1 generates a clean SVG — production-ready for logos, posters, icon sets, ads, and mockups.” Source", "Collected source links": "https://www.recraft.ai/ai-vector-generator\nhttps://www.recraft.ai/blog/recraft-v4-1-more-beautiful-by-nature", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-61d019e7de08", "Operation IDs": "285", "Citation": "Recraft — current native vector generator", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.recraft.ai/ai-vector-generator" }, { "Source ID": "collected-ae111cd442b9", "Operation IDs": "284, 285", "Citation": "Recraft — V4.1 release and branches", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.recraft.ai/blog/recraft-v4-1-more-beautiful-by-nature" } ] }, { "ID": "286", "Software": "Remini", "Operation": "Photo Enhance accepted image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b9b94476461b\ncollected-a3180abc8279\ncollected-c902137e911f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Remini Photo Enhance reconstructs image detail through the managed learned enhancement service.", "Previous record: rationale": "The documented reconstruction task supports the fidelity-to-perceptual envelope 5–8. Public service evidence does not establish an objective or checkpoint sufficient for a narrower position.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Remini's web pages describe the enhancer as reconstructing the details of low-quality images and enlarging them up to 2x, and name no model, architecture or training objective. The iOS app (Bending Spoons Apps ApS) is at version 2.10.220 in early October 2026, and remini.ai redirects to the web app at app.remini.ai. Not established: Model identity; whether the enhancer is diffusion-, GAN- or fidelity-trained; its training objective. remini.ai, not shown • apps.apple.com, version 2.10.220", "Collected source links": "https://remini.ai/\nhttps://remini.ai/low-quality-enhancer\nhttps://apps.apple.com/us/app/remini-ai-photo-enhancer/id1470373330", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b9b94476461b", "Operation IDs": "286", "Citation": "Remini · Enhance", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://remini.ai/" }, { "Source ID": "collected-a3180abc8279", "Operation IDs": "286", "Citation": "Remini — photo-quality enhancement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://remini.ai/low-quality-enhancer" }, { "Source ID": "collected-c902137e911f", "Operation IDs": "286", "Citation": "apps.apple.com, version 2.10.220", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://apps.apple.com/us/app/remini-ai-photo-enhancer/id1470373330" } ] }, { "ID": "287", "Software": "Resolume Arena / Avenue", "Operation": "Layered Alpha compositing of generated video", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-49b33f629242\ncollected-5f260e1587de\ncollected-382c4ca7dbd9", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolume Arena or Avenue composites supplied video layers with Alpha blending.", "Previous record: rationale": "Conventional blending and triggering define this operation. Generation of the supplied clips occurs outside the scope, and external neural effects are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: layer order; blend mode (Alpha); layer opacity; master fader. Returns continuous live output; editable structure: layers with blend mode and opacity. “The Alpha blend mode mixes two layers based on opacity.” Source • Screenshot", "Collected source links": "https://resolume.com/software\nhttps://resolume.com/support/en/blend-modes\nhttps://www.resolume.com/support/en/layers", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-49b33f629242", "Operation IDs": "287", "Citation": "Resolume · Software overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://resolume.com/software" }, { "Source ID": "collected-5f260e1587de", "Operation IDs": "287, 288", "Citation": "Resolume — blend-mode mechanisms", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://resolume.com/support/en/blend-modes" }, { "Source ID": "collected-382c4ca7dbd9", "Operation IDs": "287", "Citation": "www.resolume.com — layers", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.resolume.com/support/en/layers" } ] }, { "ID": "288", "Software": "Resolume Wire", "Operation": "Video Mixer Alpha composite in a working patch", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5f260e1587de\ncollected-8d109b118e71", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Resolume Wire evaluates a two-source Video Mixer patch in Alpha mode.", "Previous record: rationale": "The bounded composition is conventional arithmetic over the supplied sources. Node-based organization and earlier source generation do not alter X.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: nodes/parameters; Video Mixer node; blend mode (Alpha default); node search-and-connect; Alt-drag duplicate. Returns count not documented; editable structure: nodes. “This is default blend mode of the Video Mixer node in Resolume Wire.” Source • Screenshot", "Collected source links": "https://resolume.com/support/en/blend-modes\nhttps://resolume.com/support/en/fast-patching", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5f260e1587de", "Operation IDs": "287, 288", "Citation": "Resolume — blend-mode mechanisms", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://resolume.com/support/en/blend-modes" }, { "Source ID": "collected-8d109b118e71", "Operation IDs": "288", "Citation": "Resolume — patch connections", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://resolume.com/support/en/fast-patching" } ] }, { "ID": "289", "Software": "Autodesk Revit", "Operation": "Generative Design Dynamo Study", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-81528b312933\ncollected-953c38c073f1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Revit's Three Box Massing study uses NSGA-II to optimize a specified Dynamo definition.", "Previous record: rationale": "Evolutionary search evaluates candidate parameter combinations without a trained predictor in this example. Optimization and unexpected alternatives do not themselves satisfy the learned branch.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by search or optimization. Designer controls: inputs (box height, X/Y positions); goals (maximize floor area, minimize surface area); method Optimize; Population size; Generations; Number of outcomes; Seed. Returns several to choose from; editable structure: Revit elements from a chosen outcome. “How many outcomes (design alternatives) do you want to generate?” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://help.autodesk.com/cloudhelp/2026/ENU/Revit-GDiR/files/GUID-3B0AA77C-27BA-4967-8026-690D42611FC4.htm\nhttps://help.autodesk.com/cloudhelp/2026/ENU/Revit-GDiR/files/GUID-710D71BE-3376-4647-822A-DA4D793DCBFA.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-81528b312933", "Operation IDs": "289", "Citation": "G1 Autodesk Revit 2026 Three Box Massing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Revit-GDiR/files/GUID-3B0AA77C-27BA-4967-8026-690D42611FC4.htm" }, { "Source ID": "collected-953c38c073f1", "Operation IDs": "289", "Citation": "G2 Autodesk Revit 2026 Generation Settings", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/Revit-GDiR/files/GUID-710D71BE-3376-4647-822A-DA4D793DCBFA.htm" } ] }, { "ID": "290", "Software": "Autodesk Revit", "Operation": "Parametric Building Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-2286ad40aa8f\ncollected-f9cfb0da500c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Revit updates a native Basic Wall when its constrained top level changes.", "Previous record: rationale": "The supplied level, offsets and wall relationships determine the parametric update through conventional computation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: Base Constraint; Base Offset; Top Constraint; Unconnected Height; Top Offset. Returns one result; editable structure: BIM wall element. “whether its height will change when its base constraint or top constraint moves.” Source", "Collected source links": "https://help.autodesk.com/cloudhelp/2027/ENU/Revit-ArchDesign/files/GUID-D69E6B7D-8D8D-40E4-A6D4-7CAF1A927894.htm\nhttps://help.autodesk.com/cloudhelp/2027/ENU/Revit-ArchDesign/files/GUID-ED2045E5-90BF-418F-AAD6-3BB1544F34F5.htm", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-2286ad40aa8f", "Operation IDs": "290", "Citation": "R1 Autodesk Revit 2027 Height and Depth of Walls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Revit-ArchDesign/files/GUID-D69E6B7D-8D8D-40E4-A6D4-7CAF1A927894.htm" }, { "Source ID": "collected-f9cfb0da500c", "Operation IDs": "290", "Citation": "R2 Autodesk Revit 2027 Wall Instance Properties", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2027/ENU/Revit-ArchDesign/files/GUID-ED2045E5-90BF-418F-AAD6-3BB1544F34F5.htm" } ] }, { "ID": "291", "Software": "Rhino 8", "Operation": "Direct SubD Geometry Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-93ae13320c92\ncollected-9183d4fa1d00", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Rhino 8 performs direct editing of SubD geometry.", "Previous record: rationale": "The scoped control-point and surface evaluation operations are conventional, with no retained trained prediction.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: push and pull; conversion to solids. Returns one result; editable structure: SubD geometry. “Push and pull to explore complex, free-form shapes in realtime.” Source", "Collected source links": "https://www.rhino3d.com/features/\nhttps://www.rhino3d.com/features/subd/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-93ae13320c92", "Operation IDs": "291", "Citation": "Rhino · Rhino features", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.rhino3d.com/features/" }, { "Source ID": "collected-9183d4fa1d00", "Operation IDs": "291", "Citation": "McNeel · Direct SubD editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.rhino3d.com/features/subd/" } ] }, { "ID": "292", "Software": "Rhino 8 / Grasshopper", "Operation": "Conventional Parametric Geometry Definition", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4c7e8d4d53e4\ncollected-a6de8c4f4ea3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Grasshopper evaluates a specified graph of conventional geometry components.", "Previous record: rationale": "Graph complexity does not establish learning. Third-party learned components are outside this defined configuration.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: nodes/parameters (components). Returns one result; editable structure: nodes; geometry in Rhino. “Grasshopper® is a graphical algorithm editor tightly integrated with Rhino's 3-D modeling tools.” Source • Screenshot", "Collected source links": "https://www.rhino3d.com/6/new/grasshopper/\nhttps://www.grasshopper3d.com/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4c7e8d4d53e4", "Operation IDs": "292", "Citation": "McNeel · Grasshopper component system", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.rhino3d.com/6/new/grasshopper/" }, { "Source ID": "collected-a6de8c4f4ea3", "Operation IDs": "292", "Citation": "www.grasshopper3d.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.grasshopper3d.com/" } ] }, { "ID": "293", "Software": "Rhino 8 / Grasshopper", "Operation": "Galapagos Evolutionary Geometry Search", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f509dfb7fb4a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Grasshopper Galapagos searches parameters of a conventional geometry definition with an evolutionary solver.", "Previous record: rationale": "The specified search uses no learned surrogate. Its generated alternatives remain conventional under the equation's retained-learning criterion.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Other. Decided by search or optimization. Designer controls: genome (number sliders); fitness input; minimize/maximize; solver choice (evolutionary or simulated annealing); solver settings; Reinstate. Returns several to choose from (ranked genomes; number not documented); editable structure: parameters: slider values reinstated into the definition. “a generic platform for the application of Evolutionary Algorithms to be used on a wide variety of problems by non-programmers” Source Includes a third-party source.", "Collected source links": "https://www.grasshopper3d.com/group/galapagos", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f509dfb7fb4a", "Operation IDs": "293", "Citation": "David Rutten · Galapagos evolutionary computation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.grasshopper3d.com/group/galapagos" } ] }, { "ID": "294", "Software": "Rhino 8 / Grasshopper", "Operation": "RhinoMCP LLM Grasshopper Graph Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-0eaa74ba277f\ncollected-3becc2a18e62", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "RhinoMCP uses an LLM to choose components, connections and parameters for a Grasshopper graph.", "Previous record: rationale": "Learned structural decisions survive as the graph. The synthesis-to-region envelope records a cross-domain comparison; later conventional graph evaluation does not cancel those decisions.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). RhinoMCP 0.4.1.1 (14 September 2026) exposes Grasshopper tools that search the installed component library, add and wire components, set slider, toggle and panel values, and build or mutate a whole graph in one batched call; auto-layout is a separate tool. Code-execution tools (RhinoScript-Python, RhinoCommon C#) act on Rhino geometry outside the graph scope. Not established: The LLM is chosen by the user's MCP client. Documentation does not settle whether composing library components into a new graph sits nearer synthesis or region determination. github.com, README at release 0.4.1.1, 2026-09-14 • github.com, 2026-09-14", "Collected source links": "https://github.com/jingcheng-chen/rhinomcp\nhttps://github.com/jingcheng-chen/rhinomcp/releases", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-0eaa74ba277f", "Operation IDs": "294", "Citation": "RhinoMCP developer documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/jingcheng-chen/rhinomcp" }, { "Source ID": "collected-3becc2a18e62", "Operation IDs": "294", "Citation": "github.com, 2026-09-14", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/jingcheng-chen/rhinomcp/releases" } ] }, { "ID": "295", "Software": "Rive", "Operation": "State-machine animation of generated visual assets", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6f5b453c8cb9\ncollected-ab1136ae3101\ncollected-2b94a2029133", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Rive evaluates supplied animations through an explicitly constructed state machine.", "Previous record: rationale": "States, transitions and bindings determine the visual behavior conventionally. The imported assets' earlier generation is outside this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: states (timeline animations); transitions; inputs; conditions; layers. Returns continuous live output; editable structure: state machine of states and transitions. “State Machines are a visual way to connect animations together and define the logic that drives the transitions.” Source • Screenshot", "Collected source links": "https://rive.app/\nhttps://rive.app/docs/editor/state-machine/state-machine\nhttps://mintcdn.com/rive/IMrXM-oXoMvrTV9Y/images/editor/state-machine/307461c0-2006-4fdf-bdc3-61875d40f422.webp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6f5b453c8cb9", "Operation IDs": "295", "Citation": "Rive · State Machine", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://rive.app/" }, { "Source ID": "collected-ab1136ae3101", "Operation IDs": "295", "Citation": "Rive current state-machine construction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://rive.app/docs/editor/state-machine/state-machine" }, { "Source ID": "collected-2b94a2029133", "Operation IDs": "295", "Citation": "mintcdn.com — 307461c0-2006-4fdf-bdc3-61875d40f422.webp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://mintcdn.com/rive/IMrXM-oXoMvrTV9Y/images/editor/state-machine/307461c0-2006-4fdf-bdc3-61875d40f422.webp" } ] }, { "ID": "296", "Software": "Rodin by Hyper3D", "Operation": "Gen 2.5 Image to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ce6cc6eb6752\ncollected-6270cffedf92\ncollected-019324020fce", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Rodin Gen 2.5 generates mesh geometry from images, before separate material editing or inpainting.", "Previous record: rationale": "Current placement unresolved. K=2; M=0; unknown=0; conditional=2; A=4. Lower: 2 + 3 × 2 ÷ (4 − 1) = 4.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 4.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Rodin Gen-2.5 image to 3D). Designer controls: images (1–5); prompt (optional); tier; mesh_mode (Raw or Quad); quality; texture_mode; geometry_file_format; addons; uhd_texture; quality_override. Returns count not documented; editable structure: mesh with PBR textures; GLB, FBX, OBJ, STL, USDZ and others. “Submit one to five `images`. The first image is used for material generation.” Source", "Collected source links": "https://hyper3d.ai/blog/rodin-gen-2\nhttps://hyper3d.ai/features/image-to-3d\nhttps://docs.hyper3d.ai/en/api-specification/rodin-gen2-5", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ce6cc6eb6752", "Operation IDs": "296", "Citation": "Hyper3D · Diffusion meshes and regional 3D editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://hyper3d.ai/blog/rodin-gen-2" }, { "Source ID": "collected-6270cffedf92", "Operation IDs": "296", "Citation": "Hyper3D · Image mesh output formats", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://hyper3d.ai/features/image-to-3d" }, { "Source ID": "collected-019324020fce", "Operation IDs": "296, 297", "Citation": "Rodin Gen 2.5 API", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.hyper3d.ai/en/api-specification/rodin-gen2-5" } ] }, { "ID": "297", "Software": "Rodin by Hyper3D", "Operation": "Gen 2.5 Text to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-08235665051a\ncollected-019324020fce", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Rodin Gen 2.5 generates mesh geometry from text, before separate material editing or inpainting.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Rodin Gen-2.5 text to 3D). Designer controls: prompt; tier; mesh_mode; quality; texture_mode; geometry_file_format; addons; uhd_texture; quality_override; geometry_instruct_mode. Returns count not documented; editable structure: mesh with PBR textures; STL, GLB, OBJ, FBX. “Generate geometry and proportions from text so you can move from text to 3D model without manual modeling.” Source", "Collected source links": "https://hyper3d.ai/features/text-to-3d\nhttps://docs.hyper3d.ai/en/api-specification/rodin-gen2-5", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-08235665051a", "Operation IDs": "297", "Citation": "Hyper3D · Text mesh output formats", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://hyper3d.ai/features/text-to-3d" }, { "Source ID": "collected-019324020fce", "Operation IDs": "296, 297", "Citation": "Rodin Gen 2.5 API", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.hyper3d.ai/en/api-specification/rodin-gen2-5" } ] }, { "ID": "298", "Software": "Runway", "Operation": "Act-Two Character Video Performance Transfer", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-400e7843f0f6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Runway Act-Two synthesizes character-video facial performance from a driving performance.", "Previous record: rationale": "Current placement unresolved. K=3; M=0; unknown=3; conditional=0; A=6. Lower: 2 + 3 × 3 ÷ (6 − 1) = 3.8. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.8; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Act-Two performance transfer). Designer controls: driving performance video; character image or video; gesture toggle; facial expressiveness (default 3). Returns count not documented; a flat video. “Act-Two allows you to animate characters using driving performance videos.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://help.runwayml.com/hc/en-us/articles/42311337895827-Performance-Capture-with-Act-Two", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-400e7843f0f6", "Operation IDs": "298", "Citation": "Runway · Act-Two input modes and motion preservation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.runwayml.com/hc/en-us/articles/42311337895827-Performance-Capture-with-Act-Two" } ] }, { "ID": "299", "Software": "Runway", "Operation": "Aleph 2.0 / Edit Studio", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6a6476a8f5c5\ncollected-0aae6894f3b0\ncollected-c5fce7aa4ae4", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Runway Edit Studio edits a reference frame with Gen-4 Image and produces revised footage through Aleph 2.0.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=3; conditional=3; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Runway changelog: Aleph 2.0 and Edit Studio released May 13, 2026 (\"Edit a frame the way you want, and Aleph 2.0 edits the rest\"); available through Runway MCP from Jul 14, 2026. X stays 2. Runway changelog\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Gen-4 Image keyframe edit, then Aleph 2.0 video edit). Designer controls: video upload (2–30 s); keyframe selection; keyframe edit prompt; prompt versions tray; Extra motion prompt. Returns count not documented; a flat video. “Single edit generates an image based on a selected keyframe still from your video” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://help.runwayml.com/hc/en-us/articles/51683104370451-Creating-with-Edit-Studio\nhttps://help.runwayml.com/hc/en-us/articles/52150503729171-Aleph-2-0-Prompting-Guide\nhttps://runway.com/changelog", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6a6476a8f5c5", "Operation IDs": "299", "Citation": "Runway · Aleph 2.0 keyframe edits and prompt versions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.runwayml.com/hc/en-us/articles/51683104370451-Creating-with-Edit-Studio" }, { "Source ID": "collected-0aae6894f3b0", "Operation IDs": "299", "Citation": "Runway · Aleph selective changes and preservation claims", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.runwayml.com/hc/en-us/articles/52150503729171-Aleph-2-0-Prompting-Guide" }, { "Source ID": "collected-c5fce7aa4ae4", "Operation IDs": "299, 301, 303", "Citation": "Runway changelog", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://runway.com/changelog" } ] }, { "ID": "300", "Software": "Runway", "Operation": "Gen-4 Image reference-conditioned synthesis", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-aedb35ed3880", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Runway Gen-4 Image synthesizes an image using the recorded reference-conditioned route.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=2; conditional=3; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Gen-4 Image, reference-conditioned). Designer controls: prompt; reference images (up to 3); reference tags; aspect ratio (six options); batch size (1 or 4). Returns several to choose from (1 or 4); a flat image. “You can use up to three References for a single generation.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://help.runwayml.com/hc/en-us/articles/40042718905875-Creating-with-Gen-4-Image-References", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-aedb35ed3880", "Operation IDs": "300", "Citation": "Runway Gen-4 Image References", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.runwayml.com/hc/en-us/articles/40042718905875-Creating-with-Gen-4-Image-References" } ] }, { "ID": "301", "Software": "Runway", "Operation": "Gen-4.5 Video Generation", "Proposed X": "2", "Review status": "Proposed X calculated", "Previous X": "2.0", "S": "5", "U": "5", "J": "5", "Reviewed on": "2026-10-14", "Configuration": "Text-only video generation from a high-level scene/action request, new clip; no prescribed layout, trajectory or event staging, source-video protection or image-to-video starting frame.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "High-level scene/action prompt and video settings; text-only route without prescribed trajectories, event staging or source-frame protection.", "Learned determination": "Video diffusion synthesizes new scene content and evolving relationships across time.", "Conventional processing": "Encoding/playback presents the learned clip.", "Model contribution in the result": "Learned temporal synthesis remains; failures of consistency concern quality, not whether this responsibility was delegated.", "Why this scope score (S)": "S5 whole clip; U5 new instance; J5 learned synthesis across time supported by Runway’s video-diffusion and temporal-consistency account. Merely saving a video file would not support J5.", "Why this latitude score (U)": "S5 whole clip; U5 new instance; J5 learned synthesis across time supported by Runway’s video-diffusion and temporal-consistency account. Merely saving a video file would not support J5.", "Why this coordination score (J)": "S5 whole clip; U5 new instance; J5 learned synthesis across time supported by Runway’s video-diffusion and temporal-consistency account. Merely saving a video file would not support J5.", "Comparisons and other modes": "X2.0 versus SD3.5 still synthesis X2.9. Documented temporal failures limit quality, not the fact that temporal relationships are assigned to the learned task. Image-to-video has different supplied constraints and must be separately assessed.", "Citations": "Runway Gen4.5 research account — https://runway.com/research/introducing-runway-gen-4.5\nRunway Gen4.5 creation instructions — https://help.runwayml.com/hc/en-us/articles/46974685288467-Creating-with-Gen-4-5", "Source IDs": "CAL2-RUNWAY\nCAL2-RUNHELP", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human specifies scene and motion in language. Learned video diffusion synthesizes content and evolving relationships; encoding/playback presents that contribution.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X2.0 versus SD3.5 still synthesis X2.9. Documented temporal failures limit quality, not the fact that temporal relationships are assigned to the learned task. Image-to-video has different supplied constraints and must be separately assessed.", "Previous record: scope": "Runway Gen-4.5 synthesizes a video in the scoped generation-only route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Runway changelog: Gen-4.5 released Dec 11, 2025; image-to-video with a first frame added Jan 21, 2026. X stays 2. Runway changelog\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (Gen-4.5 video). Designer controls: prompt; duration (2–10 s); aspect ratio (six options); frame rate (24 or 25 fps). Returns count not documented; a flat video. “Gen-4.5 Text to Video supports durations of 2-10 seconds and costs 12 credits per second.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://help.runwayml.com/hc/en-us/articles/46974685288467-Creating-with-Gen-4-5\nhttps://help.runwayml.com/hc/en-us/sections/54758948966547-Third-Party-Video-Models\nhttps://runway.com/changelog\nhttps://web.archive.org/web/20260906060634id_/https://help.runwayml.com/hc/en-us/articles/37425232841875-Getting-Started-with-Generative-Video\nhttps://web.archive.org/web/20260803200249id_/https://help.runwayml.com/hc/en-us/articles/46974685288467-Creating-with-Gen-4-5\nhttps://web.archive.org/web/20250809062430id_/https://help.runwayml.com/hc/en-us/articles/26115194070675-Can-I-generate-videos-in-4K", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "2", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-301", "Sources": [ { "Source ID": "CAL2-RUNWAY", "Operation IDs": "301", "Citation": "Runway Gen4.5 research account", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Runway Gen4.5 research account", "Finding used in this review": "Video diffusion and temporally consistent evolving actions; documented limitations remain.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://runway.com/research/introducing-runway-gen-4.5" }, { "Source ID": "CAL2-RUNHELP", "Operation IDs": "301", "Citation": "Runway Gen4.5 creation instructions", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Runway Gen4.5 creation instructions", "Finding used in this review": "Text-to-video and image-to-video configurations; use text-only generation for the reference case.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://help.runwayml.com/hc/en-us/articles/46974685288467-Creating-with-Gen-4-5" } ] }, { "ID": "302", "Software": "Runway", "Operation": "Ruby SDR-to-HDR video conversion", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-276985ceb212\ncollected-4ebc367638c0\ncollected-a5464b2d5d6c\ncollected-fac0efd3f68c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Runway Ruby predicts an HDR appearance from SDR video.", "Previous record: rationale": "The learned appearance-mapping task is compared with fixed-function transformation and perceptual reconstruction, supporting 5–12. The public workflow account does not establish a precise training objective or t.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Runway's API documentation (September 2026) states Ruby converts SDR to HDR keeping the source pixels and audio; since 12 September 2026 the ACEScg EXR deliveries are the source plate passed through the inverse ACES SDR Output Transform with highlight detail recovered by Ruby added on top. The learned contribution is range expansion plus highlight recovery; no whole-frame re-synthesis is documented. Not established: Training objective of the highlight recovery (fidelity or perceptual/adversarial) and how the hdr10 display grade is built. Unclipped content follows a mapping, while recovered highlights are inferred content, so the evidence does not exclude either end of 5–12. Runway help center refused fetch. docs.dev.runwayml.com, 2026-09 • docs.dev.runwayml.com, 2026-09-12 • vp-land.com, 2026-09-10", "Collected source links": "https://help.runwayml.com/hc/en-us/articles/54749891346579-Creating-HDR-videos-with-Runway-Ruby\nhttps://docs.dev.runwayml.com/guides/models/\nhttps://docs.dev.runwayml.com/llms-full.txt\nhttps://www.vp-land.com/stories/runways-new-ruby-model-converts-sdr-video-into-16-bit-hdr-prores-and-exr", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-276985ceb212", "Operation IDs": "302", "Citation": "Runway Ruby conversion and output state", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.runwayml.com/hc/en-us/articles/54749891346579-Creating-HDR-videos-with-Runway-Ruby" }, { "Source ID": "collected-4ebc367638c0", "Operation IDs": "302", "Citation": "docs.dev.runwayml.com, 2026-09", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.dev.runwayml.com/guides/models/" }, { "Source ID": "collected-a5464b2d5d6c", "Operation IDs": "302", "Citation": "docs.dev.runwayml.com, 2026-09-12", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.dev.runwayml.com/llms-full.txt" }, { "Source ID": "collected-fac0efd3f68c", "Operation IDs": "302", "Citation": "vp-land.com, 2026-09-10", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.vp-land.com/stories/runways-new-ruby-model-converts-sdr-video-into-16-bit-hdr-prores-and-exr" } ] }, { "ID": "303", "Software": "Runway", "Operation": "Workflows Video Generation and Stitching", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-381fd8e6252a\ncollected-c5fce7aa4ae4", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Runway Workflows combines two learned video-generation nodes with a conventional stitching node.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 6 model-decided, 0 unknown and 0 setting-dependent categories among 6 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=6; unknown=0; conditional=0; A=6. Lower: 2 + 3 × 0 ÷ (6 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 3 Oct 2026 Runway changelog: Workflows (Oct 24, 2025) chain multiple models in node graphs. Model Routers (Jul 21, 2026) select models automatically by cost, latency or quality within allow/deny lists. When a router is on, the model in each node is chosen by the router. X stays 2. Runway changelog\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (two video-generation nodes; stitch node runs by rule). Designer controls: nodes/parameters; node links; templates; Stitch input order. Returns count not documented; a flat video: one stitched video. “Workflows offer a node-based approach to media generation that gives you granular control over every step of your creative process.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://help.runwayml.com/hc/en-us/articles/45763528999699-Introduction-to-Workflows\nhttps://runway.com/changelog", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-381fd8e6252a", "Operation IDs": "303", "Citation": "Runway · Workflows graph editing and output locks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.runwayml.com/hc/en-us/articles/45763528999699-Introduction-to-Workflows" }, { "Source ID": "collected-c5fce7aa4ae4", "Operation IDs": "299, 301, 303", "Citation": "Runway changelog", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://runway.com/changelog" } ] }, { "ID": "304", "Software": "Shap-E", "Operation": "Text to Mesh Research Example", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5f82bff01831", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Shap-E's research pipeline predicts an implicit representation and decodes a generated mesh.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (text-conditioned implicit 3D diffusion). Designer controls: prompt; guidance_scale; karras_steps; batch_size; render_mode. Returns several to choose from (4 in example notebook, batch_size 4); editable structure: mesh written as PLY or OBJ. “sample a 3D model, conditioned on a text prompt” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://github.com/openai/shap-e", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5f82bff01831", "Operation IDs": "304", "Citation": "GitHub · shap-e", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/openai/shap-e" } ] }, { "ID": "305", "Software": "Shapr3D", "Operation": "History Based Parametric Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-bc845678fb96\ncollected-beac6a0921b0", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Shapr3D evaluates a native history-based parametric modeling sequence.", "Previous record: rationale": "The scoped feature history and geometric constraints use conventional computation. This does not classify every imported or externally generated asset as a learned operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: history timeline; edit earlier steps; design parameters; advanced fillet settings; sketch projection parameter. Returns one result; editable structure: history timeline and parametric model. “This enables you to retroactively adjust design parameters, with subsequent features updating in response.” Source", "Collected source links": "https://support.shapr3d.com/hc/en-us/articles/13444210101788-5-590-History-Based-Parametric-Modeling-is-here\nhttps://www.shapr3d.com/blog/product-updates-history-based-parametric-modeling-progress", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-bc845678fb96", "Operation IDs": "305", "Citation": "Shapr3D · Modeling", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.shapr3d.com/hc/en-us/articles/13444210101788-5-590-History-Based-Parametric-Modeling-is-here" }, { "Source ID": "collected-beac6a0921b0", "Operation IDs": "305", "Citation": "www.shapr3d.com — product-updates-history-based-parametric-modeling-progress", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.shapr3d.com/blog/product-updates-history-based-parametric-modeling-progress" } ] }, { "ID": "306", "Software": "SHIMA SEIKI APEXFiz Design-Knit V-10", "Operation": "Virtual fabric preview with pattern and color revision", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f079c8572a6f\ncollected-4227f18bd149\ncollected-03e2932cea59\ncollected-09ae70353c2d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SHIMA SEIKI APEXFiz Design-Knit V-10 previews fabric with pattern and color changes.", "Previous record: rationale": "The current preview solver and any retained trained component are not established. A generative-image handoff is unnecessary for eligibility, but it would not resolve this mechanism gap either.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). APEXFiz V-10 was released 14 September 2026; its notice adds 3D Loop Simulation with multiple textures and texture-map export to CLO. The only disclosed AI model is in V-09's Search Similar Pattern menu (19 January 2026), a function separate from the fabric preview; public pages describe virtual fabric from actual yarn data without describing the simulation method. Not established: How Loop Simulation computes the preview and whether any trained component contributes; technical detail sits in the account-only SHIMA HelpCenter. shimaseiki.com, V-10, 2026-09-14 • shimaseiki.com, V-09, 2026-01-19 • shimaseiki.com, current page", "Collected source links": "https://www.shimaseiki.com/product/design/\nhttps://www.shimaseiki.com/product/design/software/\nhttps://www.shimaseiki.com/product/design/software/news/APEXFizDesign_V10_NewFunctions_EN_20260914.pdf\nhttps://www.shimaseiki.com/product/design/software/news/APEXFizDesign_V09_NewFunctions_EN_20260119.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f079c8572a6f", "Operation IDs": "306", "Citation": "Shima Seiki · Shima Seiki APEX Virtual Sampling", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.shimaseiki.com/product/design/" }, { "Source ID": "collected-4227f18bd149", "Operation IDs": "306", "Citation": "SHIMA SEIKI · Fabric simulation and virtual samples", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.shimaseiki.com/product/design/software/" }, { "Source ID": "collected-03e2932cea59", "Operation IDs": "306", "Citation": "shimaseiki.com, V-10, 2026-09-14", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.shimaseiki.com/product/design/software/news/APEXFizDesign_V10_NewFunctions_EN_20260914.pdf" }, { "Source ID": "collected-09ae70353c2d", "Operation IDs": "306", "Citation": "shimaseiki.com, V-09, 2026-01-19", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.shimaseiki.com/product/design/software/news/APEXFizDesign_V09_NewFunctions_EN_20260119.pdf" } ] }, { "ID": "307", "Software": "SideFX Houdini", "Operation": "Houdini Copernicus", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d60b148bd6da\ncollected-d498b222d40f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Houdini Copernicus evaluates the specified conventional image-processing graph.", "Previous record: rationale": "The scoped nodes perform prescribed image computation. Learned-model nodes are excluded; node count supplies no inference points.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: nodes/parameters (COP nodes). Returns continuous live output; editable structure: nodes. “COP nodes provide real-time image manipulation within a 3D space.” Source", "Collected source links": "https://www.sidefx.com/docs/houdini/copernicus/index.html\nhttps://www.sidefx.com/docs/houdini/nodes/cop/index.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d60b148bd6da", "Operation IDs": "307", "Citation": "SideFX · Copernicus image-processing scope", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/copernicus/index.html" }, { "Source ID": "collected-d498b222d40f", "Operation IDs": "307", "Citation": "SideFX · COP nodes and bypass controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/cop/index.html" } ] }, { "ID": "308", "Software": "SideFX Houdini", "Operation": "Houdini MCP Node Graph Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d87e06167f35\ncollected-7bc68d2d3631", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "The Houdini MCP workflow uses an LLM to construct a visual node graph.", "Previous record: rationale": "Learned graph structure survives native node creation. The synthesis-to-region envelope records the cross-domain comparison; later execution of the graph is outside this authorship scope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). A commit of 11 June 2026 to capoomgit/houdini-mcp added 13 structured tools, including node connection, batch parameter setting, network layout and VEX wrangle creation and editing, and documented raw Python execution as a last resort. Within graph construction the LLM selects library nodes and can also write new VEX code inside wrangle nodes; the repository has no tagged release and no later commits. Not established: Documentation does not settle a position within 2–10 for graphs that combine library nodes with authored VEX. github.com, 2026-06-11 • github.com", "Collected source links": "https://github.com/capoomgit/houdini-mcp\nhttps://github.com/capoomgit/houdini-mcp/commits/main", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d87e06167f35", "Operation IDs": "308", "Citation": "Project developer · Houdini MCP node creation and modification", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/capoomgit/houdini-mcp" }, { "Source ID": "collected-7bc68d2d3631", "Operation IDs": "308", "Citation": "github.com, 2026-06-11", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/capoomgit/houdini-mcp/commits/main" } ] }, { "ID": "309", "Software": "SideFX Houdini", "Operation": "Houdini ML Deformer", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-a26fd6c1e318\ncollected-7ca32a6c0a07\ncollected-7c6ca3faa104", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Houdini 21 ML Deform predicts weights that combine residual blend shapes to approximate the trained deformation mapping.", "Previous record: rationale": "The fixed-function emulation reference at 12 is the closer task comparison. Explicit blend-shape evaluation retains the learned weights and does not justify a conventional-only result.", "Previous record: calculation": "X = 12 + 0 × (12 − 12) = 12.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (12.0). Houdini 22 (covered by CG Channel on 16 July 2026) supersedes Houdini 21; the 22.0 ML Deform documentation keeps the same mechanism, a trained model predicting weights that combine residual blend shapes into a skin correction. Houdini 22 adds Agent Add ML Deformer, which applies models from the same ML Train Deformer recipe to crowd agents. sidefx.com, Houdini 22.0 docs • sidefx.com, Houdini 22.0", "Collected source links": "https://www.sidefx.com/docs/houdini/nodes/sop/ml_deform.html\nhttps://www.sidefx.com/docs/houdini/news/21/ml.html\nhttps://www.sidefx.com/docs/houdini22.0/news/22/ml.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-a26fd6c1e318", "Operation IDs": "309", "Citation": "SideFX · ML Deform weights and residual shapes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/sop/ml_deform.html" }, { "Source ID": "collected-7ca32a6c0a07", "Operation IDs": "309", "Citation": "SideFX · Houdini 21 deformer introduction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/news/21/ml.html" }, { "Source ID": "collected-7c6ca3faa104", "Operation IDs": "309", "Citation": "sidefx.com, Houdini 22.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini22.0/news/22/ml.html" } ] }, { "ID": "310", "Software": "SideFX Houdini", "Operation": "Houdini ML Pyro Upres", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b205d740b109\ncollected-87767041120d\ncollected-42ddde250fa8\ncollected-a8674ee2f616", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Houdini ML Volume Upres predicts higher-resolution density tiles from a lower-resolution pyro simulation.", "Previous record: rationale": "Learned reconstruction supports the fidelity-to-perceptual envelope 5–8. Optional sharpening and other-field upsampling are excluded, and current evidence does not fix a point within that envelope.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Houdini 22.0 documentation describes ML Volume Upres (available since 21.0) as applying an ONNX model trained with the ML Train Volume Upres recipe, which builds its data set from tiles of a cached high-res pyro simulation; SideFX states the model makes a low-res pyro sim look more like a high-res pyro sim. The recipe page names no loss; the separate ML Train Regression node lists mean squared error as its only standard loss. Not established: Whether the volume upres recipe trains with ML Train Regression's MSE loss or another objective. sidefx.com, Houdini 22.0 • sidefx.com, Houdini 22.0 • sidefx.com, Houdini 22.0", "Collected source links": "https://www.sidefx.com/docs/houdini/nodes/top/ml_trainregression.html\nhttps://www.sidefx.com/docs/houdini/nodes/sop/ml_volumeupres.html\nhttps://www.sidefx.com/docs/houdini/ml/train_solutions/nodes_volume.html\nhttps://www.sidefx.com/docs/houdini/ml/train_solutions/ml_trainvolumeupres.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b205d740b109", "Operation IDs": "310", "Citation": "SideFX · Neural regression and volume-upres training", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/top/ml_trainregression.html" }, { "Source ID": "collected-87767041120d", "Operation IDs": "310", "Citation": "SideFX · ML Volume Upres inputs and processing controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/sop/ml_volumeupres.html" }, { "Source ID": "collected-42ddde250fa8", "Operation IDs": "310", "Citation": "sidefx.com, Houdini 22.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/ml/train_solutions/nodes_volume.html" }, { "Source ID": "collected-a8674ee2f616", "Operation IDs": "310", "Citation": "sidefx.com, Houdini 22.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/ml/train_solutions/ml_trainvolumeupres.html" } ] }, { "ID": "311", "Software": "SideFX Houdini", "Operation": "Houdini ML Style Transfer", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9a249505c7fb\ncollected-694c883e5cda\ncollected-5ad04fe0f7ab\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Houdini trains and runs the documented GAN and U-Net style-transfer or colorization mapping.", "Previous record: rationale": "The learned image transformation changes appearance. The fixed-function-to-perceptual envelope 5–12 accommodates the documented task family without assuming unrestricted scene generation or a particular loss weighting.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Houdini 22.0 documentation for ML Train Style Transfer (node since 21.0) describes a generic training node that trains a GAN with a U-Net generator in paired or unpaired mode, with a user-set L1 Lambda; listed uses include terrain sketches to height fields and colorization. The mechanism matches the record. Not established: The trained task depends on the user's training pairs and L1 weighting; a model trained on pairs made by a fixed operator would emulate that operator (REF12), while a detail-synthesis pairing sits nearer REF05, so the node documentation does not fix a position. sidefx.com, Houdini 22.0", "Collected source links": "https://www.sidefx.com/docs/houdini/nodes/top/ml_trainstyletransfer.html\nhttps://www.sidefx.com/docs/houdini/ref/menus.html\nhttps://www.sidefx.com/docs/houdini/basics/intro.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9a249505c7fb", "Operation IDs": "311", "Citation": "SideFX · Style-transfer neural architecture and output tasks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/top/ml_trainstyletransfer.html" }, { "Source ID": "collected-694c883e5cda", "Operation IDs": "311, 313", "Citation": "www.sidefx.com — menus.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/ref/menus.html" }, { "Source ID": "collected-5ad04fe0f7ab", "Operation IDs": "311, 313", "Citation": "www.sidefx.com — intro.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/basics/intro.html" } ] }, { "ID": "312", "Software": "SideFX Houdini", "Operation": "Houdini Neural Point Surface", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-19de75b7c78c\ncollected-4f6b339a4b6e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Houdini Neural Point Surface predicts a VDB surface using a pretrained convolutional model.", "Previous record: rationale": "The surface reconstruction preserves geometry while resolving missing detail and material-specific appearance. A cross-domain fidelity-to-perceptual envelope 5–8 is more defensible than equating every CNN with unrestricted generation.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Houdini 22.0 documentation describes Neural Point Surface as turning a point cloud into a VDB surface with a pretrained convolutional network and offers four material models: Balanced (default), Liquid, Granular and Smooth. Training data, ground truth and loss are not published. Not established: Training data, ground truth and loss for the four pretrained models. sidefx.com, Houdini 22.0", "Collected source links": "https://www.sidefx.com/docs/houdini/nodes/sop/neuralpointsurface.html\nhttps://www.sidefx.com/docs/houdini/mpm/nps.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-19de75b7c78c", "Operation IDs": "312", "Citation": "SideFX · Neural Point Surface model and revision controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/sop/neuralpointsurface.html" }, { "Source ID": "collected-4f6b339a4b6e", "Operation IDs": "312", "Citation": "sidefx.com, Houdini 22.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/mpm/nps.html" } ] }, { "ID": "313", "Software": "SideFX Houdini", "Operation": "Houdini SOP Networks", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-109837171cb1\ncollected-694c883e5cda\ncollected-5ad04fe0f7ab\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Houdini evaluates a specified conventional SOP network for geometry and attributes.", "Previous record: rationale": "The defined nodes contain no learned model. Procedural complexity and downstream propagation do not change the branch.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: nodes/parameters (SOP nodes); node flags. Returns one result; editable structure: nodes; geometry with attributes. “Geometry nodes live inside Geo objects and generate geometry.” Source", "Collected source links": "https://www.sidefx.com/docs/houdini/nodes/sop/index.html\nhttps://www.sidefx.com/docs/houdini/ref/menus.html\nhttps://www.sidefx.com/docs/houdini/basics/intro.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-109837171cb1", "Operation IDs": "313", "Citation": "SideFX · SOP network structure", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/sop/index.html" }, { "Source ID": "collected-694c883e5cda", "Operation IDs": "311, 313", "Citation": "www.sidefx.com — menus.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/ref/menus.html" }, { "Source ID": "collected-5ad04fe0f7ab", "Operation IDs": "311, 313", "Citation": "www.sidefx.com — intro.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/basics/intro.html" } ] }, { "ID": "314", "Software": "SideFX Houdini", "Operation": "ONNX Inference SOP ML Terrain Example", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3dc77d18d9e0\ncollected-7aca5abe82b0\ncollected-39b805c0d47c\ncollected-cf957bae80a7", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SideFX's ML Terrain example uses a specified Pix2Pix ONNX model to predict detailed terrain from a painted heightfield.", "Previous record: rationale": "Learned terrain transformation supports the fixed-function-to-perceptual envelope 5–12 by cross-domain analogy. ONNX is an execution format and supplies no coordinate by itself.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). The ML Terrain example (published 2 November 2023, Houdini 20.5) trains its Pix2Pix model on image pairs exported from terrains produced by procedural terrain HDAs, then generates a detailed heightfield from a painted heightfield. The task is learned emulation of prescribed terrain processors under a conditional adversarial objective, so both the REF12 and REF05 comparisons stay relevant. Not established: Exact input/target pairing and the L1-to-adversarial weighting of the provided ONNX model. A newer SideFX example, Neural Terrain Paint (Houdini 21, August 2025), uses an ML Train Terrain HDA of unstated architecture and is outside this record. sidefx.com, 2023-11-02 • arxiv.org, v3, 2018-11-26 • sidefx.com, 2025-08-18", "Collected source links": "https://www.sidefx.com/contentlibrary/ml-terrain/\nhttps://www.sidefx.com/docs/houdini/nodes/sop/onnx.html\nhttps://arxiv.org/abs/1611.07004\nhttps://www.sidefx.com/contentlibrary/neural-terrain-paint/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3dc77d18d9e0", "Operation IDs": "314", "Citation": "SideFX · Pix2Pix ONNX terrain and painted-input workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/contentlibrary/ml-terrain/" }, { "Source ID": "collected-7aca5abe82b0", "Operation IDs": "314", "Citation": "SideFX · ONNX Inference inputs and visual examples", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/docs/houdini/nodes/sop/onnx.html" }, { "Source ID": "collected-39b805c0d47c", "Operation IDs": "314", "Citation": "arxiv.org, v3, 2018-11-26", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://arxiv.org/abs/1611.07004" }, { "Source ID": "collected-cf957bae80a7", "Operation IDs": "314", "Citation": "sidefx.com, 2025-08-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sidefx.com/contentlibrary/neural-terrain-paint/" } ] }, { "ID": "315", "Software": "Sketch", "Operation": "Native image vector and layer editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-560fcb67ad79\ncollected-ebdde4b1f0dc\ncollected-c25972313f3b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Sketch performs direct native image, vector and layer editing with conventional layout relationships.", "Previous record: rationale": "External assistants are excluded. The earlier whole-application label is confined here to these documented native editing operations, without requiring generated source material.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: selection; Magic Wand; crop; fill; Color Adjust (Inspector); background removal; replace image; resize to original; symbols and overrides. Returns one result; editable structure: layers, vector paths, symbols. “You can make non-destructive color adjustments to images using the Color Adjust effect in the Inspector.” Source • Screenshot", "Collected source links": "https://www.sketch.com/blog/sketch-and-ai/\nhttps://www.sketch.com/docs/symbols-and-styles/symbols/\nhttps://www.sketch.com/docs/designing/images/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-560fcb67ad79", "Operation IDs": "315", "Citation": "Sketch · Sketch overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sketch.com/blog/sketch-and-ai/" }, { "Source ID": "collected-ebdde4b1f0dc", "Operation IDs": "315", "Citation": "Sketch · Symbols and overrides", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sketch.com/docs/symbols-and-styles/symbols/" }, { "Source ID": "collected-c25972313f3b", "Operation IDs": "315", "Citation": "www.sketch.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sketch.com/docs/designing/images/" } ] }, { "ID": "316", "Software": "SketchUp", "Operation": "AI Generate Object Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d5cafe093c76\ncollected-f8871a449706\ncollected-2445119f2b60\ncollected-d0d8e2ff8c20\ncollected-a1a687f81b7d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SketchUp AI Generate predicts a mesh asset through the documented image or text route.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Model choice: No model choice is documented; options are image or text input, up to four images, and mesh simplification. Not established: Trimble's help pages, technical notes and the December 2025 launch press release do not name the 3D generation model or provider. help.sketchup.com, accessed 2026-10-04 • news.trimble.com, 2025-12-10\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (image or text to 3D object). Designer controls: images (up to 4) or text prompt; additional prompt; Simplify Mesh toggle. Returns one result; editable structure: mesh object placed in the SketchUp model, photoreal materials. “When your model is ready, SketchUp Assistant shows you a preview of the generated model, along with the details it used to generate it.” Source • Screenshot", "Collected source links": "https://help.sketchup.com/en/ai-generate\nhttps://news.trimble.com/Trimble-Launches-SketchUp-AI-a-New-Suite-of-AI-Powered-Modeling-Visualization-and-Help-Tools\nhttps://help.sketchup.com/en/sketchup-ai-faq\nhttps://help.sketchup.com/en/generating-object-image\nhttps://images.ctfassets.net/yd21npj2ir9g/38444cceed9346e521651b/e497d06f264e31a39ad155a4ab1c00b9/desktop-ai-assistant-generate-object-upload-window.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d5cafe093c76", "Operation IDs": "316", "Citation": "SketchUp · AI object generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/ai-generate" }, { "Source ID": "collected-f8871a449706", "Operation IDs": "316", "Citation": "news.trimble.com, 2025-12-10", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://news.trimble.com/Trimble-Launches-SketchUp-AI-a-New-Suite-of-AI-Powered-Modeling-Visualization-and-Help-Tools" }, { "Source ID": "collected-2445119f2b60", "Operation IDs": "316", "Citation": "SketchUp AI FAQ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/sketchup-ai-faq" }, { "Source ID": "collected-d0d8e2ff8c20", "Operation IDs": "316", "Citation": "help.sketchup.com — generating-object-image", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/generating-object-image" }, { "Source ID": "collected-a1a687f81b7d", "Operation IDs": "316", "Citation": "images.ctfassets.net — desktop-ai-assistant-generate-object-upload-window.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://images.ctfassets.net/yd21npj2ir9g/38444cceed9346e521651b/e497d06f264e31a39ad155a4ab1c00b9/desktop-ai-assistant-generate-object-upload-window.png" } ] }, { "ID": "317", "Software": "SketchUp", "Operation": "AI Render Image Output", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-894f8875166b\ncollected-75335f542500\ncollected-d4dbe744a19f\ncollected-d2272670c6a2\ncollected-3955260ab883", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SketchUp AI Render synthesizes an image from a viewport and the selected generation controls.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=2; conditional=2; A=5. Lower: 2 + 3 × 1 ÷ (5 − 1) = 2.75. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.75; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Gemini 3.1 Flash Image Preview (Nano Banana 2), Google, across all presets and inpainting tools from AI Render extension version 2.5.0. Model choice: No. Not established: The model behind SketchUp Diffusion (2023 to 2025) is not named; the technical notes describe AI Render as a latent text-to-image diffusion model, wording that matches the Stable Diffusion model card, but no provider is stated for that period. help.sketchup.com • extensions.sketchup.com\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (viewport-conditioned image synthesis). Designer controls: text prompt; presets (Auto or preset prompts); reference image (No Style preset); negative prompt; Erase, Paint and Sketch tools (brush or lasso); window size; settings sliders (geometry adherence). Returns count not documented; a flat image: images in Gallery; can be saved as a Scene. “Whatever you type becomes a text prompt that generates an image using the model viewport or generated image displayed in the AI Render window.” Source • Screenshot", "Collected source links": "https://sketchup.trimble.com/blog/en-US/article/realize-your-vision-faster-with-diffusion\nhttps://help.sketchup.com/en/using-ai-render\nhttps://help.sketchup.com/en/technical-notes\nhttps://extensions.sketchup.com/extension/7bfb6011-ee6b-4c16-9cce-9d35a3ad5b8c/ai-render\nhttps://help.sketchup.com/en/generating-output-text-prompts", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-894f8875166b", "Operation IDs": "317", "Citation": "Trimble · AI Render (formerly Diffusion)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://sketchup.trimble.com/blog/en-US/article/realize-your-vision-faster-with-diffusion" }, { "Source ID": "collected-75335f542500", "Operation IDs": "317", "Citation": "SketchUp · AI rendering and optional overlays", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/using-ai-render" }, { "Source ID": "collected-d4dbe744a19f", "Operation IDs": "317", "Citation": "help.sketchup.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/technical-notes" }, { "Source ID": "collected-d2272670c6a2", "Operation IDs": "317", "Citation": "extensions.sketchup.com", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://extensions.sketchup.com/extension/7bfb6011-ee6b-4c16-9cce-9d35a3ad5b8c/ai-render" }, { "Source ID": "collected-3955260ab883", "Operation IDs": "317", "Citation": "help.sketchup.com — generating-output-text-prompts", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/generating-output-text-prompts" } ] }, { "ID": "318", "Software": "SketchUp", "Operation": "Component Definition and Instance Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-57030c08200d\ncollected-3601f0786ec9\ncollected-0b2826ff8816\ncollected-bf075f74bc37\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SketchUp evaluates manually constructed component definitions and instances.", "Previous record: rationale": "Conventional geometry and instance relationships determine this scoped model operation; inference snapping is not evidence of a trained model.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: Make Component dialog; name; description; Glue To; Cut Opening; Always Face Camera; Set Component Axes; Replace selection with component; edit definition or single instance. Returns one result; editable structure: component definitions and instances. “Editing a component's definition will edit all instances of that same component in your model.” Source • Screenshot", "Collected source links": "https://help.sketchup.com/en/sketchup/creating-basic-component\nhttps://images.ctfassets.net/yd21npj2ir9g/a830d5582a4e8da6878482/ab1917c6ee2c11cb56272005835aabf0/su-3000124-create-component-2018.png\nhttps://help.sketchup.com/en/sketchup/organizing-component-collections\nhttps://help.sketchup.com/en/sketchup/basic-edits", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-57030c08200d", "Operation IDs": "318", "Citation": "SketchUp · Component definitions and instances", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/sketchup/creating-basic-component" }, { "Source ID": "collected-3601f0786ec9", "Operation IDs": "318", "Citation": "images.ctfassets.net — su-3000124-create-component-2018.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://images.ctfassets.net/yd21npj2ir9g/a830d5582a4e8da6878482/ab1917c6ee2c11cb56272005835aabf0/su-3000124-create-component-2018.png" }, { "Source ID": "collected-0b2826ff8816", "Operation IDs": "318", "Citation": "help.sketchup.com — organizing-component-collections", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/sketchup/organizing-component-collections" }, { "Source ID": "collected-bf075f74bc37", "Operation IDs": "318", "Citation": "help.sketchup.com — basic-edits", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.sketchup.com/en/sketchup/basic-edits" } ] }, { "ID": "319", "Software": "Sloyd", "Operation": "AI Template Assembly from Text", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5facf1f8c575\ncollected-002c0c2862ac", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Sloyd's Template Editor uses AI to choose and assemble parametric components from text.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. In the Template Editor route, the learned contribution selects templates, part variations and parameter values; geometry comes from artist-made parametric parts. That places the comparison nearer region determination than synthesis, at 6–10 within the same envelope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. Sloyd's text-to-3D page (October 2026) states that inside the Template Editor the AI assembles models from a library of artist-made parametric parts and steers toward a template and variation. Sloyd separately offers fully AI-generated meshes from prompts and image-to-3D as distinct routes outside this operation. Not established: The model and its training task; whether part positions are fixed by templates or set by the model. Previous reason: Learned structural decisions remain in a conventionally evaluated model. The synthesis-to-region envelope 2–10 records that cross-domain responsibility without giving native parameters an automatic X bonus. sloyd.ai, page as of 3 Oct 2026 • sloyd.ai, page as of 3 Oct 2026", "Collected source links": "https://www.sloyd.ai\nhttps://www.sloyd.ai/text-to-3d", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5facf1f8c575", "Operation IDs": "319", "Citation": "Sloyd · Product overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sloyd.ai" }, { "Source ID": "collected-002c0c2862ac", "Operation IDs": "319", "Citation": "Sloyd · Text generation and template assembly distinction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sloyd.ai/text-to-3d" } ] }, { "ID": "320", "Software": "Sloyd", "Operation": "Image to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-609c304a27d8\ncollected-9abecbf25b59\ncollected-7b286c8e433c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Sloyd's image-to-3D route predicts a mesh before later sculpting, rigging or assembly.", "Previous record: rationale": "Current placement unresolved. K=3; M=0; unknown=1; conditional=0; A=4. Lower: 2 + 3 × 3 ÷ (4 − 1) = 5.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 5.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Not documented. Model choice: No model choice is documented. Not established: Sloyd does not state whether image-to-mesh runs an in-house model, a third-party model, or a match to its parametric templates. sloyd.ai, 2025-12-16 • sloyd.gitbook.io, accessed 2026-10-04\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (image to 3D). Designer controls: image (PNG or JPEG); targetFaceCount; textureResolution; topology (auto, quads, triangles). Returns count not documented; editable structure: mesh; textured, or untextured geometry. “`'none'` produces a geometry-only model (no texture).” Source", "Collected source links": "https://api-dashboard.sloyd.ai/documentation/endpoints/3d-generation\nhttps://www.sloyd.ai/blog/turning-reference-images-into-3d-models\nhttps://sloyd.gitbook.io/documentation", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-609c304a27d8", "Operation IDs": "320", "Citation": "Sloyd · Mesh generation output settings", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://api-dashboard.sloyd.ai/documentation/endpoints/3d-generation" }, { "Source ID": "collected-9abecbf25b59", "Operation IDs": "320", "Citation": "sloyd.ai, 2025-12-16", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sloyd.ai/blog/turning-reference-images-into-3d-models" }, { "Source ID": "collected-7b286c8e433c", "Operation IDs": "320", "Citation": "sloyd.gitbook.io, accessed 2026-10-04", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://sloyd.gitbook.io/documentation" } ] }, { "ID": "321", "Software": "Sloyd", "Operation": "Manual Parametric Template Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-502bddb6cdf3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Sloyd evaluates manually chosen parametric template controls.", "Previous record: rationale": "The manual template route is conventional; the separate AI assembly and free-form mesh generators are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: template choice; sliders for shape, proportions and details. Returns one result; editable structure: parameters: template sliders. “Adjust shapes, proportions, and details with sliders” Source • Screenshot", "Collected source links": "https://www.sloyd.ai/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-502bddb6cdf3", "Operation IDs": "321", "Citation": "Sloyd · Parametric templates and separate generated meshes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.sloyd.ai/" } ] }, { "ID": "322", "Software": "SolidWorks", "Operation": "Parametric Feature Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8c975ed27912\ncollected-ae3d2a57e6fa", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SolidWorks evaluates a native parametric feature model with its design history.", "Previous record: rationale": "Feature and constraint evaluation are conventional within the scoped operation. This does not classify separate learned assistants or simulation surrogates.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: features; FeatureManager design tree; rollback bar. Returns one result; editable structure: feature tree. “SolidWorks is a solid modeler, and uses a parametric feature-based approach” Source Includes a third-party source.", "Collected source links": "https://help.solidworks.com/2025/English/SolidWorks/sldworks/c_rollback_bar.htm\nhttps://en.wikipedia.org/wiki/SolidWorks", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8c975ed27912", "Operation IDs": "322", "Citation": "SOLIDWORKS · Rollback and editable feature history", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.solidworks.com/2025/English/SolidWorks/sldworks/c_rollback_bar.htm" }, { "Source ID": "collected-ae3d2a57e6fa", "Operation IDs": "322", "Citation": "en.wikipedia.org — SolidWorks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://en.wikipedia.org/wiki/SolidWorks" } ] }, { "ID": "323", "Software": "SPAR3D (Stable Point Aware 3D)", "Operation": "Single Image to Mesh Neural Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-db513db8393e\ncollected-a3b8d4a25a2b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SPAR3D predicts a point representation and reconstructs a mesh from a single image.", "Previous record: rationale": "Current placement unresolved. K=2; M=0; unknown=2; conditional=0; A=4. Lower: 2 + 3 × 2 ÷ (4 − 1) = 4.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 4.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (point cloud prediction, then mesh). Designer controls: image; point cloud edits (delete, duplicate, stretch, add features, recolor). Returns count not documented; editable structure: mesh with UV texture; editable intermediate point cloud. “Allows users to directly edit the point cloud by deleting, duplicating, stretching, adding features, or recoloring points.” Source", "Collected source links": "https://stability.ai/news/stable-point-aware-3d\nhttps://stability.ai/news-updates/stable-point-aware-3d", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-db513db8393e", "Operation IDs": "323", "Citation": "Stability AI · Single-Image Reconstruction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news/stable-point-aware-3d" }, { "Source ID": "collected-a3b8d4a25a2b", "Operation IDs": "323", "Citation": "Stability AI · Point diffusion and triplane mesh generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news-updates/stable-point-aware-3d" } ] }, { "ID": "324", "Software": "Spline", "Operation": "AI Generated Mesh in Editor", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9e42dc018631\ncollected-d722f7f47879", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Spline's AI Generate route produces a mesh through a hosted learned generator.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Hunyuan3D (Tencent) by default; alternatives Hunyuan 3.1 Pro (Tencent), Trellis 2 (Microsoft TRELLIS.2) and Meshy (Meshy). Model choice: Yes. Not established: The version of the default Hunyuan3D entry (2.x or 3.0) and of the Meshy entry is not stated. docs.spline.design\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (hosted 3D generator). Designer controls: prompt or image; pick one of four looks. Returns several to choose from (4); editable structure: GLB mesh with PBR base color, metallic-roughness and normal maps. “Describe an object or drop in an image. Pick from four looks, then turn the one you want into a textured 3D model.” Source", "Collected source links": "https://spline.design/ai-generate\nhttps://docs.spline.design/generate/ai-3d-generation", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9e42dc018631", "Operation IDs": "324", "Citation": "Spline · AI 3D Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://spline.design/ai-generate" }, { "Source ID": "collected-d722f7f47879", "Operation IDs": "324", "Citation": "docs.spline.design", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.spline.design/generate/ai-3d-generation" } ] }, { "ID": "325", "Software": "Spline", "Operation": "Parametric Object Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-631345718892\ncollected-e6ae1a64bd2c\ncollected-8969a30c0298", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Spline evaluates manually specified native parametric objects and their properties.", "Previous record: rationale": "The defined editing route is conventional. Generated mesh imports and Spell world generation are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: parametric object creation (+ icon on toolbar); object parameters. Returns one result; editable structure: parametric objects. “A parametric object is an object that always retains its original parameters after being modified.” Source • Screenshot", "Collected source links": "https://spline.design\nhttps://docs.spline.design/designing-in-3-d/objects/working-with-parametric-objects\nhttps://cdn.spline.design/_assets/docs/4f427794-2253-42bf-b46b-26f2ae3a4abb.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-631345718892", "Operation IDs": "325", "Citation": "Spline · Design Editor", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://spline.design" }, { "Source ID": "collected-e6ae1a64bd2c", "Operation IDs": "325", "Citation": "Spline · Parametric object editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.spline.design/designing-in-3-d/objects/working-with-parametric-objects" }, { "Source ID": "collected-8969a30c0298", "Operation IDs": "325", "Citation": "cdn.spline.design — 4f427794-2253-42bf-b46b-26f2ae3a4abb.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.spline.design/_assets/docs/4f427794-2253-42bf-b46b-26f2ae3a4abb.png" } ] }, { "ID": "326", "Software": "Spline (Spell)", "Operation": "Spell World Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9dc8f74f5cdb\ncollected-9a6d61fa8fb1", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Spline Spell uses diffusion to generate an initial world volume.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Spell, Spline's own experimental diffusion model, producing a volume rendered as Gaussian splats (or NeRF). Model choice: No model choice is documented. Not established: Spline does not disclose Spell's architecture, size, base model or training data, or whether it is fine-tuned from a public model. docs.spline.design • blog.spline.design, 2025-01-22\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Spell diffusion world model). Designer controls: image. Returns count not documented; editable structure: Gaussian splat volume; also video or image sequences, convertible to mesh. “At its core, Spell is a type of diffusion model that can generate 3D worlds with realistic multi-view consistency.” Source", "Collected source links": "https://blog.spline.design/introducing-spell\nhttps://docs.spline.design/spline-ai/spell-ai-3d-worlds", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9dc8f74f5cdb", "Operation IDs": "326", "Citation": "Spline · Spell World Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.spline.design/introducing-spell" }, { "Source ID": "collected-9a6d61fa8fb1", "Operation IDs": "326", "Citation": "docs.spline.design", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.spline.design/spline-ai/spell-ai-3d-worlds" } ] }, { "ID": "327", "Software": "Stable Diffusion WebUI Forge", "Operation": "SD3.5 Large text-to-image output", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c270207d8e51\ncollected-c8f4506dc46c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Forge executes SD3.5 Large text-to-image generation without optional LoRAs, ControlNets or later editing.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (SD3.5 Large, MMDiT). Designer controls: prompt; negative prompt; model choice (checkpoint); sampler; sampling steps; CFG scale; seed; width/height; batch count; batch size. Returns several to choose from (batch count × batch size, user-set); a flat image. “class StableDiffusion3(ForgeDiffusionEngine):” Source Includes a third-party source.", "Collected source links": "https://github.com/lllyasviel/stable-diffusion-webui-forge\nhttps://github.com/lllyasviel/stable-diffusion-webui-forge/blob/main/backend/diffusion_engine/sd35.py", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c270207d8e51", "Operation IDs": "327", "Citation": "GitHub · stable-diffusion-webui-forge", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/lllyasviel/stable-diffusion-webui-forge" }, { "Source ID": "collected-c8f4506dc46c", "Operation IDs": "327", "Citation": "Forge — SD3.5 engine", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/lllyasviel/stable-diffusion-webui-forge/blob/main/backend/diffusion_engine/sd35.py" } ] }, { "ID": "328", "Software": "Adobe Substance 3D Painter", "Operation": "Procedural Smart Material Layer Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-685292aeed18\ncollected-510ad5deae52\ncollected-a9f7b137d67b\ncollected-30d2b80987b0", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Substance Painter constructs Smart Materials through conventional layers, generators and baked mesh maps.", "Previous record: rationale": "The documented procedural route contains no trained predictor. The word Smart does not establish learning.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: layer stack; generators and their parameters; baked mesh maps (Position, Curvature, World Space Normal); create smart material from folder; smart masks. Returns one result; editable structure: layer stack saved as smart material preset. “Generators are substances that generate a mask or textures based on the mesh topology using baked utility maps” Source • Screenshot", "Collected source links": "https://helpx.adobe.com/substance-3d-painter/using/materials-smart-materials.html\nhttps://experienceleague.adobe.com/en/docs/substance-3d-painter/using/getting-started/glossary\nhttps://experienceleague.adobe.com/en/docs/substance-3d-painter/using/effects/generator\nhttps://experienceleague.adobe.com/en/docs/substance-3d-painter/using/features/smart-materials-and-masks", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-685292aeed18", "Operation IDs": "328", "Citation": "Adobe · Smart Materials", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/substance-3d-painter/using/materials-smart-materials.html" }, { "Source ID": "collected-510ad5deae52", "Operation IDs": "328", "Citation": "Adobe · Procedural smart materials and masks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://experienceleague.adobe.com/en/docs/substance-3d-painter/using/getting-started/glossary" }, { "Source ID": "collected-a9f7b137d67b", "Operation IDs": "328", "Citation": "experienceleague.adobe.com — generator", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://experienceleague.adobe.com/en/docs/substance-3d-painter/using/effects/generator" }, { "Source ID": "collected-30d2b80987b0", "Operation IDs": "328", "Citation": "experienceleague.adobe.com — smart-materials-and-masks", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://experienceleague.adobe.com/en/docs/substance-3d-painter/using/features/smart-materials-and-masks" } ] }, { "ID": "329", "Software": "Adobe Substance 3D Sampler", "Operation": "Image to Material AI Filter", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-be095ed3e071\ncollected-313dfa8aeaec\ncollected-41c99006b41d\ncollected-e5492da1ea6c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Substance Sampler's Image to Material AI filter estimates material maps from a photograph.", "Previous record: rationale": "Learned decomposition of appearance into material properties crosses fixed-function mapping and perceptual reconstruction, supporting 5–12. The map count is not an inference weight.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Image to Material (AI) has no documented model change since Sampler 4.2–4.2.1 (September 2023) through 6.0.4 (24 September 2026); the 6.0.1 removal (16 April 2026) concerns the separate Generative AI panel. Adobe describes two neural models, one for de-lighting and one for geometry reconstruction, trained on procedurally generated materials; the training objective is not stated. Not established: Training objective (supervised regression of maps or generative, as in Adobe's ControlMat research) and whether the shipped 4.2 model descends from MaterIA (2022) cascaded U-Nets. experienceleague.adobe.com, 2026-04-07 • research.adobe.com, 2024-11-27 • experienceleague.adobe.com, 4.2.1, 2023-09-21", "Collected source links": "https://research.adobe.com/news/behind-the-tech-image-to-material-in-adobe-substance-3d-sampler-and-adobe-capture\nhttps://research.adobe.com/news/how-adobe-researchers-simplified-capturing-and-applying-materials-in-3d/\nhttps://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/filters/tools/image-to-material\nhttps://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/release-notes/all-changes", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-be095ed3e071", "Operation IDs": "329", "Citation": "Adobe · Image to Material", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.adobe.com/news/behind-the-tech-image-to-material-in-adobe-substance-3d-sampler-and-adobe-capture" }, { "Source ID": "collected-313dfa8aeaec", "Operation IDs": "329", "Citation": "Adobe Research · Image to material neural models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.adobe.com/news/how-adobe-researchers-simplified-capturing-and-applying-materials-in-3d/" }, { "Source ID": "collected-41c99006b41d", "Operation IDs": "329", "Citation": "experienceleague.adobe.com, 2026-04-07", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/filters/tools/image-to-material" }, { "Source ID": "collected-e5492da1ea6c", "Operation IDs": "329, 330", "Citation": "experienceleague.adobe.com, 4.2.1, 2023-09-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/release-notes/all-changes" } ] }, { "ID": "330", "Software": "Adobe Substance 3D Sampler", "Operation": "Text to Texture", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d98ecc710ac2\ncollected-e5492da1ea6c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Substance Sampler Text to Texture uses Firefly to synthesize a texture image.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Firefly). Designer controls: prompt; Type dropdown (Texture); Generate. Returns several to choose from (4); a flat image: texture image, then dragged into the material creation template or added to layers. “Each use of the feature will generate four variations.” Source", "Collected source links": "https://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/features-and-workflows/generative-workflows\nhttps://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/release-notes/all-changes", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d98ecc710ac2", "Operation IDs": "330", "Citation": "Adobe · Text to Texture", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/features-and-workflows/generative-workflows" }, { "Source ID": "collected-e5492da1ea6c", "Operation IDs": "329, 330", "Citation": "experienceleague.adobe.com, 4.2.1, 2023-09-21", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://experienceleague.adobe.com/en/docs/substance-3d-sampler/using/release-notes/all-changes" } ] }, { "ID": "331", "Software": "SuperSplat", "Operation": "Direct Gaussian Selection and Transform", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-02a3b644aba1\ncollected-7194ba658db2\ncollected-b5fb27463978", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SuperSplat performs direct selection, transformation and cleanup of supplied Gaussian primitives.", "Previous record: rationale": "These editing operations are conventional even when the supplied representation was learned earlier. This boundary excludes fitting the scene representation itself.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: selection tools (rect, brush, lasso, sphere); transform gizmo; Position, Rotation and uniform Scale fields; per-splat color, opacity and scale; delete; export (PLY, compressed PLY, SOG). Returns one result; editable structure: Gaussian splat scene; exports PLY, compressed PLY or SOG. “SuperSplat can transform an entire imported splat or only the Gaussians selected within it.” Source • Screenshot", "Collected source links": "https://playcanvas.com/products/supersplat\nhttps://developer.playcanvas.com/user-manual/supersplat/editor/transform-measure-align/\nhttps://playcanvas.com/assets/images/products/supersplat/supersplat-editor.jpg", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-02a3b644aba1", "Operation IDs": "331", "Citation": "PlayCanvas · Editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://playcanvas.com/products/supersplat" }, { "Source ID": "collected-7194ba658db2", "Operation IDs": "331", "Citation": "PlayCanvas · Direct selection and splat transforms", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.playcanvas.com/user-manual/supersplat/editor/transform-measure-align/" }, { "Source ID": "collected-b5fb27463978", "Operation IDs": "331", "Citation": "playcanvas.com — supersplat-editor.jpg", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://playcanvas.com/assets/images/products/supersplat/supersplat-editor.jpg" } ] }, { "ID": "332", "Software": "SWAPP (over Autodesk Revit / Archicad)", "Operation": "Frank Native Drawing Annotation Automation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-10b482ca6e10\ncollected-7cda753b585e\ncollected-3faeb3d4a767", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "SWAPP Frank predicts and places native drawing annotations using the BIM model and learned documentation practices.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. The learned contribution chooses which native annotations, views and sheet arrangements to create and where to place them; tag and dimension values come from the BIM model through native annotation objects. That places the comparison nearer region determination than synthesis, at 6–10 within the same envelope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. SWAPP's documentation page (October 2026) states that Frank produces dimensions, tags, views and sheets as native Revit and ArchiCAD objects grounded in a firm's production history, and adjusts annotation placement to firm standards. The FAQ states that Frank builds a persistent model of firm standards from existing projects. No dates or model names are given. Not established: Whether the persistent model of firm standards is a trained model or extracted rules; which model plans annotation choices. Previous reason: Learned visual organization survives in the drawings. The synthesis-to-region envelope is a cross-domain comparison; it does not classify the separate 2D-to-3D modeling product. swapp.ai, page as of 3 Oct 2026 • swapp.ai, page as of 3 Oct 2026", "Collected source links": "https://swapp.ai/\nhttps://swapp.ai/products/documentation/\nhttps://swapp.ai/faq/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-10b482ca6e10", "Operation IDs": "332", "Citation": "SWAPP · Documentation Automation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://swapp.ai/" }, { "Source ID": "collected-7cda753b585e", "Operation IDs": "332", "Citation": "SWAPP · Frank documentation automation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://swapp.ai/products/documentation/" }, { "Source ID": "collected-3faeb3d4a767", "Operation IDs": "332", "Citation": "swapp.ai, page as of 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://swapp.ai/faq/" } ] }, { "ID": "333", "Software": "TattoosAI", "Operation": "Prompt to tattoo design image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-eb69257c14b5\ncollected-d04c19468c9f\ncollected-9f4f7e66597c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TattoosAI generates a tattoo-concept image from text and style controls.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Stable Diffusion (Stability AI) through the DreamStudio API, as stated by the developer at launch in September 2022; version not stated. Model choice: No model choice is documented; the user picks style and color options that are composed into a prompt. Not established: The developer's 2022 launch post is the sole model statement located. news.ycombinator.com, 2022-09-08 • tattoosai.com\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (text-to-image). Designer controls: prompt; style dropdown (18+ styles). Returns several to choose from (number not stated; 'a few unique concepts'); a flat image. “a few unique concepts based on the inputs which you've provided” Source", "Collected source links": "https://tattoosai.com/\nhttps://www.tattoosai.com/\nhttps://news.ycombinator.com/item?id=32761238", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-eb69257c14b5", "Operation IDs": "333", "Citation": "TattoosAI · Prompt Generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://tattoosai.com/" }, { "Source ID": "collected-d04c19468c9f", "Operation IDs": "333", "Citation": "TattoosAI prompt-to-design workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.tattoosai.com/" }, { "Source ID": "collected-9f4f7e66597c", "Operation IDs": "333", "Citation": "news.ycombinator.com, 2022-09-08", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://news.ycombinator.com/item?id=32761238" } ] }, { "ID": "334", "Software": "TestFit", "Operation": "Site Solver Live Parametric Layout", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-dbfffa79da00\ncollected-b29b25973d2c\ncollected-63d2af1a7905\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TestFit Site Solver evaluates its documented procedural layout engine under the chosen site constraints.", "Previous record: rationale": "The vendor explicitly states that its engine does not run on a trained model. That evidence supports 15 for the solver-only boundary; deterministic execution alone would not establish absence of learning.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by procedural generation. Designer controls: site constraints; setbacks; parking ratios; unit mix presets; height; FAR; building configurators; core placement; financial parameters. Returns several to choose from (number not documented); form not documented. “Select a preferred solution and regenerate similar solutions to refine your design options.” Source", "Collected source links": "https://www.testfit.io/blog/testfit-mcp-feasibility-studies-in-plain-language\nhttps://www.testfit.io/product/site-solver\nhttps://support.testfit.io/knowledge/testfit-glossary", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-dbfffa79da00", "Operation IDs": "334", "Citation": "TestFit engine and trained-model distinction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.testfit.io/blog/testfit-mcp-feasibility-studies-in-plain-language" }, { "Source ID": "collected-b29b25973d2c", "Operation IDs": "334", "Citation": "TestFit · Site Solver", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.testfit.io/product/site-solver" }, { "Source ID": "collected-63d2af1a7905", "Operation IDs": "334", "Citation": "support.testfit.io — testfit-glossary", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://support.testfit.io/knowledge/testfit-glossary" } ] }, { "ID": "335", "Software": "Topaz Gigapixel", "Operation": "Core Upscale — Standard 2", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "7.3", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5cccdd590eb9\ncollected-507550bc216f\ncollected-37637bcfd671\ncollected-6733964ca189", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Gigapixel Standard 2 reconstructs an enlarged image with Face Recovery disabled.", "Previous record: rationale": "Rule B (3 Oct 2026): vendor documentation states fidelity-oriented training or non-generative intent for an operation that creates missing samples, which follows the Lightroom Super Resolution comparison: t = 0.75 between 5 and 8, sensitivity t = 0.5–1 (6.5–8).", "Previous record: calculation": "X = 5 + 0.75 × (8 − 5) = 7.25. Sensitivity t = 0.5–1 gives 6.5–8.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 7.3. Topaz's developer page for Standard 2 describes clean enlargement that maintains true-to-input detail and structure, and the Gigapixel family page states that precision models attempt to improve quality without changing source characteristics. The Gigapixel core-models page classifies core models, Standard 2 included, as non-generative and directs generative needs to Recover, Wonder and Redefine. Not established: Training loss (pixel, perceptual or adversarial terms) and architecture of Standard 2. Previous reason: The core model's detail-preserving reconstruction supports 5–8 between fidelity and perceptual references. The published model description does not supply a training objective precise enough to select one t. developer.topazlabs.com, not shown • developer.topazlabs.com, not shown • docs.topazlabs.com, not shown", "Collected source links": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/ai-models/core-models\nhttps://www.krea.ai/docs/api-reference/image-enhance/topaz\nhttps://developer.topazlabs.com/image-models/gigapixel/standard-2\nhttps://developer.topazlabs.com/image-models/gigapixel", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5cccdd590eb9", "Operation IDs": "145, 335", "Citation": "docs.topazlabs.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/ai-models/core-models" }, { "Source ID": "collected-507550bc216f", "Operation IDs": "335", "Citation": "Krea Topaz API model choices", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.krea.ai/docs/api-reference/image-enhance/topaz" }, { "Source ID": "collected-37637bcfd671", "Operation IDs": "335", "Citation": "developer.topazlabs.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/gigapixel/standard-2" }, { "Source ID": "collected-6733964ca189", "Operation IDs": "335", "Citation": "developer.topazlabs.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/gigapixel" } ] }, { "ID": "336", "Software": "Topaz Gigapixel", "Operation": "Face Recovery 3 with Strength Adjustment", "Proposed X": "6.3", "Review status": "Proposed X calculated", "Previous X": "5.0", "S": "3", "U": "4", "J": "3", "Reviewed on": "2026-10-14", "Configuration": "Face Recovery 3 on selected detected faces at nonzero strength; assess face-recovery contribution. Any separate upscale model must be accounted for in a combined endpoint.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Image, detected/selected faces and nonzero Face Recovery 3 strength.", "Learned determination": "Face recovery reconstructs plausible facial features.", "Conventional processing": "Detection/selection and blending govern where recovered faces are applied.", "Model contribution in the result": "The nonzero face-recovery contribution remains. Separate upscale models require their own dependency account in a combined endpoint.", "Why this scope score (S)": "S3 fixed face component; U4 plausible reconstruction; J3 coherent relationships within each face, the named reference domain. This is functional expert coding, not a disclosed architecture.", "Why this latitude score (U)": "S3 fixed face component; U4 plausible reconstruction; J3 coherent relationships within each face, the named reference domain. This is functional expert coding, not a disclosed architecture.", "Why this coordination score (J)": "S3 fixed face component; U4 plausible reconstruction; J3 coherent relationships within each face, the named reference domain. This is functional expert coding, not a disclosed architecture.", "Comparisons and other modes": "X6.3. Strength approaching zero does not continuously rescale X. At zero, remove face recovery and reassess every remaining stage; other upscalers can still be learned. J2 alternative gives X7.2 and is retained as coding sensitivity.", "Citations": "Topaz Face Recovery 3 model documentation — https://developer.topazlabs.com/image-models/gigapixel/face-recovery-3\nTopaz face recovery controls — https://docs.topazlabs.com/topaz-gigapixel/enhancements/face-recovery\nTopaz Gigapixel v1.3.0 discussion May 13 2026 — https://community.topazlabs.com/t/topaz-gigapixel-v1-3-0-recover-faces-3-300-speed-up-for-neuroserver-models/102815", "Source IDs": "CAL2-FACE\nCAL2-FACECONTROL\nCAL2-FACERELEASE", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies image, face selection and strength. Learned processing reconstructs plausible facial features; blending retains the contribution at nonzero strength.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X6.3. Strength approaching zero does not continuously rescale X. At zero, remove face recovery and reassess every remaining stage; other upscalers can still be learned. J2 alternative gives X7.2 and is retained as coding sensitivity.", "Previous record: scope": "Gigapixel Face Recovery 3 reconstructs selected faces and applies the result with strength adjustment.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Gigapixel v1.3.0 (13 May 2026) introduced Face Recovery 3, which the Gigapixel documentation calls a generative recover faces model that processes each face separately. Topaz developer documentation states that Face Recovery 3 reconstructs plausible features when facial information is missing, and the Gigapixel family page lists generative and non-generative face recovery options. Not established: Architecture and loss (GAN, diffusion or other) are undisclosed. The developer API page lists two variants that trade fidelity against recovery strength; which variant the Gigapixel desktop app runs is not stated. Previous reason: Learned face reconstruction supports 5–8. The conventional application stage retains the prediction and supplies no automatic reduction or score proportional to strength. docs.topazlabs.com, accessed 3 Oct 2026 • community.topazlabs.com, v1.3.0, 13 May 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/face-recovery\nhttps://community.topazlabs.com/t/topaz-gigapixel-v1-3-0-recover-faces-3-300-speed-up-for-neuroserver-models/102815\nhttps://developer.topazlabs.com/image-models/gigapixel/face-recovery-3", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "19/3", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-336", "Sources": [ { "Source ID": "CAL2-FACE", "Operation IDs": "336", "Citation": "Topaz Face Recovery 3 model documentation", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Topaz Face Recovery 3 model documentation", "Finding used in this review": "Reconstructs plausible facial features and coherent faces; changes to characteristics are possible.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/gigapixel/face-recovery-3" }, { "Source ID": "CAL2-FACECONTROL", "Operation IDs": "336", "Citation": "Topaz face recovery controls", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Topaz face recovery controls", "Finding used in this review": "Face selection, strength/blending and processing components; isolate face result from other enhancement stages.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/face-recovery" }, { "Source ID": "CAL2-FACERELEASE", "Operation IDs": "336", "Citation": "Topaz Gigapixel v1.3.0 discussion May 13 2026", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Topaz Gigapixel v1.3.0 discussion May 13 2026", "Finding used in this review": "Current-version release/user lead; speed is not a scope or latitude change.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-gigapixel-v1-3-0-recover-faces-3-300-speed-up-for-neuroserver-models/102815" } ] }, { "ID": "337", "Software": "Topaz Gigapixel", "Operation": "Redefine Creative Upscale", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1c1e61c91c95\ncollected-a477493bb7c5\ncollected-a7be350cd5f4\ncollected-fb6a656981a4\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Gigapixel Redefine performs the documented creative diffusion-based reconstruction of an image.", "Previous record: rationale": "The perceptual-reconstruction reference at 5 is the closest task comparison. A diffusion architecture does not by itself equate a creative enhancement task with unrestricted scene synthesis.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0). Current Gigapixel docs split Redefine into two models, one fidelity-oriented (None or Subtle levels) and one creative (Low to Max, with Texture and image-description controls), and state Redefine runs locally or in the cloud. The record's creative diffusion-based scope corresponds to the creative model; the mechanism class is unchanged. Not established: Training objective and creativity level in the recorded configuration; whether the fidelity-oriented Redefine model needs a separate record. docs.topazlabs.com, not shown", "Collected source links": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/ai-models/generative-models\nhttps://www.topazlabs.com/learn/responsible-nonfiction-ai-enhancement\nhttps://docs.topazlabs.com/topaz-gigapixel/functions/compare\nhttps://docs.topazlabs.com/topaz-gigapixel/enhancements/settings", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1c1e61c91c95", "Operation IDs": "337", "Citation": "Topaz Gigapixel current generative model guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/ai-models/generative-models" }, { "Source ID": "collected-a477493bb7c5", "Operation IDs": "337", "Citation": "Topaz explanation of precision and diffusion enhancement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.topazlabs.com/learn/responsible-nonfiction-ai-enhancement" }, { "Source ID": "collected-a7be350cd5f4", "Operation IDs": "337", "Citation": "docs.topazlabs.com — compare", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-gigapixel/functions/compare" }, { "Source ID": "collected-fb6a656981a4", "Operation IDs": "337", "Citation": "docs.topazlabs.com — settings", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-gigapixel/enhancements/settings" } ] }, { "ID": "338", "Software": "Topaz Photo", "Operation": "Adjust Lighting — Contextual Correction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-948742f8a04f\ncollected-ef667a439c13", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Adjust Lighting v2 predicts a contextual illumination correction at full application.", "Previous record: rationale": "Learned appearance mapping supports the fixed-function-to-perceptual envelope 5–12. Exposure-like controls do not establish a conventional curve or an exact comparison position.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Topaz Photo v1.1.0 (22 December 2025) added Adjust Lighting v3, trained to recover detail from dark shadows and blown highlights; v2 remains selectable with Exposure, Highlight and Shadow controls and an optional color-correction toggle. Documentation describes per-pixel contextual changes to exposure and contrast and does not state whether the model outputs pixels or tone parameters; the record's v2 scope is unchanged. Not established: Whether v2 outputs pixels directly or predicts tone parameters, and its training objective. v3 recovers highlight and shadow detail and would need its own record. docs.topazlabs.com, v1.1.0, 2025-12-22 • docs.topazlabs.com, v3 section", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/adjust-lighting\nhttps://docs.topazlabs.com/topaz-photo/release-summary", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-948742f8a04f", "Operation IDs": "338", "Citation": "Topaz · Contextual lighting adjustment", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/adjust-lighting" }, { "Source ID": "collected-ef667a439c13", "Operation IDs": "338, 341, 347", "Citation": "docs.topazlabs.com, v1.1.0, 2025-12-22", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/release-summary" } ] }, { "ID": "339", "Software": "Topaz Photo", "Operation": "Autopilot Denoise and Sharpen Workflow", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-43ff30c2795f\ncollected-0fa936126b08\ncollected-3c1dd508cbbe\ncollected-de7a4211f886\ncollected-a7e562c03858\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Autopilot selects and applies Denoise and Sharpen on non-RAW imagery in the defined filter session.", "Previous record: rationale": "Learned restoration and automatic model selection remain in the result. The 5–8 reconstruction envelope accommodates the unspecified reconstruction objectives without averaging the stages.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Topaz Photo v1.5.0 (25 April 2026) added Denoise Max, described as the first generative denoising model, as a selectable model in the Denoise enhancement; a user post (26 June 2026) locates a Denoise Max section in Autopilot preferences. Topaz contrasts Denoise Max with the existing Normal, Strong and Extreme denoise models, and the developer model-selection page separates the standard Sharpen models from generative Super Focus. Not established: Documentation does not state whether Autopilot selects Denoise Max automatically for non-RAW files, or the training losses of the standard Denoise and Sharpen models. If the defined session uses Normal, Strong or Extreme with a standard Sharpen model, vendor documentation places all of them outside its generative models, which would support a position nearer REF08. docs.topazlabs.com, accessed 3 Oct 2026 • community.topazlabs.com, v1.5.0, 25 Apr 2026 • docs.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://www.topazlabs.com/topaz-photo\nhttps://docs.topazlabs.com/topaz-photo/functions/autopilot-and-configuration\nhttps://docs.topazlabs.com/topaz-photo/enhancements/denoise-raw-and-non-raw\nhttps://community.topazlabs.com/t/topaz-photo-v1-5-0-1-5-1-high-fidelity-v3-wonder-3-denoise-max-super-focus-v3/102274\nhttps://docs.topazlabs.com/topaz-photo/functions/save-session", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-43ff30c2795f", "Operation IDs": "339", "Citation": "Topaz · Photo AI Autopilot", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.topazlabs.com/topaz-photo" }, { "Source ID": "collected-0fa936126b08", "Operation IDs": "339", "Citation": "Topaz · Autopilot filter configuration", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/functions/autopilot-and-configuration" }, { "Source ID": "collected-3c1dd508cbbe", "Operation IDs": "339, 341", "Citation": "docs.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/denoise-raw-and-non-raw" }, { "Source ID": "collected-de7a4211f886", "Operation IDs": "339, 348", "Citation": "community.topazlabs.com, v1.5.0, 25 Apr 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-photo-v1-5-0-1-5-1-high-fidelity-v3-wonder-3-denoise-max-super-focus-v3/102274" }, { "Source ID": "collected-a7e562c03858", "Operation IDs": "339, 340, 343", "Citation": "docs.topazlabs.com — save-session", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/functions/save-session" } ] }, { "ID": "340", "Software": "Topaz Photo", "Operation": "Balance Color with Source Blending", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9e996f44e6f4\ncollected-a7e562c03858\ncollected-e8d8b29f3845\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Balance Color predicts a correction and blends it with the source at nonzero intermediate opacity.", "Previous record: rationale": "The correction retains learned appearance determination, supporting 5–12. Opacity is not t; the reference position concerns the mechanism's responsibility, not a pixel contribution percentage.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Topaz documentation describes Balance Color as a per-pixel, context-adaptive color adjustment whose output is called an AI-generated image, blended with the source by opacity. No model version, architecture or training objective is disclosed, and the Topaz Photo release summary lists no Balance Color model update. Not established: Whether the network predicts a color transform applied to source pixels or synthesizes output pixels; training objective and model version. docs.topazlabs.com, not shown", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/balance-color\nhttps://docs.topazlabs.com/topaz-photo/functions/save-session\nhttps://docs.topazlabs.com/topaz-photo/quick-start", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9e996f44e6f4", "Operation IDs": "340", "Citation": "Topaz · Balance Color and opacity", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/balance-color" }, { "Source ID": "collected-a7e562c03858", "Operation IDs": "339, 340, 343", "Citation": "docs.topazlabs.com — save-session", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/functions/save-session" }, { "Source ID": "collected-e8d8b29f3845", "Operation IDs": "340", "Citation": "docs.topazlabs.com — quick-start", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/quick-start" } ] }, { "ID": "341", "Software": "Topaz Photo", "Operation": "Denoise — Non-RAW Normal Model", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3c1dd508cbbe\ncollected-74c690872455\ncollected-ef667a439c13", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Non-RAW Normal Denoise predicts a noise-reduced image at the scoped filter stage.", "Previous record: rationale": "The fidelity-restoration task matches 8 provisionally. RAW demosaicing and other model variants are excluded, and training-loss equivalence with DnCNN is not claimed.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). Topaz Photo documentation (current release v1.7.0, 27 August 2026) describes the Non-RAW Normal model as removing low-to-medium noise while keeping a natural look, and labels Denoise MAX, added in v1.5.0 (25 April 2026), as Topaz's first generative denoising model. The scoped Normal model carries no generative label, and release notes from v1.0.0 to v1.7.0 list no change to it. Not established: Training data and loss for the Normal model are not published; its non-generative status rests on vendor labelling rather than a training statement. docs.topazlabs.com, accessed 2026-10-03 • docs.topazlabs.com, 2026-08-27", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/denoise-raw-and-non-raw\nhttps://docs.topazlabs.com/topaz-photo/enhancements/raw-denoise\nhttps://docs.topazlabs.com/topaz-photo/release-summary", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3c1dd508cbbe", "Operation IDs": "339, 341", "Citation": "docs.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/denoise-raw-and-non-raw" }, { "Source ID": "collected-74c690872455", "Operation IDs": "341", "Citation": "Topaz · RAW denoise and demosaicing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/raw-denoise" }, { "Source ID": "collected-ef667a439c13", "Operation IDs": "338, 341, 347", "Citation": "docs.topazlabs.com, v1.1.0, 2025-12-22", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/release-summary" } ] }, { "ID": "342", "Software": "Topaz Photo", "Operation": "Dust and Scratch Removal", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-760f0092c230\ncollected-f563b726b644\ncollected-ceb0ba4d7b5e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Dust and Scratch v2 reconstructs regions affected by image damage.", "Previous record: rationale": "The repair task supports the fidelity-to-perceptual envelope 5–8. It is not automatically unrestricted generation merely because damaged pixels are replaced.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Topaz Photo documentation states that the v1 Dust & Scratch model is no longer available and that v2 focuses on cleanup while better preserving original image detail; the same page records user feedback that v2 can remove image detail. The Topaz API serves Dust & Scratch 2 through its Restore Generative endpoint (restore-gen). Not established: Vendor text does not describe Dust & Scratch 2 as generative or state its training objective; the endpoint grouping alone does not establish whether the repair is trained for fidelity or for perceptual reconstruction. docs.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, Image API 1.2.0, accessed 3 Oct 202", "Collected source links": "https://docs.topazlabs.com/topaz-photo/dust-and-scratch\nhttps://developer.topazlabs.com/image-models/removal-and-cleanup/dust-and-scratch-2\nhttps://developer.topazlabs.com/reference/image/restore/restore-generative", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-760f0092c230", "Operation IDs": "342", "Citation": "Topaz · Dust and Scratch reconstruction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/dust-and-scratch" }, { "Source ID": "collected-f563b726b644", "Operation IDs": "342", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/removal-and-cleanup/dust-and-scratch-2" }, { "Source ID": "collected-ceb0ba4d7b5e", "Operation IDs": "342", "Citation": "developer.topazlabs.com, Image API 1.2.0, accessed 3 Oct 202", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/reference/image/restore/restore-generative" } ] }, { "ID": "343", "Software": "Topaz Photo", "Operation": "Preserve Text — Masked Restoration", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-6fe454210c6b\ncollected-e04d1051dc1c\ncollected-a7e562c03858\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Preserve Text restores manually masked lettering and applies the restoration over other filters.", "Previous record: rationale": "Learned image-detail reconstruction supports 5–8. The explicit mask retains the restoration's influence and does not add a conventional-component bonus; OCR and text authorship are excluded.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Topaz Photo documentation lists Low Resolution and Noisy/Compressed models for Preserve Text and states that the Strength slider controls the degree of generation used to preserve the text. The 6 July 2023 Topaz launch post describes the filter as making small or degraded text sharper and more natural; no later model update is documented. Not established: Training objective and loss for the Preserve Text models are undisclosed, and the vendor's Precision/Generative/Creative tier pages do not list Preserve Text. docs.topazlabs.com, accessed 3 Oct 2026 • community.topazlabs.com, 6 Jul 2023", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/preserve-text\nhttps://community.topazlabs.com/t/preserve-text-sharper-text-with-fewer-artifacts-july-2023/47969\nhttps://docs.topazlabs.com/topaz-photo/functions/save-session", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-6fe454210c6b", "Operation IDs": "343", "Citation": "Topaz · Preserve Text mask and models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/preserve-text" }, { "Source ID": "collected-e04d1051dc1c", "Operation IDs": "343", "Citation": "community.topazlabs.com, 6 Jul 2023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/preserve-text-sharper-text-with-fewer-artifacts-july-2023/47969" }, { "Source ID": "collected-a7e562c03858", "Operation IDs": "339, 340, 343", "Citation": "docs.topazlabs.com — save-session", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/functions/save-session" } ] }, { "ID": "344", "Software": "Topaz Photo", "Operation": "Recover Faces with Source Blending", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b65277044f86\ncollected-846960007c99\ncollected-86752655b5fb", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Recover Faces v3 reconstructs selected faces and blends the reconstruction with the source.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Topaz Photo v1.6.0 (13 May 2026) introduced Recover Faces v3, which the Topaz Photo documentation calls a generative recover faces model that processes each face separately; Gigapixel Face Recovery 3 shipped the same day with the same wording. Topaz developer documentation states that Face Recovery 3 reconstructs plausible features when facial information is missing. Not established: Architecture and loss (GAN, diffusion or other) are undisclosed. The developer API page lists two variants that trade fidelity against recovery strength; which variant Topaz Photo runs is not stated. Previous reason: The fidelity-to-perceptual envelope 5–8 follows the retained face-reconstruction task. This overlaps Gigapixel's recovery family; blending does not determine t. docs.topazlabs.com, accessed 3 Oct 2026 • community.topazlabs.com, v1.6.0, 13 May 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/recover-faces\nhttps://community.topazlabs.com/t/topaz-photo-v1-6-0-recover-faces-3-sharpen-noise-aware-neuroserver-model-speed-ups/102790\nhttps://developer.topazlabs.com/image-models/gigapixel/face-recovery-3", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b65277044f86", "Operation IDs": "344", "Citation": "Topaz · Recover Faces controls", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/recover-faces" }, { "Source ID": "collected-846960007c99", "Operation IDs": "344", "Citation": "community.topazlabs.com, v1.6.0, 13 May 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-photo-v1-6-0-recover-faces-3-sharpen-noise-aware-neuroserver-model-speed-ups/102790" }, { "Source ID": "collected-86752655b5fb", "Operation IDs": "336, 344", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/gigapixel/face-recovery-3" } ] }, { "ID": "345", "Software": "Topaz Photo", "Operation": "Recover v3 Upscale", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-03f827830add\ncollected-9c23f5e3639d\ncollected-8693ad023a6c\ncollected-ae347297ca23", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Recover v3 enlarges and reconstructs degraded imagery with optional pre-downscaling excluded.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Topaz Photo documentation describes Recover v3 (released in Topaz Photo v1.2.0) as reconstructing fine surface detail, gives it a Creativity Control for how aggressively texture is added, and its usage notes attribute its weakness on text and logos to diffusion models. Topaz developer documentation lists Recover 3 as a generative diffusion model in the Wonder family (Generative tier), and Gigapixel documentation lists Recover 3 under Generative Models. Not established: Training loss and the strength of source conditioning are undisclosed. Whether diffusion sampling places this model below REF05 is not established; the vendor assigns transformation with new creative detail to a separate Creative tier (Bloom). Previous reason: Restorative enlargement supports 5–8. The current host and model are retained explicitly rather than attributing the mechanism to the older Gigapixel route. docs.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/gigapixel-ai/filters-panel/generative-models\nhttps://docs.topazlabs.com/topaz-photo/enhancements/upscale-and-resize\nhttps://developer.topazlabs.com/image-models/wonder/recover-3\nhttps://developer.topazlabs.com/image-models/wonder", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-03f827830add", "Operation IDs": "345", "Citation": "Topaz · Gigapixel generative models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/gigapixel-ai/filters-panel/generative-models" }, { "Source ID": "collected-9c23f5e3639d", "Operation IDs": "345", "Citation": "Topaz Photo current upscale models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/upscale-and-resize" }, { "Source ID": "collected-8693ad023a6c", "Operation IDs": "345", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/wonder/recover-3" }, { "Source ID": "collected-ae347297ca23", "Operation IDs": "345, 349", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/wonder" } ] }, { "ID": "346", "Software": "Topaz Photo", "Operation": "Remove Object — Regional Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3e7797c7cffa\ncollected-62a479cdfd0d\ncollected-e4a469e65166\ncollected-edc10c98ed4a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Remove v2 synthesizes replacement content for a selected object region.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 1 model-decided, 4 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–5.0; no midpoint is used.", "Previous record: calculation": "K=0; M=1; unknown=4; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Remove v2, an in-house Topaz Labs generative removal model that runs locally (default from Topaz Photo v1.2.0, 2026-01-22); architecture and base model not documented. Model choice: No. Not established: Topaz does not state whether Remove v2 is trained from scratch or built on a public inpainting model. community.topazlabs.com, 2026-01-22 • docs.topazlabs.com, accessed 2026-10-04 • community.topazlabs.com, 2023-11-03\nDesigner side, 3 Oct 2026 Works inside a region of existing work. Decided by a trained model (Remove v2, local). Designer controls: Regular brush; Object Selection brush; Super-Pixel brush; brush size; add/remove mode; preview color. Returns count not documented; form not documented. “The new model is especially strong when working with large objects and complex backgrounds” Source • Screenshot", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/remove-object\nhttps://community.topazlabs.com/t/topaz-photo-v1-2-0/100375\nhttps://community.topazlabs.com/t/generative-remove-with-local-processing-november-2023/55465\nhttps://cdn.sanity.io/images/r2plryeu/production/84f3e2c3683169baa195275e07cd344b972284ab-273x449.png?q=90&fit=max&auto=format", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3e7797c7cffa", "Operation IDs": "346", "Citation": "Topaz · Remove Object and revision limits", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/remove-object" }, { "Source ID": "collected-62a479cdfd0d", "Operation IDs": "346", "Citation": "community.topazlabs.com, 2026-01-22", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-photo-v1-2-0/100375" }, { "Source ID": "collected-e4a469e65166", "Operation IDs": "346", "Citation": "community.topazlabs.com, 2023-11-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/generative-remove-with-local-processing-november-2023/55465" }, { "Source ID": "collected-edc10c98ed4a", "Operation IDs": "346", "Citation": "cdn.sanity.io — 84f3e2c3683169baa195275e07cd344b972284ab-273x449.png?q=90&fit=max&auto=format", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.sanity.io/images/r2plryeu/production/84f3e2c3683169baa195275e07cd344b972284ab-273x449.png?q=90&fit=max&auto=format" } ] }, { "ID": "347", "Software": "Topaz Photo", "Operation": "Sharpen — Standard Model", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ad5a96745570\ncollected-ef667a439c13", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Standard Sharpen estimates detail lost through blur, without a separate denoise pass.", "Previous record: rationale": "Its scoped restoration responsibility matches the fidelity-restoration reference at 8 provisionally. This is distinct from prescribed edge-contrast sharpening.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). Topaz Photo documentation lists Standard as the all-purpose sharpening model for slight lens and motion blur and directs missed-focus cases to the separate Super Focus model. Release notes through v1.7.0 (27 August 2026) add Wildlife, Portrait and Noise-Aware sharpen models and tie Super Focus v3 to the NeuroServer route for large local generative models; Standard carries no generative label and no listed change. Not established: Training data and loss for Standard are not published; the absence of a generative label is the only vendor signal on its trained task. docs.topazlabs.com, accessed 2026-10-03 • docs.topazlabs.com, v1.5.0, 2026-04-25", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/sharpen\nhttps://docs.topazlabs.com/topaz-photo/release-summary", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ad5a96745570", "Operation IDs": "347", "Citation": "Topaz · Sharpen models and selection", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/sharpen" }, { "Source ID": "collected-ef667a439c13", "Operation IDs": "338, 341, 347", "Citation": "docs.topazlabs.com, v1.1.0, 2025-12-22", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/release-summary" } ] }, { "ID": "348", "Software": "Topaz Photo", "Operation": "Super Focus with Post-Render Selection", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-99ba2e03c265\ncollected-de7a4211f886\ncollected-10a2958fd1b6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Super Focus v3 reconstructs blurred detail and permits subsequent selection and strength adjustment.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Topaz Photo v1.5.0 (25 April 2026) release notes describe Super Focus v3 as a new generative model able to fill in missing details, and Topaz Photo documentation calls Super Focus a generative AI tool trained on missed-focus cases. Topaz developer documentation lists Super Focus 3 as generative blur recovery served through the sharpen-gen endpoint; v1.7.0 (25 August 2026) added RAW support for v3. Not established: Architecture and training loss are undisclosed. Previous reason: Learned detail reconstruction supports 5–8. Post-render application controls retain the learned result and do not independently change the comparison coordinate. community.topazlabs.com, v1.5.0, 25 Apr 2026 • docs.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/super-focus\nhttps://community.topazlabs.com/t/topaz-photo-v1-5-0-1-5-1-high-fidelity-v3-wonder-3-denoise-max-super-focus-v3/102274\nhttps://developer.topazlabs.com/image-models/sharpen/super-focus-3", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-99ba2e03c265", "Operation IDs": "348", "Citation": "Topaz · Photo AI Super Focus", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/super-focus" }, { "Source ID": "collected-de7a4211f886", "Operation IDs": "339, 348", "Citation": "community.topazlabs.com, v1.5.0, 25 Apr 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-photo-v1-5-0-1-5-1-high-fidelity-v3-wonder-3-denoise-max-super-focus-v3/102274" }, { "Source ID": "collected-10a2958fd1b6", "Operation IDs": "348", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/sharpen/super-focus-3" } ] }, { "ID": "349", "Software": "Topaz Photo", "Operation": "Wonder v3.5 Image Repair", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b71ba59da653\ncollected-f6398acd5387\ncollected-4acd283b54f1\ncollected-ae347297ca23", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Photo Wonder v3.5 performs an integrated learned repair and enlargement pass.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Topaz Photo v1.7.0 (25 August 2026) introduced Wonder 3.5, which the release notes call the same generative upscaling model as Wonder 3 with improved detail and text clarity. Topaz developer documentation lists Wonder 3.5 in the Wonder family, whose tier is Generative: adding recovered detail and texture while preserving image intent. Not established: Architecture (diffusion or other) and training loss are undisclosed for Wonder 3.5. Previous reason: The fidelity-to-perceptual reconstruction envelope 5–8 is supported; the available objective evidence does not select a precise t. Multiple repairs within one model do not require an arithmetic sum. community.topazlabs.com, v1.7.0, 25 Aug 2026 • developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-photo/enhancements/wonder\nhttps://community.topazlabs.com/t/topaz-photo-v1-7-0-wonder-3-5/104557\nhttps://developer.topazlabs.com/image-models/wonder/wonder-3.5-new\nhttps://developer.topazlabs.com/image-models/wonder", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b71ba59da653", "Operation IDs": "349", "Citation": "Topaz · Wonder image repair", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-photo/enhancements/wonder" }, { "Source ID": "collected-f6398acd5387", "Operation IDs": "349", "Citation": "community.topazlabs.com, v1.7.0, 25 Aug 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-photo-v1-7-0-wonder-3-5/104557" }, { "Source ID": "collected-4acd283b54f1", "Operation IDs": "349", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/wonder/wonder-3.5-new" }, { "Source ID": "collected-ae347297ca23", "Operation IDs": "345, 349", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/image-models/wonder" } ] }, { "ID": "350", "Software": "Topaz Video", "Operation": "Deinterlace — Dione DV", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "7.3", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-36369543cf6d\ncollected-697d7781cfb3\ncollected-ed67d56ebdd4\ncollected-df022a157842", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Video Dione DV deinterlaces and enlarges footage with added noise disabled.", "Previous record: rationale": "Rule B (3 Oct 2026): vendor documentation states fidelity-oriented training or non-generative intent for an operation that creates missing samples, which follows the Lightroom Super Resolution comparison: t = 0.75 between 5 and 8, sensitivity t = 0.5–1 (6.5–8).", "Previous record: calculation": "X = 5 + 0.75 × (8 − 5) = 7.25. Sensitivity t = 0.5–1 gives 6.5–8.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 7.3. Topaz developer documentation places Dione (API model ddv-3 for Dione DV) in the Proteus family, whose tier is Precision with the stated intent to improve resolution while preserving source characteristics; the Generative tier goes to the Starlight family. Topaz Video documentation states that Dione DV is trained for DV, DVCPRO, DVCAM, 720i and 1080i footage. Not established: Training loss is undisclosed; the Precision label is a family-level statement of intent. Previous reason: Learned temporal and spatial reconstruction supports 5–8 by comparison with fidelity and perceptual reconstruction. Field order and output dimensions do not set t. developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-video/filters/enhancement\nhttps://developer.topazlabs.com/getting-started/model-selection\nhttps://developer.topazlabs.com/video-models/proteus\nhttps://developer.topazlabs.com/video-models/proteus/dione", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-36369543cf6d", "Operation IDs": "350, 353", "Citation": "Topaz · Video enhancement and deinterlacing models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/filters/enhancement" }, { "Source ID": "collected-697d7781cfb3", "Operation IDs": "350, 351, 353", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/getting-started/model-selection" }, { "Source ID": "collected-ed67d56ebdd4", "Operation IDs": "350, 353", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/proteus" }, { "Source ID": "collected-df022a157842", "Operation IDs": "350", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/proteus/dione" } ] }, { "ID": "351", "Software": "Topaz Video", "Operation": "Frame Interpolation — Apollo", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-67f3d368ec91\ncollected-27f7cacad7a4\ncollected-9058547d6401\ncollected-697d7781cfb3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Video Apollo estimates motion and constructs intermediate frames for the scoped twofold rate conversion.", "Previous record: rationale": "The learned temporal reconstruction supports a cross-domain 5–8 envelope. Duplicate replacement and slow motion are excluded; frame-count multiplication is not an inference weight.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Topaz Video documentation describes Apollo as specializing in non-linear motion and slightly blurry inputs and generating 8 interpolated frames per run; the developer API lists Apollo as the default Quality frame-interpolation model. Topaz's Precision/Generative/Creative tiers cover the upscaling families and do not classify the frame-interpolation models. Not established: Architecture and training objective for Apollo are undisclosed. docs.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-video/filters/frame-interpolation\nhttps://docs.topazlabs.com/topaz-video/reference-guide/importing-previewing-and-exporting\nhttps://developer.topazlabs.com/video-models/frame-interpolation/apollo\nhttps://developer.topazlabs.com/getting-started/model-selection", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-67f3d368ec91", "Operation IDs": "351", "Citation": "Topaz Video frame interpolation guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/filters/frame-interpolation" }, { "Source ID": "collected-27f7cacad7a4", "Operation IDs": "351, 356", "Citation": "Topaz Video importing previewing and exporting", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/reference-guide/importing-previewing-and-exporting" }, { "Source ID": "collected-9058547d6401", "Operation IDs": "351", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/frame-interpolation/apollo" }, { "Source ID": "collected-697d7781cfb3", "Operation IDs": "350, 351, 353", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/getting-started/model-selection" } ] }, { "ID": "352", "Software": "Topaz Video", "Operation": "Motion Deblur — Themis", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e7f19ae7f931\ncollected-fac3d28434fe", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Video Themis performs learned motion-deblur restoration with other filters disabled.", "Previous record: rationale": "The stated fidelity-restoration responsibility supports 8 provisionally. Stabilization and frame interpolation are different scoped tasks.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). Topaz Video documentation names Themis as the motion-deblur model that removes blur instead of raising edge contrast, for blur from panning, rotation or zoom; the model dates from Topaz Video AI v3.1 (January 2023). Neither the documentation nor the launch note states training data, loss or a generative component. Not established: Training data, loss, and whether Themis has been retrained since 2023 are not published. docs.topazlabs.com, accessed 2026-10-03 • community.topazlabs.com, 2023-01-18", "Collected source links": "https://docs.topazlabs.com/topaz-video/filters/motion-deblur\nhttps://community.topazlabs.com/t/topaz-video-ai-v3-1-faster-exports-faster-playback-and-new-motion-deblur-model/38926", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e7f19ae7f931", "Operation IDs": "352", "Citation": "Topaz · Themis motion deblur", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/filters/motion-deblur" }, { "Source ID": "collected-fac3d28434fe", "Operation IDs": "352", "Citation": "community.topazlabs.com, 2023-01-18", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-video-ai-v3-1-faster-exports-faster-playback-and-new-motion-deblur-model/38926" } ] }, { "ID": "353", "Software": "Topaz Video", "Operation": "Proteus Natural Progressive Video Enhancement", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "7.3", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-447ede523747\ncollected-36369543cf6d\ncollected-697d7781cfb3\ncollected-ed67d56ebdd4\ncollected-1dec76057e83", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Video Proteus Natural reconstructs progressive footage without separate source blending or other filters.", "Previous record: rationale": "Rule B (3 Oct 2026): vendor documentation states fidelity-oriented training or non-generative intent for an operation that creates missing samples, which follows the Lightroom Super Resolution comparison: t = 0.75 between 5 and 8, sensitivity t = 0.5–1 (6.5–8).", "Previous record: calculation": "X = 5 + 0.75 × (8 − 5) = 7.25. Sensitivity t = 0.5–1 gives 6.5–8.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 7.3. Topaz developer documentation places Proteus Natural in the Proteus family, whose tier is Precision ('Improve resolution, preserve source characteristics'), and describes it as lighter processing that stays closest to the original. Topaz announced Proteus Natural in its January 2026 Fidelity Update as a more natural, softer upgrade of Proteus. Not established: Training loss is undisclosed; the Precision label is a family-level statement of intent. Previous reason: Learned image-detail enhancement supports the fidelity-to-perceptual envelope 5–8. Public task documentation does not establish one training-objective position. developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://www.topazlabs.com/topaz-video-ai\nhttps://docs.topazlabs.com/topaz-video/filters/enhancement\nhttps://developer.topazlabs.com/getting-started/model-selection\nhttps://developer.topazlabs.com/video-models/proteus\nhttps://developer.topazlabs.com/video-models/proteus/proteus-natural", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-447ede523747", "Operation IDs": "353", "Citation": "Topaz · Video AI overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.topazlabs.com/topaz-video-ai" }, { "Source ID": "collected-36369543cf6d", "Operation IDs": "350, 353", "Citation": "Topaz · Video enhancement and deinterlacing models", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/filters/enhancement" }, { "Source ID": "collected-697d7781cfb3", "Operation IDs": "350, 351, 353", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/getting-started/model-selection" }, { "Source ID": "collected-ed67d56ebdd4", "Operation IDs": "350, 353", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/proteus" }, { "Source ID": "collected-1dec76057e83", "Operation IDs": "353", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/proteus/proteus-natural" } ] }, { "ID": "354", "Software": "Topaz Video", "Operation": "Hyperion 2.5 SDR to HDR", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3659c79fec1f\ncollected-7c5a5e6adcd6\ncollected-5b915e8c664e", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Video Hyperion 2.5 predicts an expanded luminance representation from SDR footage.", "Previous record: rationale": "Learned appearance mapping supports 5–12 between fixed-function transformation and perceptual reconstruction. The change is more than export metadata, but its model identity does not determine t.", "Previous record: calculation": "X = 5 + t × (12 − 5); t = 0–1; X = 5–12. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–12.0). Hyperion 2.5 is available in Topaz Video 1.7.1 and later and through the Topaz API; Topaz describes redistribution of luminance and color with fine detail preserved and the source kept intact, alongside unspecified built-in enhancements. No model architecture or training objective is disclosed. Not established: What the built-in enhancements do and their training objective; whether clipped regions receive synthesized detail. docs.topazlabs.com, Hyperion 2.5, release 1.7.1-current • developer.topazlabs.com, not shown", "Collected source links": "https://docs.topazlabs.com/topaz-video/filters/sdr-to-hdr/hyperion\nhttps://docs.topazlabs.com/topaz-video/filters/sdr-to-hdr/hyperion-2\nhttps://developer.topazlabs.com/video-models/video-utilities/hyperion-2.5-sdr-to-hdr-new", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3659c79fec1f", "Operation IDs": "354", "Citation": "Topaz · Hyperion SDR to HDR", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/filters/sdr-to-hdr/hyperion" }, { "Source ID": "collected-7c5a5e6adcd6", "Operation IDs": "354", "Citation": "Topaz current Hyperion 2.5", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/filters/sdr-to-hdr/hyperion-2" }, { "Source ID": "collected-5b915e8c664e", "Operation IDs": "354", "Citation": "developer.topazlabs.com, not shown", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/video-utilities/hyperion-2.5-sdr-to-hdr-new" } ] }, { "ID": "355", "Software": "Topaz Video", "Operation": "Stabilization with Full-Frame Border Synthesis", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-659987c0e89f\ncollected-e74ee5e0288f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Video Full Frame stabilization combines geometric stabilization with learned reconstruction of exposed borders.", "Previous record: rationale": "The border-reconstruction task supports 5–8 provisionally. Conventional geometric motion does not dilute the retained prediction; Auto Crop is outside this scoped branch.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). Topaz Video documentation states that Full Frame mode fills blank edge areas using frame interpolation and deep learning, and a Topaz Learn article describes the filled edges as AI-generated and possibly soft. Neither source names the fill model or its training objective. Not established: The border-fill model, its training objective, and how much fill content comes from neighbouring frames compared with learned synthesis are undisclosed. docs.topazlabs.com, accessed 3 Oct 2026 • topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-video/filters/stabilization\nhttps://www.topazlabs.com/learn/how-to-use-stabilization-motion-deblur-topaz-video", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-659987c0e89f", "Operation IDs": "355", "Citation": "Topaz · Stabilization geometry and generated borders", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/filters/stabilization" }, { "Source ID": "collected-e74ee5e0288f", "Operation IDs": "355", "Citation": "topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.topazlabs.com/learn/how-to-use-stabilization-motion-deblur-topaz-video" } ] }, { "ID": "356", "Software": "Topaz Video", "Operation": "Starlight Precise 2.6 Enhancement", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-114c05dec50c\ncollected-680281180621\ncollected-f47701a4c1bc\ncollected-7a7232c574ae\ncollected-27f7cacad7a4\ncollected-398a97ce3d8d\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Topaz Video Starlight Precise 2.6 reconstructs degraded video through its documented diffusion-based enhancement route.", "Previous record: rationale": "Rule A (3 Oct 2026): vendor documentation names a generative model that reconstructs detail in supplied media, which matches REF05 (ESRGAN perceptual reconstruction), X = 5. GAN or diffusion architecture adds no points under the comparison policy.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 5.0. Topaz developer documentation describes Starlight Precise 2.6 as a diffusion video enhancement model whose Sharpen parameter balances a softer look against sharper generative enhancement, and places the Starlight family in its Generative tier. Precise 2.6 replaced 2.5 from Topaz Video 1.7.0, and the September 2026 Topaz for Web release added a Speed Boost option. Not established: Training loss and the strength of source conditioning in the diffusion process are undisclosed. Previous reason: The restorative task supports 5–8. Diffusion architecture, processing cost and pixel-count increase do not automatically establish unrestricted generation or a numerical inference amount. developer.topazlabs.com, accessed 3 Oct 2026 • developer.topazlabs.com, accessed 3 Oct 2026 • docs.topazlabs.com, accessed 3 Oct 2026", "Collected source links": "https://docs.topazlabs.com/topaz-video/project-starlight-series/starlight-precise-25\nhttps://community.topazlabs.com/t/topaz-video-1-7-0/104328\nhttps://developer.topazlabs.com/video-models/starlight/starlight-precise-2.6\nhttps://developer.topazlabs.com/video-models/starlight\nhttps://docs.topazlabs.com/topaz-video/reference-guide/importing-previewing-and-exporting\nhttps://docs.topazlabs.com/topaz-video/advanced-functions/second-pass-enhancement", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-114c05dec50c", "Operation IDs": "155, 356", "Citation": "Topaz Starlight Precise 2.5 and 2.6 native documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/project-starlight-series/starlight-precise-25" }, { "Source ID": "collected-680281180621", "Operation IDs": "356", "Citation": "Topaz staff Video 1.7 release introducing Precise 2.6", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.topazlabs.com/t/topaz-video-1-7-0/104328" }, { "Source ID": "collected-f47701a4c1bc", "Operation IDs": "155, 356", "Citation": "developer.topazlabs.com, Starlight Precise 2.6", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/starlight/starlight-precise-2.6" }, { "Source ID": "collected-7a7232c574ae", "Operation IDs": "356", "Citation": "developer.topazlabs.com, accessed 3 Oct 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developer.topazlabs.com/video-models/starlight" }, { "Source ID": "collected-27f7cacad7a4", "Operation IDs": "351, 356", "Citation": "Topaz Video importing previewing and exporting", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/reference-guide/importing-previewing-and-exporting" }, { "Source ID": "collected-398a97ce3d8d", "Operation IDs": "356", "Citation": "docs.topazlabs.com — second-pass-enhancement", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.topazlabs.com/topaz-video/advanced-functions/second-pass-enhancement" } ] }, { "ID": "357", "Software": "TouchDesigner", "Operation": "MediaPipe Segmentation and Video Cutout", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-49244b2855f3\ncollected-6c20b2e634be\ncollected-9db1f5c52b39", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TouchDesigner uses MediaPipe image segmentation to key source video and composite a background.", "Previous record: rationale": "The retained learned region determination matches 10. Conventional compositing does not erase the predicted mask's role.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). The mediapipe-touchdesigner README lists Selfie, Selfie Multiclass, Selfie Landscape, Hair Segmenter and DeepLab V3 segmentation models that output confidence masks. Release v0.5.3 (10 August 2026) is tested on TouchDesigner 2025.33070 and adds wiring conveniences only. github.com, README, fetched 2026-10-03 • github.com, 2026-08-10 (v0.5.3)", "Collected source links": "https://github.com/torinmb/mediapipe-touchdesigner\nhttps://derivative.ca/UserGuide/Composite_TOP\nhttps://github.com/torinmb/mediapipe-touchdesigner/releases/tag/v0.5.3", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-49244b2855f3", "Operation IDs": "357", "Citation": "GitHub · mediapipe-touchdesigner", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/torinmb/mediapipe-touchdesigner" }, { "Source ID": "collected-6c20b2e634be", "Operation IDs": "357, 361", "Citation": "Derivative · Composite TOP image construction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://derivative.ca/UserGuide/Composite_TOP" }, { "Source ID": "collected-9db1f5c52b39", "Operation IDs": "357", "Citation": "github.com, 2026-08-10 (v0.5.3)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/torinmb/mediapipe-touchdesigner/releases/tag/v0.5.3" } ] }, { "ID": "358", "Software": "TouchDesigner", "Operation": "NVIDIA Background Segmentation and Composite", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-1fcee20d0c5c\ncollected-f816e9af4d05\ncollected-906b3e561dcb\ncollected-5cff2a12e181\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TouchDesigner uses NVIDIA Background TOP to predict a mask for a synchronized video composite.", "Previous record: rationale": "Learned region determination matches 10; the following composite adds no hybrid offset. Face tracking is not the mechanism assessed here.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Derivative's NVIDIA Background TOP page describes AI image segmentation from the Maxine Video Effects AI Green Screen module, outputting a monochrome person key, with Quality and Performance model modes. No relighting or generative feature is documented for this TOP. Not established: The Maxine model version bundled with current TouchDesigner builds is not stated. docs.derivative.ca, fetched 2026-10-03", "Collected source links": "https://docs.derivative.ca/Face_Track_CHOP\nhttps://docs.derivative.ca/Nvidia_Background_TOP\nhttps://docs.derivative.ca/.toe\nhttps://docs.derivative.ca/Cook", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-1fcee20d0c5c", "Operation IDs": "358", "Citation": "Derivative · NVIDIA Maxine Operators", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.derivative.ca/Face_Track_CHOP" }, { "Source ID": "collected-f816e9af4d05", "Operation IDs": "358", "Citation": "Derivative · NVIDIA Background TOP", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.derivative.ca/Nvidia_Background_TOP" }, { "Source ID": "collected-906b3e561dcb", "Operation IDs": "358", "Citation": "docs.derivative.ca — .toe", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.derivative.ca/.toe" }, { "Source ID": "collected-5cff2a12e181", "Operation IDs": "358", "Citation": "docs.derivative.ca — Cook", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.derivative.ca/Cook" } ] }, { "ID": "359", "Software": "TouchDesigner", "Operation": "NVIDIA Denoise TOP — Video Noise Removal", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-b3ee289788ac\ncollected-c4f3220704b3\ncollected-cdbe31876bc7", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TouchDesigner NVIDIA Denoise TOP predicts noise-reduced video at unchanged resolution in Denoising mode.", "Previous record: rationale": "The fidelity-restoration responsibility matches 8. Artifact Reduction and later compositing are excluded; source subject matter and image size supply no score components.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). The TouchDesigner wiki describes Denoising mode as an AI model tuned to remove low-light camera noise, and build 2025.30280 (30 June 2025) moved the TOP to a newer Maxine release with Blackwell GPU support. NVIDIA's Video Effects SDK guide (updated 22 September 2026) describes the webcam denoise filter as removing noise while preserving details, with a strength setting that trades texture preservation against noise removal. Not established: NVIDIA does not publish the denoise model's architecture, training data or loss. The wiki also names the AI Green Screening module as the engine component, which conflicts with the Denoising mode description and is unresolved. derivative.ca, accessed 2026-10-03 • docs.nvidia.com, Last updated 2026-09-22 • docs.derivative.ca, Build 2025.30280, 2025-06-30", "Collected source links": "https://derivative.ca/UserGuide/NVIDIA_Denoise_TOP\nhttps://docs.nvidia.com/maxine/vfx/latest/Filters/WebcamDenoiser.html\nhttps://docs.derivative.ca/Release_Notes", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-b3ee289788ac", "Operation IDs": "359", "Citation": "Derivative · NVIDIA Denoise TOP", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://derivative.ca/UserGuide/NVIDIA_Denoise_TOP" }, { "Source ID": "collected-c4f3220704b3", "Operation IDs": "359", "Citation": "NVIDIA Webcam Denoise filter guide, updated 09102026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.nvidia.com/maxine/vfx/latest/Filters/WebcamDenoiser.html" }, { "Source ID": "collected-cdbe31876bc7", "Operation IDs": "359", "Citation": "docs.derivative.ca, Build 2025.30280, 2025-06-30", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.derivative.ca/Release_Notes" } ] }, { "ID": "360", "Software": "TouchDesigner", "Operation": "NVIDIA Upscaler TOP — Super-Resolution", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0–8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-65b234d848f1\ncollected-e846d990d8a9\ncollected-b309e2db6022", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TouchDesigner NVIDIA Upscaler TOP Super Res 2 reconstructs higher-resolution frames with Artifact Reduction disabled.", "Previous record: rationale": "The learned reconstruction task supports 5–8. Available documentation does not establish a precise loss-function comparison; the separate Upscale filter is not silently substituted.", "Previous record: calculation": "X = 5 + t × (8 − 5); t = 0–1; X = 5–8. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0–8.0). NVIDIA's Maxine VFX SDK guide (Super Resolution page, updated 8 January 2026) describes the filter as improving resolution, enhancing detail and sharpening while preserving content, with one mode for lossy and one for lossless sources. The TouchDesigner wiki lists Super Res 2 with an optional Artifact Reduction toggle; its hardware note names the AI Green Screening module, an inconsistency in the wiki text. Not established: The training loss (fidelity or perceptual/adversarial) of the Maxine Super Resolution model is undisclosed. Searches for an NVIDIA model card were not executed because the session search budget ran out. docs.nvidia.com, last updated 8 Jan 2026 • derivative.ca, accessed 3 Oct 2026", "Collected source links": "https://derivative.ca/UserGuide/NVIDIA_Upscaler_TOP\nhttps://docs.nvidia.com/deeplearning/maxine/vfx-sdk-programming-guide/index.html\nhttps://docs.nvidia.com/maxine/vfx/latest/Filters/SuperResolution.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-65b234d848f1", "Operation IDs": "360", "Citation": "Derivative · NVIDIA Upscaler TOP modes", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://derivative.ca/UserGuide/NVIDIA_Upscaler_TOP" }, { "Source ID": "collected-e846d990d8a9", "Operation IDs": "360", "Citation": "NVIDIA · Video Effects SDK filters", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.nvidia.com/deeplearning/maxine/vfx-sdk-programming-guide/index.html" }, { "Source ID": "collected-b309e2db6022", "Operation IDs": "360", "Citation": "docs.nvidia.com, last updated 8 Jan 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.nvidia.com/maxine/vfx/latest/Filters/SuperResolution.html" } ] }, { "ID": "361", "Software": "TouchDesigner", "Operation": "Transform and Composite Two Images", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-da60924c8ceb\ncollected-014cd02daa53\ncollected-6c20b2e634be", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TouchDesigner transforms one supplied image and composites it over another with the specified TOP nodes.", "Previous record: rationale": "Prescribed geometry and compositing arithmetic contain no in-scope trained prediction. External image generation is outside this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: nodes/parameters; Transform TOP (Translate, Rotate, Scale, Pivot, Extend); Composite TOP (Operation, Pre-Fit Overlay, Justify, Translate, Scale). Returns one result; editable structure: nodes. “The Composite TOP is a multi-input TOP that will perform a composite operation for each input.” Source", "Collected source links": "https://docs.derivative.ca/Network_Editor\nhttps://derivative.ca/UserGuide/Transform_TOP\nhttps://derivative.ca/UserGuide/Composite_TOP", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-da60924c8ceb", "Operation IDs": "361", "Citation": "Derivative · Network Editor", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.derivative.ca/Network_Editor" }, { "Source ID": "collected-014cd02daa53", "Operation IDs": "361", "Citation": "Derivative · Transform TOP", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://derivative.ca/UserGuide/Transform_TOP" }, { "Source ID": "collected-6c20b2e634be", "Operation IDs": "357, 361", "Citation": "Derivative · Composite TOP image construction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://derivative.ca/UserGuide/Composite_TOP" } ] }, { "ID": "362", "Software": "Tripo AI", "Operation": "H3.1 Image to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-ebab39537e56\ncollected-5284a77a01a3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Tripo H3.1 predicts mesh geometry from an image before retopology, rigging or segmentation.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=2; conditional=1; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Tripo H3.1 image to 3D). Designer controls: image; model version; model seed; texture toggle; texture seed; texture alignment; texture quality; PBR toggle; geometry quality; face limit; quad; smart low poly; generate parts; orientation; auto size; UV unwrap; compression. Returns count not documented; editable structure: mesh with PBR textures; quad mesh as FBX. “delivers top-tier geometry with PBR textures, with face counts reaching up to 2 million” Source", "Collected source links": "https://www.tripo3d.ai/blog/introducing-tripo-new-algorithm3\nhttps://developers.tripo3d.ai/en/models/v3-1", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-ebab39537e56", "Operation IDs": "362", "Citation": "Tripo · v3.0 model generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.tripo3d.ai/blog/introducing-tripo-new-algorithm3" }, { "Source ID": "collected-5284a77a01a3", "Operation IDs": "362, 363", "Citation": "Tripo H3.1 model", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developers.tripo3d.ai/en/models/v3-1" } ] }, { "ID": "363", "Software": "Tripo AI", "Operation": "H3.1 Text to Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4f84a0f15afe\ncollected-5284a77a01a3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Tripo H3.1 predicts a mesh from text with optional part generation disabled.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 4 model-decided, 0 unknown and 0 setting-dependent categories among 4 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=4; unknown=0; conditional=0; A=4. Lower: 2 + 3 × 0 ÷ (4 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Tripo H3.1 text to 3D). Designer controls: prompt (up to 1024 characters); negative prompt; model version; image seed; model seed; texture toggle; texture seed; texture quality; PBR toggle; geometry quality; face limit; quad; smart low poly; generate parts; auto size; UV unwrap. Returns count not documented; editable structure: mesh with PBR textures; quad mesh as FBX. “The maximum prompt length is 1024 characters, equivalent to approximately 100 words.” Source", "Collected source links": "https://docs.tripo3d.ai/model-generation/text-to-model-v3-0-v3-1.html\nhttps://developers.tripo3d.ai/en/models/v3-1", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4f84a0f15afe", "Operation IDs": "363", "Citation": "Tripo · Text to model and optional segmented parts", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.tripo3d.ai/model-generation/text-to-model-v3-0-v3-1.html" }, { "Source ID": "collected-5284a77a01a3", "Operation IDs": "362, 363", "Citation": "Tripo H3.1 model", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://developers.tripo3d.ai/en/models/v3-1" } ] }, { "ID": "364", "Software": "TripoSR (open-source code)", "Operation": "Single Image to Mesh Reconstruction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c6a7b1c0a8da\ncollected-f4e7d302c752", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "TripoSR predicts a 3D representation from one image and decodes a mesh.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=3; conditional=0; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (feed-forward image to 3D). Designer controls: image path; --bake-texture; --texture-resolution; --output-dir. Returns one result (per input image); editable structure: mesh with vertex colors or baked texture. “generating high-quality 3D models in less than 0.5 seconds on an NVIDIA A100 GPU” Source", "Collected source links": "https://stability.ai/news-updates/triposr-3d-generation\nhttps://github.com/VAST-AI-Research/TripoSR", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c6a7b1c0a8da", "Operation IDs": "364", "Citation": "Stability AI · Single-Image Reconstruction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://stability.ai/news-updates/triposr-3d-generation" }, { "Source ID": "collected-f4e7d302c752", "Operation IDs": "364", "Citation": "VAST and Stability AI · TripoSR single image reconstruction", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/VAST-AI-Research/TripoSR" } ] }, { "ID": "365", "Software": "Unity", "Operation": "AI Sprite Generation and Inpainting", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-4a7ef8e17cd0\ncollected-4f70e20aad57\ncollected-c7a1618768a1\ncollected-36578d9091d1\ncollected-e99ec8d644f8", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Unity's Sprite Generator synthesizes or inpaints sprite imagery through the selected learned image model.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Third-party partner models: Layer models built on FLUX (Black Forest Labs) with pre-trained LoRAs, and Scenario models built on FLUX, GPT-Image (OpenAI), Google Gemini and Seedream (ByteDance) with LoRAs. At the Unity 6.2 launch the sprite models were LoRAs on Stable Diffusion and Flux. Model choice: Yes. Not established: Specific model versions (FLUX.1 vs FLUX.2, GPT-Image 1 vs later, which Gemini image model, which Seedream version) are not listed. unity.com • unity.com, 2026-04-28 • cgchannel.com, 2025-08-19\nDesigner side, 3 Oct 2026 Makes new output. Decided by a trained model (selected third-party image model; generate or inpaint). Designer controls: prompt; negative prompt; model choice; reference images (image, style, composition, depth, line art, feature); number of images; dimensions; custom seed; inpaint mask brush; remove background; upscale; pixelate; recolor. Returns several to choose from (Images setting; range not stated); a flat image: PNG sprites in /GeneratedAssets. “lets you refine a specific area of the sprite by painting a mask over it and generating new content for that region only” Source • Screenshot", "Collected source links": "https://unity.com/blog/unity-ai-sprite-generator\nhttps://docs.unity3d.com/Packages/com.unity.ai.assistant@2.19/manual/Sprite/visual-assets.html\nhttps://unity.com/legal/third-party-ai-terms\nhttps://www.cgchannel.com/2025/08/unity-rolls-out-unity-ai-in-unity-6-2/\nhttps://cdn.sanity.io/images/fuvbjjlp/production/a7b83de36adf828b48c77648cb814438dc163b3c-1922x1054.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-4a7ef8e17cd0", "Operation IDs": "365", "Citation": "Unity · Sprite inpainting and alternative assets", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://unity.com/blog/unity-ai-sprite-generator" }, { "Source ID": "collected-4f70e20aad57", "Operation IDs": "365", "Citation": "Unity · Visual generators and output types", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity3d.com/Packages/com.unity.ai.assistant@2.19/manual/Sprite/visual-assets.html" }, { "Source ID": "collected-c7a1618768a1", "Operation IDs": "365", "Citation": "unity.com — third-party-ai-terms", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://unity.com/legal/third-party-ai-terms" }, { "Source ID": "collected-36578d9091d1", "Operation IDs": "365", "Citation": "cgchannel.com, 2025-08-19", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2025/08/unity-rolls-out-unity-ai-in-unity-6-2/" }, { "Source ID": "collected-e99ec8d644f8", "Operation IDs": "365", "Citation": "cdn.sanity.io — a7b83de36adf828b48c77648cb814438dc163b3c-1922x1054.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://cdn.sanity.io/images/fuvbjjlp/production/a7b83de36adf828b48c77648cb814438dc163b3c-1922x1054.png" } ] }, { "ID": "366", "Software": "Unity", "Operation": "Assistant Scene Object Placement", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "6.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-41cb1c2613df\ncollected-3946e26950c3\ncollected-70d89b997397\ncollected-c452a910624d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Unity Assistant selects and places scene objects through its documented Run action.", "Previous record: rationale": "Rule D (3 Oct 2026): the REF02–REF10 envelope for language-model construction narrows by the documented role of the model; t = 0.5–1 (6–10) where the model selects or places existing elements, t = 0–0.5 (2–6) where it creates new content. For the scoped placement of scene objects, the learned contribution selects existing objects or prefabs and sets positions, rotations and scales through tool calls; object content comes from existing assets. That places the comparison nearer region determination than synthesis, at 6–10 within the same envelope.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0.5–1; X = 6–10. No midpoint assigned.", "Previous record: notes": "Changed 3 Oct 2026 X 2.0–10.0 → 6.0–10.0. Unity's Assistant package 2.20.0-pre.1 (23 September 2026) documents Ask, Plan and Agent modes; the Run mode documented for 1.0.0-pre.12 on Unity 6.3 no longer appears, and scene changes go through Agent-mode calls to predefined Unity tools such as Create GameObject. The documentation gives scattering attached tree prefabs at varied sizes and placements as an example. Not established: Which model tier serves a given request; whether a placement request can also trigger asset generation in the same turn. Previous reason: Learned structural choices remain in the scene. The synthesis-to-region envelope 2–10 is a cross-domain comparison; informational answers and file organization are excluded. docs.unity3d.com, 2.20.0-pre.1, 2026-09-23 • docs.unity3d.com, 2.20.0-pre.1, 2026-09-23 • docs.unity3d.com, 2.20.0-pre.1, 2026-09-23", "Collected source links": "https://docs.unity.cn/6000.3/Documentation/Manual/com.unity.ai.assistant.html\nhttps://docs.unity3d.com/Packages/com.unity.ai.assistant@2.20/manual/about/assistant-modes.html\nhttps://docs.unity3d.com/Packages/com.unity.ai.assistant@2.20/manual/assets/assistant-tools.html\nhttps://docs.unity3d.com/Packages/com.unity.ai.assistant@2.20/manual/assets/unity-object-query.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-41cb1c2613df", "Operation IDs": "366", "Citation": "Unity · Assistant scene actions", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity.cn/6000.3/Documentation/Manual/com.unity.ai.assistant.html" }, { "Source ID": "collected-3946e26950c3", "Operation IDs": "366", "Citation": "docs.unity3d.com, 2.20.0-pre.1, 2026-09-23", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity3d.com/Packages/com.unity.ai.assistant@2.20/manual/about/assistant-modes.html" }, { "Source ID": "collected-70d89b997397", "Operation IDs": "366", "Citation": "docs.unity3d.com, 2.20.0-pre.1, 2026-09-23", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity3d.com/Packages/com.unity.ai.assistant@2.20/manual/assets/assistant-tools.html" }, { "Source ID": "collected-c452a910624d", "Operation IDs": "366", "Citation": "docs.unity3d.com, 2.20.0-pre.1, 2026-09-23", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity3d.com/Packages/com.unity.ai.assistant@2.20/manual/assets/unity-object-query.html" } ] }, { "ID": "367", "Software": "Unity", "Operation": "GameObject Hierarchy and Component Editing", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-c265c3b9526a\ncollected-ffaa99c92d0e\ncollected-cc817ddaeddf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Unity evaluates manually edited GameObjects, components and hierarchy relationships.", "Previous record: rationale": "The scoped native editing is conventional. Runtime learned behavior and prior asset generation are separate operations.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: Hierarchy window (parenting); components; Inspector. Returns one result; editable structure: GameObject hierarchy with components. “Child GameObjects inherit the movement and rotation of the parent GameObject.” Source • Screenshot", "Collected source links": "https://docs.unity3d.com/6000.2/Documentation/Manual/unity-editor.html\nhttps://docs.unity3d.com/6000.2/Documentation/Manual/Hierarchy.html\nhttps://docs.unity3d.com/6000.2/Documentation/uploads/Main/hierarchy-window.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-c265c3b9526a", "Operation IDs": "367", "Citation": "Unity · Unity overview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity3d.com/6000.2/Documentation/Manual/unity-editor.html" }, { "Source ID": "collected-ffaa99c92d0e", "Operation IDs": "367", "Citation": "docs.unity3d.com — Hierarchy.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity3d.com/6000.2/Documentation/Manual/Hierarchy.html" }, { "Source ID": "collected-cc817ddaeddf", "Operation IDs": "367", "Citation": "docs.unity3d.com — hierarchy-window.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.unity3d.com/6000.2/Documentation/uploads/Main/hierarchy-window.png" } ] }, { "ID": "368", "Software": "Unity Editor", "Operation": "MCP LLM Scene Hierarchy Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-7a6a36314d76\ncollected-4e01e8b8b328\ncollected-cbb502961bef", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "The CoplayDev Unity MCP integration uses an LLM to construct GameObjects and their hierarchy.", "Previous record: rationale": "Learned scene-structure choices survive conventional API execution. The 2–10 envelope records that responsibility without treating the MCP transport as a model.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). MCP for Unity v10.2.0 (1 September 2026) exposes manage_gameobject with create, modify, duplicate and relative-move actions taking primitive type, parent, position, rotation, scale and components. The LLM can compose new forms from primitives as well as place existing prefabs within one hierarchy; separate tools create C# scripts and call Tripo or Meshy for 3D generation, outside this scope. Not established: Documentation does not separate composition of new forms from placement of existing assets for a given session, so the 2–10 envelope stays. github.com, v10.2.0, 2026-09-01 • github.com, main branch, 2026-10-03 • github.com, main branch, 2026-10-03", "Collected source links": "https://github.com/CoplayDev/unity-mcp\nhttps://github.com/CoplayDev/unity-mcp/blob/main/website/docs/reference/tools/core/manage_gameobject.md\nhttps://github.com/CoplayDev/unity-mcp/blob/main/website/docs/reference/tools/asset_gen/generate_model.md", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-7a6a36314d76", "Operation IDs": "368", "Citation": "GitHub · unity-mcp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/CoplayDev/unity-mcp" }, { "Source ID": "collected-4e01e8b8b328", "Operation IDs": "368", "Citation": "github.com, main branch, 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/CoplayDev/unity-mcp/blob/main/website/docs/reference/tools/core/manage_gameobject.md" }, { "Source ID": "collected-cbb502961bef", "Operation IDs": "368", "Citation": "github.com, main branch, 2026-10-03", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/CoplayDev/unity-mcp/blob/main/website/docs/reference/tools/asset_gen/generate_model.md" } ] }, { "ID": "369", "Software": "Unreal Engine", "Operation": "Blueprint Scene Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e872383f3851\ncollected-78c31f6aff2e\ncollected-ce9b355fbd4b\ncollected-12192d8312cb\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Unreal Engine evaluates explicitly authored Blueprint scene construction and actor-property edits.", "Previous record: rationale": "The scoped graph and property operations are conventional. Runtime neural inference and performance capture are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: nodes/parameters (Blueprint graph); actor-property edits. Returns one result; editable structure: nodes; actors. “The Blueprint Visual Scripting system in Unreal Engine is a complete gameplay scripting system based on the concept of using a node-based interface” Source", "Collected source links": "https://dev.epicgames.com/documentation/en-us/unreal-engine/blueprints-visual-scripting-in-unreal-engine\nhttps://dev.epicgames.com/documentation/unreal-engine/playing-and-simulating-in-unreal-engine\nhttps://dev.epicgames.com/documentation/en-us/unreal-engine/levels-in-unreal-engine\nhttps://dev.epicgames.com/documentation/en-us/unreal-engine/overview-of-blueprints-visual-scripting-in-unreal-engine", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e872383f3851", "Operation IDs": "369", "Citation": "Epic Games · Unreal Blueprint / Level Editing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/blueprints-visual-scripting-in-unreal-engine" }, { "Source ID": "collected-78c31f6aff2e", "Operation IDs": "369", "Citation": "Epic Games · Editor simulation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/unreal-engine/playing-and-simulating-in-unreal-engine" }, { "Source ID": "collected-ce9b355fbd4b", "Operation IDs": "369", "Citation": "dev.epicgames.com — levels-in-unreal-engine", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/levels-in-unreal-engine" }, { "Source ID": "collected-12192d8312cb", "Operation IDs": "369", "Citation": "dev.epicgames.com — overview-of-blueprints-visual-scripting-in-unreal-engine", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/overview-of-blueprints-visual-scripting-in-unreal-engine" } ] }, { "ID": "370", "Software": "Unreal Engine", "Operation": "Neural Morph Correction with Skinning", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "12.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-bdf2ba160c12\ncollected-b72cfb59950e\ncollected-7836bfd289cb\ncollected-4017c66e4be6\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Unreal Neural Morph predicts corrective deformation combined with conventional skinning.", "Previous record: rationale": "The trained approximation of a deformation mapping supports the fixed-function emulation reference at 12. Skinning retains the predicted correction and does not establish a conventional-only output.", "Previous record: calculation": "X = 12 + 0 × (12 − 12) = 12.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (12.0). The Unreal Engine 5.8 ML Deformer documentation describes the Neural Morph Model as producing morph target weights that correct a linearly skinned mesh, trained from an animation sequence and a geometry cache. No change to this mechanism is documented. dev.epicgames.com, UE 5.8 docs", "Collected source links": "https://dev.epicgames.com/documentation/unreal-engine/ml-deformer-framework-in-unreal-engine\nhttps://dev.epicgames.com/documentation/en-us/unreal-engine/ml-deformer-framework-in-unreal-engine\nhttps://dev.epicgames.com/documentation/en-us/unreal-engine/working-with-assets-in-unreal-engine\nhttps://dev.epicgames.com/documentation/en-us/unreal-engine/API/Plugins/MLDeformerFramework/UMLDeformerComponent", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-bdf2ba160c12", "Operation IDs": "370", "Citation": "Epic Games · Unreal ML Deformer", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/unreal-engine/ml-deformer-framework-in-unreal-engine" }, { "Source ID": "collected-b72cfb59950e", "Operation IDs": "370", "Citation": "Epic · Neural Morph and skinning framework", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/ml-deformer-framework-in-unreal-engine" }, { "Source ID": "collected-7836bfd289cb", "Operation IDs": "370", "Citation": "dev.epicgames.com — working-with-assets-in-unreal-engine", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/working-with-assets-in-unreal-engine" }, { "Source ID": "collected-4017c66e4be6", "Operation IDs": "370", "Citation": "dev.epicgames.com — UMLDeformerComponent", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/API/Plugins/MLDeformerFramework/UMLDeformerComponent" } ] }, { "ID": "371", "Software": "Unreal Engine", "Operation": "NNE Intel Open Image Denoise Spatial Pass", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "8.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-49cc605f566a\ncollected-73d79e85d28b\ncollected-be5f35fa2f0a\ncollected-d08801f2754a\ncollected-d0dcbc90916d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Unreal's specified NNE route runs Intel Open Image Denoise as a spatial image pass.", "Previous record: rationale": "The learned fidelity-restoration task matches 8. NNE is an execution framework, and neither its name nor other possible models are part of this assignment.", "Previous record: calculation": "X = 5 + 1 × (8 − 5) = 8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (8.0). The Unreal Engine 5.8 NNE Denoiser page states that the plugin ships several versions of Intel's Open Image Denoiser (fast, balanced and high quality, each with and without alpha) and accepts custom networks; UE 5.5 moved these to OIDN 2.3.0 models, and temporal denoising uses the separate NFOR denoiser. OIDN 3, announced for the second half of 2026 with kernel prediction, is unreleased as of 3 October 2026. Not established: The OIDN model generation shipped with UE 5.8 is not stated on the page. Adoption of OIDN 3 in NNE is not documented. dev.epicgames.com, Unreal Engine 5.8 documentation • dev.epicgames.com, UE 5.5 • cgchannel.com, 2026-01-30", "Collected source links": "https://dev.epicgames.com/documentation/en-us/unreal-engine/neural-network-engine-overview-in-unreal-engine\nhttps://dev.epicgames.com/documentation/unreal-engine/path-tracer-in-unreal-engine\nhttps://dev.epicgames.com/documentation/en-us/unreal-engine/nne-denoiser-in-unreal-engine\nhttps://dev.epicgames.com/documentation/unreal-engine/unreal-engine-5-5-release-notes?lang=en-US\nhttps://www.cgchannel.com/2026/01/open-image-denoise-3-will-support-temporal-denoising/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-49cc605f566a", "Operation IDs": "371", "Citation": "Epic Games · Unreal Neural Network Engine (NNE)", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/neural-network-engine-overview-in-unreal-engine" }, { "Source ID": "collected-73d79e85d28b", "Operation IDs": "371", "Citation": "Unreal · Path tracer NNE denoiser", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/unreal-engine/path-tracer-in-unreal-engine" }, { "Source ID": "collected-be5f35fa2f0a", "Operation IDs": "371", "Citation": "dev.epicgames.com, Unreal Engine 5.8 documentation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/nne-denoiser-in-unreal-engine" }, { "Source ID": "collected-d08801f2754a", "Operation IDs": "371", "Citation": "dev.epicgames.com, UE 5.5", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/unreal-engine/unreal-engine-5-5-release-notes?lang=en-US" }, { "Source ID": "collected-d0dcbc90916d", "Operation IDs": "371", "Citation": "cgchannel.com, 2026-01-30", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2026/01/open-image-denoise-3-will-support-temporal-denoising/" } ] }, { "ID": "372", "Software": "Unreal Engine", "Operation": "Sequencer 3D Transform Interpolation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-46e44fe6ee2e\ncollected-59df3e45fdb3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Unreal Sequencer interpolates specified object and camera transform keys.", "Previous record: rationale": "Prescribed temporal interpolation contains no trained predictor in this scope. Intermediate poses are computed without learned reconstruction.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: transform track; location, rotation and scale keys; custom tangents; curves; trajectory display. Returns one result; editable structure: timeline tracks with keys. “Float keyframes interpolate and can use custom tangents and curves.” Source • Screenshot", "Collected source links": "https://dev.epicgames.com/documentation/en-us/unreal-engine/transform-and-property-tracks?application_version=4.27\nhttps://d1iv7db44yhgxn.cloudfront.net/documentation/images/a99f0b18-4b37-4453-b472-2274e36af613/transformtrack.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-46e44fe6ee2e", "Operation IDs": "372", "Citation": "Epic · Transform channels and quaternion interpolation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/unreal-engine/transform-and-property-tracks?application_version=4.27" }, { "Source ID": "collected-59df3e45fdb3", "Operation IDs": "372", "Citation": "d1iv7db44yhgxn.cloudfront.net — transformtrack.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://d1iv7db44yhgxn.cloudfront.net/documentation/images/a99f0b18-4b37-4453-b472-2274e36af613/transformtrack.png" } ] }, { "ID": "373", "Software": "Unreal Engine MetaHuman", "Operation": "Current Animator Video and Depth Facial Solve", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-cfbdf95f71e7\ncollected-7de2cf9671e4\ncollected-1a19b010067c\ncollected-bb1c76648399\ncollected-a9c3a8164717", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "MetaHuman Animator solves facial animation through the current video-and-depth workflow.", "Previous record: rationale": "The saved evidence establishes learned dependencies but does not separate their retained roles from performer fitting and geometric solving sufficiently to select comparison references. Audio-only routes are excluded.", "Previous record: calculation": "", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (Unresolved). Epic's depth-data workflow page requires Unreal Engine 5.6 or later and lists depth generation, MetaHuman Identity creation and animation solve without describing their components. Epic staff stated in April 2023 that machine learning is the core of MetaHuman Animator, without assigning roles; other Epic documentation pages did not render outside a browser and the Unreal Engine blog refused access. Not established: Which solve stages (stereo depth generation, identity fitting, per-frame animation solve) use trained models, and what share of the animation curves they determine. dev.epicgames.com, current docs • 80.lv, 2023-04-06 • cgchannel.com, 2023-06", "Collected source links": "https://dev.epicgames.com/documentation/en-us/metahuman/metahuman-animator\nhttps://www.metahuman.com/news/delivering-high-quality-facial-animation-in-minutes-metahuman-animator-is-now-available\nhttps://dev.epicgames.com/documentation/en-us/metahuman/animation-from-depth-data\nhttps://80.lv/articles/the-epic-games-team-on-lumen-metahuman-ue5-2-demo-ai\nhttps://www.cgchannel.com/2023/06/epic-games-releases-metahuman-animator/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-cfbdf95f71e7", "Operation IDs": "373", "Citation": "Epic Games · Unreal MetaHuman Animator", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/metahuman/metahuman-animator" }, { "Source ID": "collected-7de2cf9671e4", "Operation IDs": "373", "Citation": "Epic · MetaHuman facial performance solving", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.metahuman.com/news/delivering-high-quality-facial-animation-in-minutes-metahuman-animator-is-now-available" }, { "Source ID": "collected-1a19b010067c", "Operation IDs": "373", "Citation": "dev.epicgames.com, current docs", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://dev.epicgames.com/documentation/en-us/metahuman/animation-from-depth-data" }, { "Source ID": "collected-bb1c76648399", "Operation IDs": "373", "Citation": "80.lv, 2023-04-06", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://80.lv/articles/the-epic-games-team-on-lumen-metahuman-ue5-2-demo-ai" }, { "Source ID": "collected-a9c3a8164717", "Operation IDs": "373", "Citation": "cgchannel.com, 2023-06", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.cgchannel.com/2023/06/epic-games-releases-metahuman-animator/" } ] }, { "ID": "374", "Software": "Unreal Engine", "Operation": "MCP LLM Blueprint Graph Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–10.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-5f02b686450b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "The experimental Unreal MCP integration uses an LLM to construct Blueprint nodes and connections.", "Previous record: rationale": "Learned graph choices support a synthesis-to-region envelope 2–10. This is a documentation-based proposed configuration; current host compatibility has not been executed or verified.", "Previous record: calculation": "X = 2 + t × (10 − 2); t = 0–1; X = 2–10. No midpoint assigned.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (2.0–10.0). chongdashu/unreal-mcp has no commits after 22 April 2025 and remains marked experimental for Unreal Engine 5.5+. Its Blueprint graph tools add event nodes, add and connect function-call nodes, and add typed variables, so the LLM composes Blueprint logic from engine node types. Not established: Compatibility with current Unreal Engine releases is unverified. Documentation does not settle a position within 2–10 for Blueprint logic composed from engine nodes. github.com, last commit 2025-04-22", "Collected source links": "https://github.com/chongdashu/unreal-mcp", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-5f02b686450b", "Operation IDs": "374", "Citation": "GitHub · unreal-mcp", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/chongdashu/unreal-mcp" } ] }, { "ID": "375", "Software": "Upscayl", "Operation": "RealESRGAN AnimeVideo v3 still-image upscale", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8b93800629a0\ncollected-0cce1c946983\ncollected-e215fa9a6670\ncollected-565850c3a37e\ncollected-b973909d0fa0\ncollected-523bf372b146\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Upscayl runs the bundled RealESRGAN AnimeVideo v3 model for native fourfold still-image enlargement.", "Previous record: rationale": "The model's perceptual reconstruction supports the ESRGAN-family reference at 5. The video-derived model name does not make the scoped output a video, and this host overlaps the same mechanism family.", "Previous record: calculation": "X = 5 + 0 × (8 − 5) = 5.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (5.0). Upscayl v2.15.0 (25 December 2024, latest release; repository pushed September 2026) packages seven models from resources/models (upscayl-standard-4x, upscayl-lite-4x, high-fidelity-4x, remacri-4x, ultramix-balanced-4x, ultrasharp-4x, digital-art-4x). The realesr-animevideov3 files sit in the top-level models folder (last changed 12 April 2023), which the build configuration does not package, so running AnimeVideo v3 in the current app requires loading it as a custom model. The model's mechanism is unchanged. Not established: Whether any released Upscayl version packaged AnimeVideo v3 (tags v2.5.5, v2.9.0, v2.11.5 and v2.15.0 do not include it in resources/models); the record's word bundled may need rescoping to custom-model use. github.com, v2.15.0 • github.com, v2.15.0 • github.com, main, 2026", "Collected source links": "https://github.com/upscayl/upscayl\nhttps://github.com/upscayl/upscayl/tree/main/models\nhttps://github.com/upscayl/upscayl/blob/v2.15.0/package.json\nhttps://github.com/upscayl/upscayl/blob/v2.15.0/common/models-list.ts\nhttps://github.com/upscayl/upscayl/blob/main/renderer/locales/en.json\nhttps://raw.githubusercontent.com/upscayl/upscayl/main/renderer/locales/en.json", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8b93800629a0", "Operation IDs": "375", "Citation": "GitHub · upscayl", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/upscayl/upscayl" }, { "Source ID": "collected-0cce1c946983", "Operation IDs": "375", "Citation": "Upscayl bundled model files", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/upscayl/upscayl/tree/main/models" }, { "Source ID": "collected-e215fa9a6670", "Operation IDs": "375", "Citation": "github.com, v2.15.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/upscayl/upscayl/blob/v2.15.0/package.json" }, { "Source ID": "collected-565850c3a37e", "Operation IDs": "375", "Citation": "github.com, v2.15.0", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/upscayl/upscayl/blob/v2.15.0/common/models-list.ts" }, { "Source ID": "collected-b973909d0fa0", "Operation IDs": "375", "Citation": "github.com — en.json", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/upscayl/upscayl/blob/main/renderer/locales/en.json" }, { "Source ID": "collected-523bf372b146", "Operation IDs": "375", "Citation": "raw.githubusercontent.com — en.json", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/upscayl/upscayl/main/renderer/locales/en.json" } ] }, { "ID": "376", "Software": "Vectorworks", "Operation": "AI Visualizer Generate Image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "3.5–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8405e8286672\ncollected-4b81b537927d\ncollected-4d7d8350edf7\ncollected-95c315f40522\ncollected-4716bb9b140f", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Vectorworks AI Visualizer Generate synthesizes an image from a supplied view.", "Previous record: rationale": "Working v5 documentation coding: 2 kept, 1 model-decided, 2 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 3.5–5.0; no midpoint is used.", "Previous record: calculation": "K=2; M=1; unknown=2; conditional=0; A=5. Lower: 2 + 3 × 2 ÷ (5 − 1) = 3.5. Upper: 2 + 3 × 4 ÷ (5 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Stable Diffusion (Stability AI), version not stated, from the 2024 Update 4 launch. In Vectorworks 2027 (2026-09-16) subscription customers get Google's Nano Banana models: Nano Banana Pro and Nano Banana 2 (credit-limited) plus an unnamed free, unlimited model. Model choice: Yes in Vectorworks 2027: a free, unlimited model, Nano Banana Pro, or Nano Banana 2. Not established: The Stable Diffusion version used from 2024 to 2026 is not stated, and later updates mention an 'updated AI model' without naming it. nemetschek.com, 2024-03-13 • manilatimes.net • vectorworks.net\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (view-conditioned image synthesis). Designer controls: prompt; negative prompt; composition source (viewport, bitmap, design-layer view or image file); style source image; creativity slider; generate similar; aspect ratio; seamless tiling; alpha channel. Returns count not documented; a flat image: PNG; can be placed as bitmap, texture, image fill or image prop. “Indicates how far the generated image should deviate from the source image” Source • Screenshot", "Collected source links": "https://www.vectorworks.net/en-US/newsroom/2025-update-4\nhttps://app-help.vectorworks.net/2025/eng/VW2025_Guide/Rendering2/Generating_AI_images.htm\nhttps://www.nemetschek.com/en/news-media/future-design-here-ai-vectorworks-2024-update-4\nhttps://www.manilatimes.net/2026/09/16/tmt-newswire/globenewswire/vectorworks-2027-delivers-what-designers-need-where-it-matters-most/2426786\nhttps://www.vectorworks.net/en-US/whats-new", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8405e8286672", "Operation IDs": "376", "Citation": "Vectorworks · AI Visualizer", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.vectorworks.net/en-US/newsroom/2025-update-4" }, { "Source ID": "collected-4b81b537927d", "Operation IDs": "376", "Citation": "Vectorworks · Stable Diffusion image generation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://app-help.vectorworks.net/2025/eng/VW2025_Guide/Rendering2/Generating_AI_images.htm" }, { "Source ID": "collected-4d7d8350edf7", "Operation IDs": "376", "Citation": "nemetschek.com, 2024-03-13", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.nemetschek.com/en/news-media/future-design-here-ai-vectorworks-2024-update-4" }, { "Source ID": "collected-95c315f40522", "Operation IDs": "376", "Citation": "manilatimes.net", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.manilatimes.net/2026/09/16/tmt-newswire/globenewswire/vectorworks-2027-delivers-what-designers-need-where-it-matters-most/2426786" }, { "Source ID": "collected-4716bb9b140f", "Operation IDs": "376", "Citation": "vectorworks.net", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.vectorworks.net/en-US/whats-new" } ] }, { "ID": "377", "Software": "Vectorworks Architect", "Operation": "Native BIM Element Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-99de9a67e4db\ncollected-bb3a4c6f7ad6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Vectorworks Architect evaluates manually constructed native BIM elements and their parameters.", "Previous record: rationale": "The scoped model construction is conventional. AI Visualizer and automatic documentation are separate operations.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: Wall, Slab, Window, Door and Space tools; object parameters (height, width, material). Returns one result; editable structure: parametric BIM objects. “these objects are parametric, containing information like height, width, material, and so on.” Source", "Collected source links": "https://www.vectorworks.net/en-US/architect\nhttps://www.vectorworks.net/newsroom/simple-bim-for-the-project-architect", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-99de9a67e4db", "Operation IDs": "377", "Citation": "Vectorworks · Vectorworks BIM Modeling", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.vectorworks.net/en-US/architect" }, { "Source ID": "collected-bb3a4c6f7ad6", "Operation IDs": "377", "Citation": "www.vectorworks.net — simple-bim-for-the-project-architect", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.vectorworks.net/newsroom/simple-bim-for-the-project-architect" } ] }, { "ID": "378", "Software": "Veras in supported CAD hosts", "Operation": "Image to Video Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.6–5.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-529618646dd4\ncollected-d14b9de87fea\ncollected-499fd0ced605\ncollected-6d6ac25881f1\ncollected-b533567130a4\ncollected-e655f43f726e\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Veras produces video from an image through the selected hosted generation route.", "Previous record: rationale": "Working v5 documentation coding: 1 kept, 1 model-decided, 3 unknown and 1 setting-dependent categories among 6 applicable. The recorded uncertainty gives 2.6–5.0; no midpoint is used.", "Previous record: calculation": "K=1; M=1; unknown=3; conditional=1; A=6. Lower: 2 + 3 × 1 ÷ (6 − 1) = 2.6. Upper: 2 + 3 × 5 ÷ (6 − 1) = 5.0. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Checked 4 Oct 2026 Model: Google Veo 3.1 in three tiers (Veo 3.1, Veo 3.1 Fast, Veo 3.1 Lite) from Veras 4.6 (2026-07-09); Wan 2.2 (Alibaba) remains in the model menu. Model choice: Yes. Not established: The model behind the Veras 3.0 image-to-video launch (September 2025) is not named; blog.chaos.com, 2026-07-09 • chaos.com, accessed 2026-10-04 • chaos.com\nDesigner side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (selected hosted video model, e.g. Veo 3.1). Designer controls: image; prompt; preset; model choice. Returns count not documented; a flat video: with generated audio for Veo 3.1. “Turn any image into a dynamic video with Veras using a prompt or preset to instantly bring your visuals to life.” Source • Screenshot", "Collected source links": "https://www.chaos.com/veras\nhttps://blog.chaos.com/veras-4.6-veo-3.1\nhttps://www.chaos.com/press/chaos-unveils-several-new-ai-tools-led-by-veras-30-with-image-to-video-generator-for-aec\nhttps://static.chaos.com/images/assets/000/027/463/articles_2_columns/image-video-690x380.png?1781509368\nhttps://docs-chaos.atlassian.net/wiki/rest/api/content/855867697?expand=body.view\nhttps://docs-chaos.atlassian.net/wiki/rest/api/content/856227941?expand=body.view", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-529618646dd4", "Operation IDs": "378", "Citation": "Chaos · Veras image video and 3D workflows", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.chaos.com/veras" }, { "Source ID": "collected-d14b9de87fea", "Operation IDs": "378", "Citation": "blog.chaos.com, 2026-07-09", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.chaos.com/veras-4.6-veo-3.1" }, { "Source ID": "collected-499fd0ced605", "Operation IDs": "378", "Citation": "www.chaos.com — chaos-unveils-several-new-ai-tools-led-by-veras-30-with-image-to-video-generator-for-aec", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.chaos.com/press/chaos-unveils-several-new-ai-tools-led-by-veras-30-with-image-to-video-generator-for-aec" }, { "Source ID": "collected-6d6ac25881f1", "Operation IDs": "378", "Citation": "static.chaos.com — image-video-690x380.png?1781509368", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://static.chaos.com/images/assets/000/027/463/articles_2_columns/image-video-690x380.png?1781509368" }, { "Source ID": "collected-b533567130a4", "Operation IDs": "378", "Citation": "docs-chaos.atlassian.net — 855867697?expand=body.view", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs-chaos.atlassian.net/wiki/rest/api/content/855867697?expand=body.view" }, { "Source ID": "collected-e655f43f726e", "Operation IDs": "378", "Citation": "docs-chaos.atlassian.net — 856227941?expand=body.view", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs-chaos.atlassian.net/wiki/rest/api/content/856227941?expand=body.view" } ] }, { "ID": "379", "Software": "Veras in supported CAD hosts", "Operation": "Veras 5.0 Trellis 2 Image to 3D Mesh", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-f330c7ed66e2\ncollected-cfc34972542b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Veras 5.0 invokes the specified Trellis 2 image-to-3D option to generate a mesh.", "Previous record: rationale": "Current placement unresolved. K=1; M=0; unknown=3; conditional=0; A=4. Lower: 2 + 3 × 1 ÷ (4 − 1) = 3.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 3.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Trellis2 image to 3D). Designer controls: image; model choice (Trellis2 or Tencent HY3D). Returns count not documented; editable structure: mesh placed into Revit, Rhino or SketchUp at scale. “generate a 3D asset from an image using two AI models, either Tencent's HY3D or Microsoft's Trellis2” Source", "Collected source links": "https://blog.chaos.com/veras-5.0\nhttps://www.chaos.com/veras/whats-new", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-f330c7ed66e2", "Operation IDs": "379", "Citation": "Chaos · Veras 5.0 model choices", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.chaos.com/veras-5.0" }, { "Source ID": "collected-cfc34972542b", "Operation IDs": "379", "Citation": "Chaos · Image-to-3D beta workflow", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.chaos.com/veras/whats-new" } ] }, { "ID": "380", "Software": "Veras in supported CAD hosts", "Operation": "Viewport to Image Generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "Unresolved", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-e1c5a32004b6\ncollected-9285592d768a", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Veras uses the recorded Nano Banana image-model option to synthesize an image from a viewport.", "Previous record: rationale": "Current placement unresolved. K=0; M=0; unknown=1; conditional=4; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. All uncertain categories kept would give K=A; the upper endpoint is outside this equation’s domain.", "Previous record: calculation": "Lower bound 2.0; upper endpoint unresolved. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by a trained model (Nano Banana Pro or Nano Banana 2). Designer controls: prompt; presets; typed reference images (Style, Material, Color, Object, Design Lock, Custom); seed; model version (Nano Banana Pro or Nano Banana 2); sketch and annotation; canvas filters; render selection (Modify, Replace). Returns count not documented; a flat image. “Built-in sketching and annotation tools: mark up and guide the AI directly on your viewport without leaving the plugin.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://www.evolvelab.io/veras\nhttps://blog.chaos.com/veras-v7-vs-standalone-nano-banana", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-e1c5a32004b6", "Operation IDs": "380", "Citation": "EvolveLAB · Viewport Rendering", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.evolvelab.io/veras" }, { "Source ID": "collected-9285592d768a", "Operation IDs": "380", "Citation": "blog.chaos.com — veras-v7-vs-standalone-nano-banana", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.chaos.com/veras-v7-vs-standalone-nano-banana" } ] }, { "ID": "381", "Software": "vvvv gamma", "Operation": "Stride Scene Graph Construction", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d40490a23920\ncollected-e7917d0d0cf2\ncollected-936c380a5c37\ncollected-e8e3a21a0769\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "vvvv gamma evaluates an explicitly constructed VL.Stride scene graph.", "Previous record: rationale": "The bounded native scene construction is conventional. The platform's ability to host learned processing does not establish learning in this graph.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: nodes/parameters (visual patching); models; lights; C# code. Returns continuous live output; editable structure: patch and scene graph. “No matter if you 'patch' visually or you write parts of your code in C#, any change you make is live.” Source", "Collected source links": "https://vvvv.org/\nhttps://thegraybook.vvvv.org/reference/libraries/graphics-3d.html\nhttps://thegraybook.vvvv.org/reference/language/nodes.html\nhttps://thegraybook.vvvv.org/reference/libraries/3d/rendering.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d40490a23920", "Operation IDs": "381", "Citation": "vvvv · Patching", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://vvvv.org/" }, { "Source ID": "collected-e7917d0d0cf2", "Operation IDs": "381", "Citation": "vvvv · Stride 3D scene graph", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://thegraybook.vvvv.org/reference/libraries/graphics-3d.html" }, { "Source ID": "collected-936c380a5c37", "Operation IDs": "381", "Citation": "thegraybook.vvvv.org — nodes.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://thegraybook.vvvv.org/reference/language/nodes.html" }, { "Source ID": "collected-e8e3a21a0769", "Operation IDs": "381", "Citation": "thegraybook.vvvv.org — rendering.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://thegraybook.vvvv.org/reference/libraries/3d/rendering.html" } ] }, { "ID": "382", "Software": "waifu2x in nunif", "Operation": "SwinUNet image upscaling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "7.3", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-421319cded33\ncollected-01baaddcc0fc\ncollected-9449f3f98d72\ncollected-3c2703cb53fa\ncollected-8c59f889ff63", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "waifu2x in nunif runs the specified SwinUNet art model at native twofold enlargement.", "Previous record: rationale": "Rule B (3 Oct 2026): vendor documentation states fidelity-oriented training or non-generative intent for an operation that creates missing samples, which follows the Lightroom Super Resolution comparison: t = 0.75 between 5 and 8, sensitivity t = 0.5–1 (6.5–8).", "Previous record: calculation": "X = 5 + 0.75 × (8 − 5) = 7.25. Sensitivity t = 0.5–1 gives 6.5–8.", "Previous record: notes": "Changed 3 Oct 2026 X 5.0–8.0 → 7.3. The nunif waifu2x CLI documentation (commit of 9 July 2023, current on master) names photo and art_scan as the GAN-based models, and the waifu2x project description calls the photo models GAN based, which leaves the art models outside the GAN set. The SwinUNet art fine-tuning script trains 2x models with the lbp loss and no discriminator, and nunif's LBPLoss applies Charbonnier loss to fixed binary-pattern filter outputs; the pretrained model package is version 20250502. Not established: The exact recipe for the released swin_unet/art weights is documented through example scripts and notes; a per-model training log is not published. Previous reason: Learned detail reconstruction supports 5–8. The current model family is retained without assigning the older implementation's architecture or an undisclosed training objective to it. github.com, commit 9 Jul 2023 • github.com, commit 17 Dec 2023 • github.com, commit 5 Apr 2023", "Collected source links": "https://github.com/nagadomi/waifu2x\nhttps://github.com/nagadomi/nunif/blob/master/waifu2x/hub.py\nhttps://github.com/nagadomi/nunif/blob/master/waifu2x/docs/cli.md\nhttps://github.com/nagadomi/nunif/blob/master/waifu2x/README.md\nhttps://github.com/nagadomi/nunif/blob/master/waifu2x/appendix/finetune_swin_unet_art.sh", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-421319cded33", "Operation IDs": "382", "Citation": "GitHub · waifu2x", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/nagadomi/waifu2x" }, { "Source ID": "collected-01baaddcc0fc", "Operation IDs": "382", "Citation": "waifu2x current model selection", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/nagadomi/nunif/blob/master/waifu2x/hub.py" }, { "Source ID": "collected-9449f3f98d72", "Operation IDs": "382", "Citation": "github.com, commit 9 Jul 2023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/nagadomi/nunif/blob/master/waifu2x/docs/cli.md" }, { "Source ID": "collected-3c2703cb53fa", "Operation IDs": "382", "Citation": "github.com, commit 17 Dec 2023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/nagadomi/nunif/blob/master/waifu2x/README.md" }, { "Source ID": "collected-8c59f889ff63", "Operation IDs": "382", "Citation": "github.com, commit 5 Apr 2023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/nagadomi/nunif/blob/master/waifu2x/appendix/finetune_swin_unet_art.sh" } ] }, { "ID": "383", "Software": "Wilcom EmbroideryStudio 2026", "Operation": "Manual digitizing and visual stitch-design revision", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-46ca2745483f\ncollected-6658c7fa8303", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Wilcom EmbroideryStudio evaluates manually specified outlines, stitch types, densities and sequencing for visual design revision.", "Previous record: rationale": "The scoped digitizing and preview construction contains no documented trained predictor. The artwork's earlier provenance does not change X.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by the designer’s marks. Designer controls: digitizing input tools (Column A/B/C, Complex fill, Complex turning, Ring, Star); outline tools; manual stitch tools; stitch types (satin, tatami); process stitches (density, scale); Color-object list sequencing; resequence by color. Returns one result; editable structure: embroidery objects with stitch settings. “The Traditional Digitizing toolset provides all the input methods embroidery digitizers are familiar with.” Source", "Collected source links": "https://www.wilcom.com/\nhttps://docs.wilcom.com/embroiderystudio/28/en/downloads/QuickReference.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-46ca2745483f", "Operation IDs": "383", "Citation": "Wilcom · Wilcom EmbroideryStudio Digitizing", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://www.wilcom.com/" }, { "Source ID": "collected-6658c7fa8303", "Operation IDs": "383", "Citation": "Wilcom current 2026 visual design commands", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.wilcom.com/embroiderystudio/28/en/downloads/QuickReference.pdf" } ] }, { "ID": "384", "Software": "Womp", "Operation": "Parametric Liquid SDF and CSG Modeling", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-44f20475c284\ncollected-dd1faed90dd3", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Womp evaluates manually specified signed-distance fields and constructive solid geometry with its Liquid 3D controls.", "Previous record: rationale": "The documented combination and blending rules support conventional computation. Smooth or surprising shapes do not establish a model trained to predict them.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by parameters the designer sets, run by rules. Designer controls: SDF/CSG shapes; Goop strength (0 to 200+); negative form (subtract); scene list order; sketch strokes onto model. Returns one result; editable structure: SDF/CSG shapes. “Build with SDF/CSG shapes that blend, merge, and carve naturally. Like clay, but parametric.” Source", "Collected source links": "https://womp.com/\nhttps://womp.com/blogs/the-ultimate-guide-to-goop-liquid-3d-modeling-thats-changing-everything/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-44f20475c284", "Operation IDs": "384", "Citation": "Womp · Liquid 3D Modeling", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://womp.com/" }, { "Source ID": "collected-dd1faed90dd3", "Operation IDs": "384", "Citation": "Womp team — Goop controls and scene-list rules, 05072025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://womp.com/blogs/the-ultimate-guide-to-goop-liquid-3d-modeling-thats-changing-everything/" } ] }, { "ID": "385", "Software": "Zapier MCP / OpenAI", "Operation": "Generate An Image", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-8edc4af09b72\ncollected-c364070efd03\ncollected-c1622d2b8143", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Zapier MCP invokes the OpenAI Generate An Image action.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (OpenAI image model via Zapier action). Designer controls: prompt; number of variations; size; quality; style. Returns several to choose from (number of variations, user-set); form not documented. “you can tell ChatGPT the number of image variations you'd like it to generate, the sizes of the images, the quality, and the style.” Source", "Collected source links": "https://help.zapier.com/hc/en-us/articles/14860148802829-How-to-get-started-with-ChatGPT-OpenAI-on-Zapier\nhttps://docs.zapier.com/mcp/quickstart\nhttps://zapier.com/blog/automate-ai-images/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-8edc4af09b72", "Operation IDs": "385", "Citation": "Zapier · OpenAI image-generation action", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.zapier.com/hc/en-us/articles/14860148802829-How-to-get-started-with-ChatGPT-OpenAI-on-Zapier" }, { "Source ID": "collected-c364070efd03", "Operation IDs": "385", "Citation": "Zapier · MCP tools and client connection", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://docs.zapier.com/mcp/quickstart" }, { "Source ID": "collected-c1622d2b8143", "Operation IDs": "385", "Citation": "zapier.com — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://zapier.com/blog/automate-ai-images/" } ] }, { "ID": "386", "Software": "Adobe Lightroom / Adobe Lightroom Classic", "Operation": "Super Resolution", "Proposed X": "8.9", "Review status": "Proposed X calculated", "Previous X": "7.3Range 6.5–8", "S": "4", "U": "2", "J": "1", "Reviewed on": "2026-10-14", "Configuration": "Raw photograph processed by Super Resolution, including its raw enhancement path; return enlarged reconstructed image.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Raw photograph and Super Resolution enhancement choice.", "Learned determination": "The enhancement model estimates higher-resolution image samples under a fidelity target.", "Conventional processing": "Output handling retains the reconstructed image.", "Model contribution in the result": "Learned raw enhancement and increased image sampling remain in the enlarged photograph.", "Why this scope score (S)": "S4 observed-image reconstruction; U2 fidelity-led sampling recovery; J1 neighboring-pixel context is the positively documented coordination here. A deep CNN label alone does not justify J2 or J3.", "Why this latitude score (U)": "S4 observed-image reconstruction; U2 fidelity-led sampling recovery; J1 neighboring-pixel context is the positively documented coordination here. A deep CNN label alone does not justify J2 or J3.", "Why this coordination score (J)": "S4 observed-image reconstruction; U2 fidelity-led sampling recovery; J1 neighboring-pixel context is the positively documented coordination here. A deep CNN label alone does not justify J2 or J3.", "Comparisons and other modes": "X8.9. Face Recovery is X6.3 under equal weights: it has more latitude and coordination but narrower scope. The order survives the three tested doubled-weight alternatives; it reverses if scope weight exceeds twice the sum of the other two weights.", "Citations": "Adobe Super Resolution development account — https://blog.adobe.com/en/publish/2021/03/10/from-the-acr-team-super-resolution\nAdobe Enhance documentation — https://helpx.adobe.com/lightroom-classic/help/enhance-details.html", "Source IDs": "CAL2-SUPER\nCAL2-SUPERHELP", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies the photograph and chooses enhancement. Learned recovery estimates higher sampling detail; conventional output handling retains it.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. X8.9. Face Recovery is X6.3 under equal weights: it has more latitude and coordination but narrower scope. The order survives the three tested doubled-weight alternatives; it reverses if scope weight exceeds twice the sum of the other two weights.", "Previous record: scope": "Lightroom or Lightroom Classic Super Resolution applies its learned reconstruction to a raw photograph; Denoise and later editing are excluded.", "Previous record: rationale": "Paired low- and high-resolution training and the fidelity-oriented task place this comparison closer to DnCNN-S than ESRGAN: t = 0.75 between 5 and 8. The retained sensitivity interval t = 0.5–1 yields 6.5–8; this is expert calibration, not an Adobe-reported measurement.", "Previous record: calculation": "X = 5 + 0.75 × (8 − 5) = 7.25. Sensitivity range 6.5–8.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (7.3). No newer Adobe documentation on Super Resolution was retrieved: Adobe helpx pages refused fetch and the session's web-search budget ran out before this operation. The 10 March 2021 Adobe account of paired low- and high-resolution training, in the record, remains the mechanism evidence. Not established: Whether Adobe updated the Super Resolution model in Lightroom or Camera Raw releases after 2021; the training loss remains undisclosed. blog.adobe.com, 2021-03-10", "Collected source links": "https://blog.adobe.com/en/publish/2021/03/10/from-the-acr-team-super-resolution\nhttps://helpx.adobe.com/lightroom-classic/help/enhance-details.html\nhttps://web.archive.org/web/20260422080740id_/https://helpx.adobe.com/lightroom-classic/help/photos.html\nhttps://web.archive.org/web/20260922110218id_/https://helpx.adobe.com/lightroom-classic/desktop/process-and-develop-photos/enhance-details.html\nhttps://web.archive.org/web/20251128223715id_/https://helpx.adobe.com/camera-raw/using/ai-edit-status.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "134/15", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-386", "Sources": [ { "Source ID": "CAL2-SUPER", "Operation IDs": "386", "Citation": "Adobe Super Resolution development account", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Adobe Super Resolution development account", "Finding used in this review": "Supervised paired sampling recovery and local neighboring-pixel context; no disclosed whole-image joint mechanism here.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2021/03/10/from-the-acr-team-super-resolution" }, { "Source ID": "CAL2-SUPERHELP", "Operation IDs": "386", "Citation": "Adobe Enhance documentation", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Adobe Enhance documentation", "Finding used in this review": "Raw Super Resolution includes enhancement processing; specify the raw workflow.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://helpx.adobe.com/lightroom-classic/help/enhance-details.html" } ] }, { "ID": "387", "Software": "Google Nano Banana Pro", "Operation": "Gemini 3 Pro Image full-image generation", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-d1f8145d064c\ncollected-32f83c1b97ce", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Google Nano Banana Pro uses the recorded Gemini 3 Pro Image model for full-image generation.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 5 model-decided, 0 unknown and 0 setting-dependent categories among 5 applicable. The category formula gives 2.0.", "Previous record: calculation": "K=0; M=5; unknown=0; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Makes new output. Decided by a trained model (Gemini 3 Pro Image). Designer controls: prompt; reference images (up to 14 objects, 5 characters); aspect ratio; resolution (1k, 2k, 4k). Returns count not documented; a flat image. “Use Gemini's real-world knowledge to deliver precise results, from complex infographics to historically accurate scenes” Source", "Collected source links": "https://deepmind.google/models/gemini-image/pro/\nhttps://deepmind.google/models/model-cards/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-d1f8145d064c", "Operation IDs": "387", "Citation": "Google model account", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://deepmind.google/models/gemini-image/pro/" }, { "Source ID": "collected-32f83c1b97ce", "Operation IDs": "387", "Citation": "Google model cards", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://deepmind.google/models/model-cards/" } ] }, { "ID": "388", "Software": "Adobe Photoshop", "Operation": "Select Subject — automatic selection (Cloud)", "Proposed X": "10.7", "Review status": "Proposed X calculated", "Previous X": "10.0", "S": "2", "U": "0", "J": "3", "Reviewed on": "2026-10-14", "Configuration": "Cloud mode explicitly selected; initial automatic selection on a supplied image, before manual refinement.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Supplied image and explicit Cloud processing choice; initial selection before manual refinement.", "Learned determination": "The model locates the image’s main subject and predicts selection membership.", "Conventional processing": "Selection/mask application retains predicted membership.", "Model contribution in the result": "The returned selection carries semantic subject membership determined by learned processing.", "Why this scope score (S)": "S2 selection field; U0 recognition of observed content; J3 integration of the subject within the image. J3 is a functional inference from the documented main-subject task, not an architecture claim.", "Why this latitude score (U)": "S2 selection field; U0 recognition of observed content; J3 integration of the subject within the image. J3 is a functional inference from the documented main-subject task, not an architecture claim.", "Why this coordination score (J)": "S2 selection field; U0 recognition of observed content; J3 integration of the subject within the image. J3 is a functional inference from the documented main-subject task, not an architecture claim.", "Comparisons and other modes": "Magic Wand locates colors by supplied rules and is X15; Select Subject resolves semantic membership and is X10.7. J2 would give X11.5, so keep the stated whole-subject premise visible.", "Citations": "Adobe Select Subject instructions — https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/detect-subject-using-select-subject.html\nAdobe cloud and device selection modes — https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/improved-select-subject-and-remove-background-results.html\nAdobe Select Subject update report — https://community.adobe.com/questions-712/select-subject-no-longer-working-after-updating-to-ps-2026-1619694", "Source IDs": "CAL2-SUBJECT\nCAL2-CLOUD\nCAL2-REPORTSUBJECT", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies image and chooses the tool. Learned processing identifies the main subject and returns a selection; subsequent masking retains it.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. Magic Wand locates colors by supplied rules and is X15; Select Subject resolves semantic membership and is X10.7. J2 would give X11.5, so keep the stated whole-subject premise visible.", "Previous record: scope": "Photoshop Select Subject in Cloud mode predicts initial selection coverage through the documented learned selection lineage.", "Previous record: rationale": "Region determination matches the SAM task reference at 10. This does not claim that Adobe deploys SAM, and later painting or edge refinement is excluded.", "Previous record: calculation": "X = 8 + 1 × (10 − 8) = 10.", "Previous record: notes": "Checked 3 Oct 2026 X unchanged (10.0). Photoshop 2026 (version 27.x) ships improved Select Subject and Remove Background models; the on-device model was updated toward cloud-model quality, and the Cloud option remains available. The task remains selection; no generative step is documented for Select Subject. Not established: Adobe does not disclose the cloud model architecture or training data. photoshopcafe.com, Photoshop 2026 (27.0) • photoshopcafe.com, 2025-09-11 (Photoshop Beta)", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/detect-subject-using-select-subject.html\nhttps://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/improved-select-subject-and-remove-background-results.html\nhttps://photoshopcafe.com/whats-new-in-photoshop-2026-full-release-overview/\nhttps://photoshopcafe.com/on-device-select-subject-in-photoshop-improved-ai-selections-without-the-cloud/\nhttps://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/history-panel-overview.html\nhttps://web.archive.org/web/20260926081858id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/manage-image-states.html\nhttps://web.archive.org/web/20251113031521id_/https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/detect-subject-using-select-subject.html\nhttps://web.archive.org/web/20260925075845id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/history-panel-overview.html\nhttps://web.archive.org/web/20260922053813id_/https://helpx.adobe.com/photoshop/desktop/get-started/settings-and-preferences/adjust-preferences.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "32/3", "Retained learning": "True", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-388", "Sources": [ { "Source ID": "CAL2-SUBJECT", "Operation IDs": "388", "Citation": "Adobe Select Subject instructions", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Adobe Select Subject instructions", "Finding used in this review": "Automatic main-subject selection; current page updated February 23 2026.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/detect-subject-using-select-subject.html" }, { "Source ID": "CAL2-CLOUD", "Operation IDs": "388", "Citation": "Adobe cloud and device selection modes", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Adobe cloud and device selection modes", "Finding used in this review": "Cloud is a consequential mode; assessment specifies cloud explicitly rather than relying on defaults.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/improved-select-subject-and-remove-background-results.html" }, { "Source ID": "CAL2-REPORTSUBJECT", "Operation IDs": "388", "Citation": "Adobe Select Subject update report", "Type": "user_report_or_forum", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Adobe Select Subject update report", "Finding used in this review": "Version/mode failure report is a reason to pin cloud configuration; it is not evidence of internal architecture.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://community.adobe.com/questions-712/select-subject-no-longer-working-after-updating-to-ps-2026-1619694" } ] }, { "ID": "389", "Software": "Adobe Photoshop", "Operation": "Manual Brush editing of an existing layer mask", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-57a36df33987\ncollected-ef165773d5b7\ncollected-7e584a41a9c6\ncollected-aada92b2030e\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop's ordinary Brush tool changes a supplied layer mask with a manually specified stroke.", "Previous record: rationale": "Prescribed brush application contains no in-scope learned determination. Any earlier automatic selection is outside this separate edit boundary.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Brush tool; Brush Picker size and hardness; Opacity; foreground color black/white/gray. Returns one result; editable structure: layer mask. “black on a layer mask hides the layer that contains the mask, so you can see what is underneath that layer.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/using/editing-layer-masks.html\nhttps://web.archive.org/web/20260926081858id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/manage-image-states.html\nhttps://web.archive.org/web/20260131081935id_/https://helpx.adobe.com/photoshop/using/editing-layer-masks.html\nhttps://web.archive.org/web/20260925075845id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/history-panel-overview.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-57a36df33987", "Operation IDs": "255, 389", "Citation": "Adobe Edit layer masks, July 18, 2024", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/using/editing-layer-masks.html" }, { "Source ID": "collected-ef165773d5b7", "Operation IDs": "257, 388, 389", "Citation": "web.archive.org — manage-image-states.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260926081858id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/manage-image-states.html" }, { "Source ID": "collected-7e584a41a9c6", "Operation IDs": "389", "Citation": "web.archive.org — editing-layer-masks.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260131081935id_/https://helpx.adobe.com/photoshop/using/editing-layer-masks.html" }, { "Source ID": "collected-aada92b2030e", "Operation IDs": "257, 388, 389", "Citation": "web.archive.org — history-panel-overview.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260925075845id_/https://helpx.adobe.com/photoshop/desktop/get-started/set-up-toolbars-panels/history-panel-overview.html" } ] }, { "ID": "390", "Software": "Adobe Photoshop", "Operation": "Magic Wand — color-based selection", "Proposed X": "15", "Review status": "Proposed X calculated", "Previous X": "15.0", "S": "0", "U": "0", "J": "0", "Reviewed on": "2026-10-14", "Configuration": "Sample color, tolerance, contiguity, antialiasing and layer options fixed; return color-based selection.", "Next step": "Review this v2 working proposal and its configuration limits. Final method adoption remains open.", "Human input and controls": "Sampled color, tolerance, contiguity, antialiasing and layer options.", "Learned determination": "None in the scoped Magic Wand path.", "Conventional processing": "Color comparisons and connectivity produce the selection.", "Model contribution in the result": "No learned prediction contributes.", "Why this scope score (S)": "S0/U0/J0: semantic subject identity is not inferred by the scoped color-selection function.", "Why this latitude score (U)": "S0/U0/J0: semantic subject identity is not inferred by the scoped color-selection function.", "Why this coordination score (J)": "S0/U0/J0: semantic subject identity is not inferred by the scoped color-selection function.", "Comparisons and other modes": "A complicated boundary or one-click use does not make the selection learned. This is the matched conventional comparator for Select Subject.", "Citations": "Adobe Magic Wand instructions — https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/select-areas-by-color-with-the-magic-wand-tool.html", "Source IDs": "CAL2-WAND", "Supplemental findings": "Completed documentary calibration transferred under the author’s publication/incorporation request. Human supplies sampling rules. Conventional comparisons and connectivity determine selected pixels.", "Evidence limits": "Source-linked expert coding, not software trials or measured discovery probability. A complicated boundary or one-click use does not make the selection learned. This is the matched conventional comparator for Select Subject.", "Previous record: scope": "Photoshop Magic Wand selects colors using the supplied sample and documented tolerance and connectivity rules.", "Previous record: rationale": "Conventional color comparison determines the region without a documented trained predictor. Automatic selection is therefore insufficient evidence for the learned branch.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by parameters the designer sets, run by rules. Designer controls: click sample point; Tolerance (0-255); Contiguous; Anti-alias; Sample All Layers. Returns one result; editable structure: pixel selection. “Tolerance sets the range of the selected color values from 0 (narrow) to 255 (broad).” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/select-areas-by-color-with-the-magic-wand-tool.html", "Evidence status": "Mechanism evidence supports proposed ratings", "Raw X (fraction)": "15", "Retained learning": "False", "Method ID": "irm-x-calibration-proposal-20261014-2", "Method SHA": "7d6564b1c5e3a6ba04c3bad31788fb173321afaf6ef707de7d866ec968a9eed0", "Assessment ID": "calibration-v2-390", "Sources": [ { "Source ID": "CAL2-WAND", "Operation IDs": "390", "Citation": "Adobe Magic Wand instructions", "Type": "primary", "Source date": "", "Accessed": "2026-10-14", "Locator": "Calibration source record: Adobe Magic Wand instructions", "Finding used in this review": "Sampled color, tolerance, contiguity, antialiasing and layer sampling.", "Evidence limits": "Documentary evidence; no local software trial or accuracy benchmark.", "Review status": "Calibration evidence; see source limits", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/make-selections/automatic-color-based-selections/select-areas-by-color-with-the-magic-wand-tool.html" } ] }, { "ID": "391", "Software": "Adobe Photoshop", "Operation": "Pen tool — manually specified path", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-62ed56c7dec6", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop's standard Pen tool constructs a path from supplied anchors and direction handles.", "Previous record: rationale": "Conventional curve evaluation determines the path. Content-Aware Tracing and automatic boundary detection are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Builds structure from the designer’s input. Decided by the designer’s marks. Designer controls: Pen tool; click to place anchor points; drag to create direction handles; corner and smooth points. Returns one result; editable structure: vector path in Paths panel. “The positions of direction lines and points determine the size and shape of a curved segment” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/draw-shapes-paths/draw-lines-curves/draw-paths-with-the-pen-tool.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-62ed56c7dec6", "Operation IDs": "391", "Citation": "Adobe Pen tool guide, February 23, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/draw-shapes-paths/draw-lines-curves/draw-paths-with-the-pen-tool.html" } ] }, { "ID": "392", "Software": "Adobe Photoshop", "Operation": "Path to selection — Make Selection", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-cf63a2ef2baf", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop Make Selection rasterizes a supplied closed path with specified feathering and anti-aliasing.", "Previous record: rationale": "The path conversion is prescribed computation; construction of the path and any preceding learned boundary prediction are outside this stage.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Make Selection dialog; Feather Radius (pixels); Anti-aliased; Operation (New, Add, Subtract, Intersect); Load Path As A Selection button. Returns one result; editable structure: selection border. “Defines how far inside and outside the selection border the feather edge extends.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/using/converting-paths-selection-borders.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-cf63a2ef2baf", "Operation IDs": "392", "Citation": "Adobe path/selection conversion guide, May 24, 2023", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/using/converting-paths-selection-borders.html" } ] }, { "ID": "393", "Software": "Adobe Premiere", "Operation": "Razor cut at a manually specified position", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-dd9019889b1d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere Razor splits a supplied timeline clip at the editor's specified position while retaining both pieces.", "Previous record: rationale": "The split changes source-interval bookkeeping conventionally. No trained prediction determines the cut position in this operation.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Razor tool click position; Alt/Option-click for audio or video only; Sequence > Add Edit (Ctrl/Command+K); Add Edit to All Tracks. Returns one result; editable structure: two timeline clip instances. “Cutting a clip creates a new, separate instance of the original, with updated In and Out points” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/premiere/desktop/edit-projects/trim-clips/cut-clips.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-dd9019889b1d", "Operation IDs": "393", "Citation": "Adobe Cut clips, August 22, 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/edit-projects/trim-clips/cut-clips.html" } ] }, { "ID": "394", "Software": "Adobe Premiere", "Operation": "Manual trim of a clip edge", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-fc202ccaa3fb", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere trims one clip edge to a manually specified frame through ordinary edge dragging.", "Previous record: rationale": "The supplied endpoint changes a source interval conventionally. Generative Extend and other trim modes are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Selection tool; drag clip edge at trim icon; Trim mode (Shift+T). Returns one result; editable structure: timeline clip In/Out points. “Hover your mouse over the edge of the clip until the trim icon appears.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/premiere/desktop/edit-projects/trim-clips/trim-a-clip.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-fc202ccaa3fb", "Operation IDs": "394", "Citation": "Adobe Trim a clip, August 22, 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/edit-projects/trim-clips/trim-a-clip.html" } ] }, { "ID": "395", "Software": "Adobe Premiere", "Operation": "Manual clip rearrangement — insert move", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-9aeea20c60dd\ncollected-d92b2f4b9961\ncollected-6a320b9624b0\ncollected-0fd36e49e2d8\ncollected-1ab0f12c39a7\ncollected-e613a38b4206", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere performs an insert move of an existing clip to a specified timeline destination.", "Previous record: rationale": "Documented track and interval rules determine the rearrangement. No learned choice of destination is included.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: drag clip; Ctrl/Command on release (insert); Ctrl+Alt/Command+Option (rearrange); Shift while dragging. Returns one result; editable structure: timeline clips. “When you perform an insert edit this way, only clips in the destination track are shifted; other tracks are unaffected.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/premiere/desktop/edit-projects/change-clip-sequence/rearrange-clips-on-the-timeline.html\nhttps://web.archive.org/web/20260823142358id_/https://helpx.adobe.com/premiere/desktop/edit-projects/correct-mistakes/view-or-make-changes-in-the-history-panel.html\nhttps://web.archive.org/web/20260927163748id_/https://helpx.adobe.com/premiere/desktop/collaborate-with-others/collaborate-using-productions/copy-clips-across-projects-in-a-production.html\nhttps://web.archive.org/web/20260929204046id_/https://helpx.adobe.com/premiere/desktop/edit-projects/change-clip-sequence/rearrange-clips-on-the-timeline.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-9aeea20c60dd", "Operation IDs": "395", "Citation": "Adobe Rearrange clips, January 7, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/premiere/desktop/edit-projects/change-clip-sequence/rearrange-clips-on-the-timeline.html" }, { "Source ID": "collected-d92b2f4b9961", "Operation IDs": "395", "Citation": "web.archive.org — view-or-make-changes-in-the-history-panel.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260823142358id_/https://helpx.adobe.com/premiere/desktop/edit-projects/correct-mistakes/view-or-make-changes-in-the-history-panel.html" }, { "Source ID": "collected-6a320b9624b0", "Operation IDs": "395", "Citation": "web.archive.org — copy-clips-across-projects-in-a-production.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260927163748id_/https://helpx.adobe.com/premiere/desktop/collaborate-with-others/collaborate-using-productions/copy-clips-across-projects-in-a-production.html" }, { "Source ID": "collected-0fd36e49e2d8", "Operation IDs": "273, 395", "Citation": "web.archive.org — rearrange-clips-on-the-timeline.html", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://web.archive.org/web/20260929204046id_/https://helpx.adobe.com/premiere/desktop/edit-projects/change-clip-sequence/rearrange-clips-on-the-timeline.html" } ] }, { "ID": "396", "Software": "Adobe Premiere", "Operation": "Close Gap — Sequence command", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-07e07cd75757\ncollected-134bb0a47a08\ncollected-d5b0f252f46d", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Premiere Sequence Close Gap closes eligible empty timeline intervals after prior edits.", "Previous record: rationale": "The documented command applies conventional interval handling. Exact multitrack eligibility requires configuration checks, but the available operation account provides no trained content decision.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Sequence > Close Gap menu command; assignable keyboard shortcut. Returns one result; editable structure: timeline clips. “It removes all gaps between clips, no matter what video or audio layer they are on.” Source", "Collected source links": "https://community.adobe.com/questions-729/transfering-edits-to-different-tracks-1416345\nhttps://blog.adobe.com/en/publish/2025/01/26/sundance-2025-behind-the-scenes-with-the-editor-of-bunnylovr\nhttps://blog.frame.io/2024/01/10/insider-tips-premiere-pros-automations-for-cleaning-timelines/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-07e07cd75757", "Operation IDs": "396", "Citation": "Adobe staff explanation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://community.adobe.com/questions-729/transfering-edits-to-different-tracks-1416345" }, { "Source ID": "collected-134bb0a47a08", "Operation IDs": "396", "Citation": "Adobe editor interview", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.adobe.com/en/publish/2025/01/26/sundance-2025-behind-the-scenes-with-the-editor-of-bunnylovr" }, { "Source ID": "collected-d5b0f252f46d", "Operation IDs": "396", "Citation": "blog.frame.io — ", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://blog.frame.io/2024/01/10/insider-tips-premiere-pros-automations-for-cleaning-timelines/" } ] }, { "ID": "397", "Software": "Adobe Photoshop", "Operation": "Generative Expand with Firefly", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "2.0–3.5", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-dc90669db1e8\ncollected-897a3b739bb0", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "Photoshop Generative Expand with Firefly synthesizes image content outside the original canvas.", "Previous record: rationale": "Working v5 documentation coding: 0 kept, 3 model-decided, 2 unknown and 0 setting-dependent categories among 5 applicable. The recorded uncertainty gives 2.0–3.5; no midpoint is used.", "Previous record: calculation": "K=0; M=3; unknown=2; conditional=0; A=5. Lower: 2 + 3 × 0 ÷ (5 − 1) = 2.0. Upper: 2 + 3 × 2 ÷ (5 − 1) = 3.5. No midpoint is assigned. Current profile and next action", "Previous record: notes": "Designer side, 3 Oct 2026 Other. Decided by a trained model (Firefly Fill & Expand, Image 3 or Image 1). Designer controls: Crop tool handles (new canvas size); prompt (optional); model choice (Firefly Fill & Expand, Firefly Image 3, Firefly Image 1). Returns several to choose from (count not stated); editable structure: Generative Layer. “Select Generate and choose from the variations that appear in the Properties panel.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://helpx.adobe.com/photoshop/desktop/create-open-import-images/create-images/explore-beyond-the-canvas-with-generative-expand.html\nhttps://research.adobe.com/news/photoshop-turns-35/", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-dc90669db1e8", "Operation IDs": "397", "Citation": "Adobe Photoshop Generative Expand guide, updated February 23, 2026", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://helpx.adobe.com/photoshop/desktop/create-open-import-images/create-images/explore-beyond-the-canvas-with-generative-expand.html" }, { "Source ID": "collected-897a3b739bb0", "Operation IDs": "253, 397", "Citation": "Adobe Research, September 15, 2025", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://research.adobe.com/news/photoshop-turns-35/" } ] }, { "ID": "398", "Software": "Autodesk 3ds Max", "Operation": "Manual move of one Editable Poly vertex", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-3cdd0d29418a\ncollected-9e4c4926954c\ncollected-8c94cc21042c", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "3ds Max moves one specified Editable Poly vertex by a supplied displacement, excluding Soft Selection and downstream modifiers.", "Previous record: rationale": "The coordinate update and connected geometry evaluation are conventional. This is a direct edit, separate from automatic retopology.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Vertex sub-object level; vertex selection; Move transform (standard methods). Returns one result; editable structure: Editable Poly mesh. “At the Vertex sub-object level, you can select single and multiple vertices and move them using standard methods.” Source • Screenshot", "Collected source links": "https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Modifiers/files/GUID-047B00EC-D4E3-4D05-8FB7-220795D13874.htm\nhttps://help.autodesk.com/cloudhelp/2026/ENU/MAXScript-Help/files/3ds-Max-Objects-and-Interfaces/Editable-Meshes-Splines-Patches/Editable_Poly/Editable_Poly-Methods/GUID-07318A19-8091-4B3E-BB46-463942B0F39B.html\nhttps://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Modifiers/images/GUID-924F0ACB-B988-4FD6-8DC7-EAAF9A97E1D3.png", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-3cdd0d29418a", "Operation IDs": "398", "Citation": "Autodesk 3ds Max 2026 Editable Poly Vertex guide", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Modifiers/files/GUID-047B00EC-D4E3-4D05-8FB7-220795D13874.htm" }, { "Source ID": "collected-9e4c4926954c", "Operation IDs": "398", "Citation": "Autodesk 2026 vertex-method reference", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/MAXScript-Help/files/3ds-Max-Objects-and-Interfaces/Editable-Meshes-Splines-Patches/Editable_Poly/Editable_Poly-Methods/GUID-07318A19-8091-4B3E-BB46-463942B0F39B.html" }, { "Source ID": "collected-8c94cc21042c", "Operation IDs": "398", "Citation": "help.autodesk.com — GUID-924F0ACB-B988-4FD6-8DC7-EAAF9A97E1D3.png", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.autodesk.com/cloudhelp/2026/ENU/3DSMax-Modifiers/images/GUID-924F0ACB-B988-4FD6-8DC7-EAAF9A97E1D3.png" } ] }, { "ID": "399", "Software": "Maxon ZBrush", "Operation": "ZModeler Move Point — single Point target", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-45cbce9e8331\ncollected-2992285c5a24", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ZBrush ZModeler Move with a single Point target displaces the selected point through a user-directed drag.", "Previous record: rationale": "The recorded view and movement determine the geometric update. Broader targets, snapping, symmetry and later sculpting are excluded.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Applies the designer’s marks to existing work. Decided by the designer’s marks. Designer controls: Move action; Point target; click-drag; Shift to constrain to the X, Y or Z working plane. Returns one result; editable structure: mesh. “The Move Action lets you freely move the selected point relative to the screen's working plane.” Source", "Collected source links": "https://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/modeling-basics/creating-meshes/zmodeler/zmodeler-actions/point-actions/point-actions.html\nhttps://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/modeling-basics/creating-meshes/zmodeler/targets/point-targets/point-targets.html", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-45cbce9e8331", "Operation IDs": "399", "Citation": "Maxon desktop ZModeler Point Actions, Move section", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/modeling-basics/creating-meshes/zmodeler/zmodeler-actions/point-actions/point-actions.html" }, { "Source ID": "collected-2992285c5a24", "Operation IDs": "399", "Citation": "Maxon desktop ZModeler Point Targets", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://help.maxon.net/zbr/en-us/Content/html/user-guide/3d-modeling/modeling-basics/creating-meshes/zmodeler/targets/point-targets/point-targets.html" } ] }, { "ID": "400", "Software": "ComfyUI", "Operation": "ImageCompositeMasked with supplied images and mask", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-84f461b67be0\ncollected-38047174c32b", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "ComfyUI ImageCompositeMasked combines supplied images and mask by the pinned node's conventional alpha arithmetic.", "Previous record: rationale": "The node neither predicts the mask nor regenerates the source. The blend retains no newly learned determination within this isolated operation; generation-plus-compositing would be a different boundary.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: destination image; source image; x; y; resize_source; mask. Returns one result; a flat image. “destination[..., top:bottom, left:right] = source_portion + destination_portion” Source", "Collected source links": "https://github.com/Comfy-Org/ComfyUI/blob/9c90340bc26ea69c2029ee6e882acbbb06e157b4/comfy_extras/nodes_mask.py\nhttps://raw.githubusercontent.com/Comfy-Org/ComfyUI/9c90340bc26ea69c2029ee6e882acbbb06e157b4/node_helpers.py", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-84f461b67be0", "Operation IDs": "400", "Citation": "Pinned implementation", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://github.com/Comfy-Org/ComfyUI/blob/9c90340bc26ea69c2029ee6e882acbbb06e157b4/comfy_extras/nodes_mask.py" }, { "Source ID": "collected-38047174c32b", "Operation IDs": "400", "Citation": "pinned helper", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://raw.githubusercontent.com/Comfy-Org/ComfyUI/9c90340bc26ea69c2029ee6e882acbbb06e157b4/node_helpers.py" } ] }, { "ID": "401", "Software": "Blackmagic Design DaVinci Resolve", "Operation": "Manual primary master Gain adjustment", "Proposed X": "", "Review status": "Awaiting reassessment", "Previous X": "15.0", "S": "", "U": "", "J": "", "Reviewed on": "", "Configuration": "", "Next step": "Research configurations and calculate proposed X", "Human input and controls": "", "Learned determination": "", "Conventional processing": "", "Model contribution in the result": "", "Why this scope score (S)": "", "Why this latitude score (U)": "", "Why this coordination score (J)": "", "Comparisons and other modes": "", "Citations": "", "Source IDs": "collected-138132678af4\ncollected-5667a78a3d1b\ncollected-fd049ba63edb", "Supplemental findings": "", "Evidence limits": "", "Previous record: scope": "DaVinci Resolve applies a manually selected primary master Gain adjustment in a fixed color-management context.", "Previous record: rationale": "The tonal operation is prescribed, not a learned estimate of a desired correction. The unpublished exact shader prevents claiming one universal channel multiplier but does not establish a trained predictor.", "Previous record: calculation": "X = 15 by the no retained learning branch.", "Previous record: notes": "Designer side, 3 Oct 2026 Transforms supplied material. Decided by parameters the designer sets, run by rules. Designer controls: Gain master wheel; Gain color balance wheel; Lift and Gamma master wheels. Returns one result; editable structure: correction node parameters. “You can use the Gain master wheel to brighten the lighter areas of the image.” Source Quote taken from a search result for the vendor page.", "Collected source links": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf\nhttps://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-20-Colorist-Guide.pdf\nhttps://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-18-Colorist-Guide.pdf", "Evidence status": "Awaiting reassessment under proposed method", "Raw X (fraction)": "", "Retained learning": "", "Method ID": "", "Method SHA": "", "Assessment ID": "", "Sources": [ { "Source ID": "collected-138132678af4", "Operation IDs": "71, 72, 74, 75, 76, 77, 401", "Citation": "Resolve21 manual, p3204", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf" }, { "Source ID": "collected-5667a78a3d1b", "Operation IDs": "401", "Citation": "Resolve20 Colorist guide, pp28–31", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-20-Colorist-Guide.pdf" }, { "Source ID": "collected-fd049ba63edb", "Operation IDs": "401", "Citation": "documents.blackmagicdesign.com — DaVinci-Resolve-18-Colorist-Guide.pdf", "Type": "Collected lead", "Source date": "", "Accessed": "", "Locator": "Link imported from previous Word record", "Finding used in this review": "", "Evidence limits": "", "Review status": "Awaiting source review", "Open source": "Open source", "URL": "https://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-18-Colorist-Guide.pdf" } ] } ] }