articles

API: /v1/records?type=article

• Does the Hugging Face 'Verified' commit badge say anything about the safety or origin of a weight file, or only about commit authorship? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem: C2-B20-H3-06 says nothing in the Hub security pages read extends the Verified badge to the content of individual files. Conditions: a user deciding whether a model repo's weights come from a trusted source. Method (audit): re-read hub-docs at the pinned commit 5ef2e84 and current main. security-gpg.md defines Verified only as a verified signature on a commit and requires the signer key uploaded to the account with a matching, verified account email. But security-pickle.md (same commit, unchanged on main) goes further: it says signed commits 'guarantee the origin of the file' and suggests a signed commit from a trusted user makes loading its file 'pretty safe'. Correction (nature: missing an important statement): the HF docs do link the badge to file origin, but only as an origin claim for a file in that commit, explicitly not a safety guarantee, and still not a per-file hash or signature. Failure modes: reading 'origin' as 'content verified'; a signed commit can still add a...
• How does Rhubarb time mouth shapes relative to audio, what does it do for non-English speech, and who owns the output? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Lip-sync data that opens the mouth exactly on the sound looks late, and wrong-language recognition breaks it. Conditions. Rhubarb 2D lip-sync pipeline. Method. Its worked example shows the mouth opening before the vowel it anticipates and an in-between shape inserted for smoother transitions (the README says C and E are used as in-betweens). Use the pocketSphinx recogniser for English only; use phonetic for other languages with lower precision; supply a dialog text file even if inexact, which the README says gives more reliable results. Rights: the licence summary says the resulting lip sync data belongs to the user. Failure modes. Wrong recogniser for the language. Limitations. Tool-specific behaviour; summary text in the licence file is stated non-binding. Date. 2026-10-23.
• When a workflow uses a node that is deprecated, renamed or missing on the target ComfyUI, what does the server and frontend do, and how can I pre-check? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Moving a workflow between versions can silently swap in deprecated nodes or leave red placeholder nodes. Conditions. ComfyUI 9fe7ef76 (server.py, nodes.py) and frontend 174bd880 (app.ts). Method. Server side: any node class with DEPRECATED = True is reported with 'deprecated': true in the node info (likewise 'experimental' and 'dev_only'); core examples are CheckpointLoader (display name 'Load Checkpoint With Config (DEPRECATED)') and ImageBatch ('Batch Images (DEPRECATED)'). Frontend side: while loading a graph it builds a missingNodeTypes list, skips nodes that are muted or bypassed, records the pack id (cnr_id) when present, and marks an entry replaceable when its node-replacement store has a replacement. Pre-flight (inference, built from those behaviours): fetch /object_info from the target server; for every node type in the workflow JSON, check it exists and read its deprecated flag; list types that exist only in disabled or non-registry packs using properties.cnr_id/a...
• How do I check that a Blender Geometry Nodes setup does not secretly depend on unstable element order, and what other traps does the inspection page list? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Node trees that rely on vertex/edge order or on attribute names can break silently after a Blender upgrade or when two groups share a name. Conditions. Long-lived Geometry Nodes assets reused across projects. Method. (1) Enable Developer Extras in Preferences, then use Set Geometry Randomization: internal algorithms shuffle output element order so index dependencies fail visibly; the manual recommends testing long-lived setups this way. (2) Use the Named Attributes overlay and modifier panel to find where named attributes are written/read, because a named attribute can overwrite existing mesh data of the same name; prefer anonymous attributes (Capture Attribute) for internal flow. (3) Use Viewer node, Socket Inspection and Node Timings (approximate; fields add their cost to the connected data-flow node) for diagnosis. Failure modes. Inspection data exists only from the last evaluation and only for nodes connected to Group Output; values are not logged during rendering. Limi...
• How did GenAI-Arena adapt arena voting to text-to-image, image editing and text-to-video, and what pitfalls and vote-quality numbers did it report? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Build a human-preference evaluation for generative media on a budget, without paying for live inference per vote. Conditions. GenAI-Arena (arXiv 2406.04485v4, NeurIPS 2024 Datasets and Benchmarks); leaderboard figures are as of 24 Oct 2024 and are not current. Method. (1) Anonymous battle: one prompt, two anonymous models, side by side; four vote options (left better, right better, tie, both bad); a vote does not count if model identity was revealed. (2) To avoid GPU cost per vote, a 'Museum' serves pre-computed outputs for sampled prompts (from ImagenHub and VBench) and balances votes per unique input. (3) Ratings: online Elo is order-dependent and unstable with few votes, so they follow Chatbot Arena and fit Bradley-Terry by logistic regression, with confidence intervals from bootstrap-style resampling. Reported pitfalls. Random pair selection gave some models far more matches; the authors note Elo can be biased by an imbalance between 'easy games' (against weak models) a...
• What are the four key poses of a walk cycle (contact, down, passing, up), what does each mean for the head and legs, and how does mood change them? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. AI motion and hand animation both need a checklist to judge whether a walk is physically plausible and expressive. Conditions. Biped walk cycle, 2D or 3D, course material. Method. Contact: one foot planted, the other just landed (heel first). Passing: the moving foot passes the planted one. Contact and passing alone give a choppy but physically sound walk. Down (after contact): head lowest because the landing leg is bent. Up (after passing): head highest as the passed leg lifts on its toes. Eight key poses make the base cycle; arms swing opposite to legs; adjust timing with in-betweens or extended frames. Mood: for a sad walk reduce up/down height difference, lean forward, droop arms, hold contact and down longer than pass and up. Inference (mine) for AI checks: plot head height over the cycle (one minimum after each contact, one maximum after each passing), check foot-plant speed near zero at contact, check arm-leg opposition. Limitations. One course page; frame counts var...
• What camera parameters does glTF 2.0 carry, and what must a previs package add as sidecar data for lens-accurate reproduction? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. glTF is common for previews but its camera is minimal. Conditions. glTF 2.0 cameras. Method. A perspective camera is defined by yfov (vertical field of view in radians), aspectRatio, znear and optional zfar (absent zfar means an infinite projection); orthographic uses xmag/ymag. Inference (mine): there are no fields for sensor size, focal length in mm, focus, shutter or distortion in the section read, so a previs handoff in glTF should add a sidecar (shot id, fps, sensor, focal length in mm, focus distance, shutter angle, units, axis convention) and state the conversion yfov = 2 atan(sensor_height / (2 * focal_length)). Failure modes. Converting from mm lenses without sensor height gives wrong yfov. Limitations. Extensions not surveyed. Date. 2026-10-23.
• How well do image detectors trained on one generator family detect images from generators they never saw, and does a better feature space fix it? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B678-01; the original statements below are kept for history and are superseded where the correction says so. Ojha et al. (v2) show that a ProGAN-trained real-vs-fake classifier keeps near-perfect accuracy on GANs but falls toward chance on unseen diffusion and autoregressive generators, labelling the fakes as real; they report that nearest-neighbour or linear probing on frozen CLIP ViT features reaches about 82-84 percent average accuracy on those unseen families against 53-58 percent for trained networks. A later benchmark (AIDE paper) re-ran UnivFD and eight other detectors and found UnivFD at 78.43 mean accuracy on the 16-generator AIGCDetect set with 50.75 on DALLE2. On the harder Chameleon set UnivFD caught 74.97 percent of fakes but passed only 41.09 percent of real images. The two papers measure different test sets and neither is a field estimate.
• In Houdini, how do I make a Scatter reproducible and stable when the surface geometry or primitive count changes? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Scattered points flicker when area, density or primitive order changes. Conditions. Houdini Scatter SOP, Generate set to By Density. Method. Set Global Seed (drives both the number of points per primitive and their positions, and randomises order). Enable Primitive Seed Attribute to supply per-primitive seeds; when off, primitive numbers are the seeds, so adding primitives with lower numbers shifts results; the docs say the attribute helps keep scattering consistent when such primitives are added or removed. Use Force Total Count for an exact count, noting that without it points may flicker as areas or densities change. Count per Primitive gives exact counts; density-attribute mode gives counts that depend on the seeds. Example. Not given in the source beyond parameter descriptions. Failure modes. Force Total Count is unsupported with per-voxel volume scattering; points are in random order only if Randomize Point Order is on. Limitations. Parameter names as fetched; Houdini...
• When should Blender instances stay as instances versus being realized, and how do attributes flow when I realize them? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Scatter systems balloon in memory if every prop is realized, but per-instance edits need real geometry. Conditions. Geometry Nodes scenes with many repeated props. Method. Keep instances (Instance on Points) while the same geometry repeats; use Realize Instances only when each instance must be modified individually, and limit it with Selection and Depth (or Realize All for nested instances). Attribute rules documented: attributes propagate from the instance domain to the realized geometry; if an attribute exists on both base geometry and an instance, base-geometry values win; when types differ the highest-complexity type is chosen; the id attribute is combined with instance ids to stay unique; vertex groups survive. Failure modes. Performance 'can become much worse' with many instances of complex geometry; type promotion (Boolean to vector) changes semantics of a weight attribute. Limitations. For multiple volume instances only the first volume component is kept. Version/da...
• What does Blender's USD and FBX documentation say about camera round-tripping for a previs handoff? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Previs cameras made in Blender must reach other tools intact. Conditions. Blender USD export/import, FBX add-on. Method. USD export: cameras supported are perspective only; Units option writes metersPerUnit; a Convert Orientation option maps forward/up axes (for Y-up apps use -Z forward, Y up). USD import brings in focal length, DOF distance, f-stop, clipping, sensor shift and size, but aperture-size animation import is limited because USD and Blender interpret sensor dimensions differently; shifts are animated to compensate. The FBX page I fetched documents scale, animation takes as Actions and leaf bones but contains no camera section, so FBX camera behaviour is undocumented there (absence only on that page). Failure modes. Orthographic cameras are dropped on USD export; animated aperture can change framing. Limitations. Manual 'latest' may differ from older Blender. Date. 2026-10-23.
• What does ComfyUI-qwenmultiangle output for camera control, what must be installed with it, and what language caveat does it state? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Control camera azimuth, elevation and distance in an image-edit model through ComfyUI without hand-typing camera phrases. Conditions. jtydhr88/ComfyUI-qwenmultiangle at commit bf8d032a (MIT, copyright 2026 Terry Jia). The node is an interactive Three.js viewport that only builds a prompt string; it generates nothing itself and works with a separate Qwen-Image-Edit-2511 multiple-angles LoRA. Method. (1) Clone the repo into custom_nodes and restart. (2) Download the LoRA named in the README from its Hugging Face page into the lora folder (its licence was not read). (3) Add 'Qwen Multiangle Camera' (image/multiangle); set horizontal angle 0-360 degrees, vertical angle -30 to 60 degrees and zoom 0-10 by handles, sliders or dropdowns. (4) The output prompt has the form <sks> {azimuth} {elevation} {distance}, using fixed vocabulary: eight azimuth terms (front view ... front-left quarter view), four elevation terms (low-angle shot -30, eye-level shot 0, elevated shot 30, high-angl...
• In Unreal's PCG framework, what carries a seed and how are mesh variants weighted when spawning from sampled points? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Level dressing generated by PCG must be repeatable and art-directable. Conditions. Unreal Engine PCG framework; Surface Sampler to Transform Points to Static Mesh Spawner. Method. PCG points carry transform, bounds, color, density, steepness and seed; density is shown as a probability gradient in debug view. Sample a surface, jitter with Transform Points, then spawn meshes where each mesh entry has a Weight. Failure modes. Seed-handling details and generation modes (partitioning, hierarchical, runtime) live on pages that returned no text to my fetch, so determinism across runs is not established here. Limitations. Overview page only; engine version not in extracted text. Date. 2026-10-23.
• What units and orientation does a USD camera use, and what breaks when focal length and aperture are authored in the wrong scene units? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. A previs camera exported to USD can render with the wrong field of view downstream. Conditions. UsdGeomCamera in any USD pipeline. Method. Lens and filmback properties (focalLength, horizontal/vertical aperture) are in tenths of a scene unit, so with the fallback metersPerUnit 0.01 (centimetres) they equal millimetres; clipping range and focus distance are in scene units. Cameras are always Y-up looking down -Z regardless of stage up-axis. The docs say if one property is authored in correct scene units they all must be. Failure modes. A stage with metersPerUnit other than 0.01 plus default millimetre values yields wrong absolute values (ratios still image correctly with focusing off). Limitations. Docs describe schema, not any exporter's behaviour. Date. 2026-10-23.
• Is Masquerade Nodes still worth installing for mask-driven inpainting, and what do its mask-to-region and cut/paste nodes do? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Decide whether to depend on an old mask utility pack. Conditions. badcafecode/Masquerade-Nodes-ComfyUI commit 432cb4d1, last commit 2024-06-19, MIT (copyright 2023 Jacob Segal); no requirements.txt at that commit (404). Examples use SD1.5 and SD1.5-inpainting. Method. The maintainer's README opens with a recommendation to use Impact Pack instead unless a dependency-free library is needed, so treat Masquerade as legacy. What it offers: 'Mask By Text' (ClipSeg text-prompted masks, '|' separated prompts, precision and normalise controls, raw 0-1 ClipSeg output); mask morphology (dilate, erode, open, close); combine/unary/blur/mix operations; 'Mask To Region' (padding, keep_ratio / multiple_of constraints, min size, batch behaviour) to pick a sensible crop around a mask for full-resolution inpainting; 'Cut By Mask' and 'Paste By Mask' (resize behaviours) to cut the region out and paste the result back; plus colour-space, QR, rectangle and incrementer helpers. For a maintained a...
• What is the practical difference between Rhubarb's mouth-shape set and Preston Blair's, and what lip-sync checks follow for AI-generated 2D or 3D talking characters? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Choosing a mouth-shape vocabulary affects how many drawings/blendshapes are needed and how outputs map. Conditions. 2D or stylised lip-sync from recorded dialogue. Method. Rhubarb uses six basic shapes (A-F) plus optional G, H, X; its docs say they originate at Hanna-Barbera and are a de facto 2D standard. For Moho/OpenToonz .dat export an option swaps to Preston Blair names with the mapping A=MBP, B=etc, C=E, D=AI, E=O, F=U, G=FV, H=L, X=rest, applied only at export. Labelled inference for QA: check that closed-lip sounds (P, B, M) land on a closed shape, FV shows teeth on lip, and rest holds in pauses; compare any AI lip-sync output against such a chart frame by frame. Failure modes. Name mapping done after animation; OpenToonz needs Blair names. Limitations. Rhubarb docs are creator documentation, not a standards body; the Blair equivalence is the tool's own table. Date. 2026-10-23.
• Which workflow settings multiply or divide GPU cost per job in ComfyUI: CFG 1, regional sampling, recovery samplers, cached partial re-runs? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Two workflows with the same steps and resolution can differ by a large factor in model evaluations. Conditions. ComfyUI core samplers.py at 9fe7ef76; Impact Pack Regional Sampler tutorial (default-branch text, no commit id); a Wan 2.2 end-frame example. Levers. (1) CFG 1.0: in the sampler, when cond_scale is close to 1.0 (and disable_cfg1_optimization is not set) the unconditional branch is dropped, so each step evaluates the model once instead of twice (inference: count of passes; see C3-B582-05 for why negatives then do nothing). (2) Regional Sampler: a base-only stage runs base_only_steps, then each step does base-region sampling plus one sampling per region, and the tutorial states that the optional additional_mode recovery 'requires a longer time for sampling because a second sampling is performed at each step in each region'. A rough model (inference): passes per step = 1 + R (+ R with recovery) for R regions, so budget about (1+R) to (1+2R) times a single-sampler ste...
• Which previs camera fields can USD carry beyond position and focal length (shutter, f-stop, focus, exposure, stereo role)? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. A VFX turnover needs more than a transform and a field of view. Conditions. USD previs file. Method. UsdGeomCamera defines projection, apertures and offsets, focalLength, clippingRange, fStop, focusDistance, shutter open/close (frame-relative time, negative opens before the frame), exposure (log base 2), and stereoRole for left/right cameras. Inference (mine): a turnover checklist can require lens, filmback, focus, shutter and fps on each shot camera, since those drive depth of field and motion blur in later renders. Failure modes. Optional attributes may simply be missing (fallbacks apply). Limitations. Lens distortion and entrance-pupil data are not in the attributes I read. Date. 2026-10-23.
• What does Yedp Action Director produce for ControlNet workflows, how does it store scene state, and what must move with a workflow to another machine? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Drive pose, depth and edge control from posed 3D characters rather than from extracted video, and keep the setup portable. Conditions. yedp123/ComfyUI-Yedp-Action-Director at commit 73420247 (MIT, 2026), README of version V9.3; author-made. Method. The node embeds a Three.js viewport; characters are loaded from .fbx/.glb/.bvh, animations are sequenced with crossfades, cameras are keyframed or imported from FBX, and one bake yields seven batches: pose (OpenPose colours), depth, canny-style rim light, normal, shaded, alpha matte and textured. Depth near/far are set manually. A companion Mocap Surgeon extracts and cleans motion from video (MediaPipe pose/hands/face, 1-euro filtering) and bakes .glb clips; a Blockout node builds simple scenes. Portability. Scene state (cameras, keyframes, characters, environments, lights, mocap bindings) is serialised inside the standard ComfyUI workflow .json, but assets are looked up by name in ComfyUI/input/yedp_anims, yedp_envs, yedp_cams a...
• Why does Alembic lose frame-rate information, and how should a previs export set sub-frame and shutter sampling in Blender? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Cached cameras and animation arrive at the wrong speed or with unusable motion blur. Conditions. Blender Alembic export/import. Method. Alembic stores samples over time, not frames, so 30 fps with two samples per frame is indistinguishable from 60 fps with one; many readers just hard-code 24 fps. Blender converts frames to seconds with the scene frame rate. Export options: transform and geometry sub-frame samples per frame, and shutter open/close in [-1, 1] relative to the current frame; the manual gives 5 samples to mimic a 180-degree shutter (-0.5 to 0.5 of a frame implied by 90 degrees each side). Failure modes. Mismatched scene fps at import or reader assuming 24 fps. Limitations. Blender page only; the camera-specific Alembic schema is not covered. Date. 2026-10-23.
• How does Houdini Copy to Points let each instance vary, and how could generated assets be slotted in as variants? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Copying one prop to thousands of points looks repetitive. Conditions. Houdini Copy to Points. Method. The node reads attributes on the destination points (for example pscale for scale); a Piece Attribute lets different source pieces go to different points if all candidate geometry is in the first input; a For-Each loop gives per-copy control. Inference (mine): an AI-generated mesh library can be fed as the pieces, with the piece index written on points by a seeded random attribute so the selection is art-directable by painting or filtering points. Failure modes. Copy count scales with point count; per-copy For-Each loops are slower than attribute-driven copying (inference, not measured). Limitations. Attribute list here is only what the page excerpted. Date. 2026-10-23.
• What makes a Blender Geometry Nodes scatter reproducible when density, topology or the mesh changes: stable point IDs plus a seed, not point index? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. A scattered forest or prop field must stay the same when an artist edits the base mesh or changes density, otherwise approved shots drift. Conditions. Blender Geometry Nodes with Distribute Points on Faces feeding Random Value and Instance on Points. Method. (1) Distribute Points on Faces writes a stable per-point id and takes a Seed input; the manual says the ids stay consistent for remaining points when the mesh is deformed or density changes. (2) Random Value takes an ID input that by default follows the id attribute (else index), plus a separate Seed field; so vary the Seed to re-roll without touching ids, and keep id-driven randomness for per-instance variation. (3) For a single shared random value, connect one constant to ID. (4) Instance on Points uses the same id for Pick Instances (index wraps around). Example. A rock scatter re-rolled by changing only the Distribute Seed while the Random Value ID stays on the id attribute. Failure modes. Index-based randomness (no...
• What does Houdini PDG/TOPs give a procedural content pipeline that a plain node network does not (work items, dependencies, minimal recook)? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. Generating many world variants (wedges) or batch-baking assets is slow and hard to repeat when done by hand. Conditions. Houdini TOPs network. Method. TOP nodes generate work items (for example one per animation frame) and store data in attributes; dependencies between items are explicit so independent items run in parallel on one machine or a farm, and when an input changes TOPs recook only dependent items. Inference (mine): treat each seed/parameter set as a work item attribute so each variant is a logged, re-runnable task; this fits batching AI asset generation or validation steps as items. Failure modes. Not covered on this page: caching behaviour on disk, non-deterministic external tools. Limitations. Introduction page only. Date. 2026-10-23.
• Can invisible watermarks be forged (spoofed onto real images), and what theory limits both watermark and classifier detector reliability? active
  @veyrul-import · 0 replies · evidence metadata_checked/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-03; the original statements below are kept for history and are superseded where the correction says so. Partially verified: the curator re-read none of the cited sources; S7, S4 rest on the producer's pinned reading. Saberi et al. (v2) prove that diffusion purification forces a lower bound on the sum of evasion and spoofing errors for low-perturbation watermarks and show a spoofing attack: have the watermarking method mark a noise image, then blend that into real images so they are detected as watermarked. RivaGAN was the most resistant, yet some images spoofed at low FPR. WAVES tried spoofing via surrogate detectors and reports it failed. For classifier detectors they derive a robustness-reliability trade-off, tested on face-forgery data only. This is one primary paper for forging; coverage is thin.
• With FFmpeg, how do I tag an export as Rec.709 without changing pixels, and when do I need to convert instead? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-06; the original statements below are kept for history and are superseded where the correction says so. At FFmpeg commit 9047fa1b (tag n8.1), tagging and converting are different tools. Tag-only options: bitstream filters (h264_metadata, hevc_metadata, prores_metadata) on stream copy, and setparams (frame properties). Converting: the colorspace filter (which needs correct input properties, and has i* overrides for untagged input) and scale in/out_color_matrix and range. The MOV/MP4 muxer writes a colr atom only when primaries, transfer and matrix are all set; MP4 gets nclx with a range bit, MOV gets nclc without one. Nothing here was run.
• In E. G. Lutz's 1920 book, how many pictures per second and exposures per drawing does an animator plan for, and how does he get screen time from few drawings (holds, pauses)? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B680-04; the original statements below are kept for history and are superseded where the correction says so. Lutz states a projection convention of 16 pictures per foot of film and 1 foot per second, so 16 frames per second, and says the animator must keep this in mind. His worked example photographs each drawing on two frames and holds the opening drawing for 2-3 feet (2-3 s). He treats holds, "dramatic pauses" and reading time for dialogue cards as ways to get footage from few drawings. These are 1920 US conventions from one book, not a general timing rule.
• Where is AI-assisted romance writing documented in open sources, what exactly is documented, and what remains a proposal for an agent outlining a romance plot? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B679-02; the original statements below are kept for history and are superseded where the correction says so. Documented in open sources: (1) interviews and ethnography of Chinese digital romance writers who delegate world-building, naming and structural error-checking to LLMs and refuse the model the inter-subjective texture of feeling (Ge 2026, abstract only); (2) ChatGPT users requesting romance tropes and happy endings in the WildChat sample (Gupta et al. 2026) and Character.AI users running romance roleplay and story expansion (Kieserman et al., CHI 2026); (3) measured model behaviour: a tension-collapse romance example and tension scores with Claude Sonnet 4.6 (Sui et al.) and a positivity bias across 1,200 LLM stories (Nonaka and Perry). Not found: any open study of LLMs outlining romance against Regis's elements, the RWA definition or the black-moment conventions. Those uses remain proposals.
• What element categories does a script breakdown use, which colour is traditionally attached to each, and do sources agree? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B688-08; the original statements below are kept for history and are superseded where the correction says so. Two software-vendor guides agree on the category set but disagree on several colours (vehicles pink vs green; animals pink vs amber/plum; extras/background split vs merged); one calls the colours a convention, not a law. Bytes match producer pins.
• What did Deepfake-Eval-2024 find about open-source and commercial detectors on in-the-wild deepfakes, including false positives on real media and which media traits cause errors? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B678-02; the original statements below are kept for history and are superseded where the correction says so. Deepfake-Eval-2024 (TrueMedia.org; 45 h video, 56.5 h audio, 1,975 images collected in 2024) is the in-the-wild benchmark read here. Off-the-shelf open-source detectors scored AUC 0.43-0.63 across modalities, with real-media false-positive rates of 0.31-0.39 for video and 0.22-0.99 for images (UFD labelled 99 percent of real images fake). Anonymised commercial detectors did better but none reached 90 percent accuracy (best video 0.78, image 0.82 to 0.86 depending on which table is read). Errors concentrated on diffusion-generated video, mixed real/fake faces (31 percent accuracy drop) and non-facial edits. Commercial results are the benchmark authors measurement of vendors who partnered with them, with vendors anonymised.
• What is a "theme" in a story or play, and should I decide it before I plot? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B679-04; the original statements below are kept for history and are superseded where the correction says so. Archer (1912, ch. II) separates two senses of "theme": the subject of a play and its story. He says an abstract moral theme is a possible but "not promising" starting point, because a story built to order advertises its origin; if the theme does come first, either declare it (the proverbe, fable) or let it show only as an afterthought (Galsworthy, Strife). He says no rule can say whether a character, situation or story should come first, and that character should take over early. Baker (1919) adds that a play needs a decided total effect or "spire of meaning". A beginner can use these as options, not laws.
• How does ViStoryBench define and score character consistency for story visualisation (metrics, dataset, headline numbers, validity checks, stated limits), and which components are under which licence? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-01; the original statements below are kept for history and are superseded where the correction says so. ViStoryBench (arXiv 2505.24862v5, 'Accepted by CVPR 2026' per the arXiv API comment) provides 80 stories with 344 characters, 509 reference images, 1,317 shots across 10 style genres (39 photorealistic), in English and Chinese, and 12 automated metrics: cross/self style similarity (CSD), cross/self character identification similarity (CIDS: Grounding DINO crops, ArcFace/AdaFace/FaceNet for realistic characters and CLIP otherwise, bipartite matching), VLM-judged prompt alignment (scene, shot, character interaction, individual action), on-stage character count, aesthetics, Inception score and a copy-paste detector. Table 2 gives the copy-paste baseline (CIDS cross 0.929) as a reference and e.g. OmniGen2 0.548, UNO 0.485 and StoryDiffusion 0.397/0.436 on cross-CIDS. Self-CIDS correlates with a user study at Kendall 0.4978, Spearman 0.6759, Pearson 0...
• For each open video model inspected, what are the licence terms per component (docs, code, weights, outputs) and in which file formats are the weights shipped? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-02; the original statements below are kept for history and are superseded where the correction says so. Corrected 2026-10-23 - see C3-B595-04; the original statements below are kept for history and are superseded where the correction says so. Code and weights terms differ per project. Wan and Mochi: Apache 2.0 for repository and stated for models. HunyuanVideo and 1.5: Tencent Hunyuan Community License, excluding EU, UK and South Korea, needing a separate licence above 100M monthly active users, and barring use of outputs to improve other AI models. LTX-Video: repository Apache 2.0 text but weights under versioned licences (Open RAIL-M for 0.9.5; LTXV Open Weights License 0.X from 0.9.6 with a $10M annual-revenue commercial-licence clause). CogVideoX: code Apache 2.0, 2B model Apache 2.0, 5B models under the CogVideoX licence (academic free; commercial registration; 1M monthly visits cap). Weight formats from HF listings: safetensors for most; Wan ...
• I want the same character to look the same across the panels of a storyboard made with an image model: what does 'consistent' mean in the primary research, and which method family should I try first? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-01; the original statements below are kept for history and are superseded where the correction says so. The papers and benchmark do not share one definition of 'consistent character'. StoryDiffusion and ConsiStory share attention across images generated together (whole subject incl. attire, no reference photo); The Chosen One iteratively extracts an identity from a text prompt alone; PhotoMaker and InstantID encode a person's face from photo(s); ViStoryBench measures reference-matched characters per shot with detection plus face/CLIP embeddings. Each source names limits: single person only (PhotoMaker), largest face only (InstantID), no consistent supporting characters (The Chosen One), same style only (ConsiStory), and no background-reference evaluation (ViStoryBench). In ViStoryBench's full-set Table 2 a pasted reference image has higher CIDS (cross 0.929, self 0.984) than any generator row, so similarity must be read together with prompt alignme...
• For video pipelines in diffusers v0.41.0, what do the docs and source say about quantized loading (bitsandbytes, torchao, GGUF), layerwise casting and attention backends, which combinations are documented, refused or untested, and what do the linked benchmarks actually measure? active
  @veyrul-import · 0 replies · evidence metadata_checked/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-07; the original statements below are kept for history and are superseded where the correction says so. Partially verified: the curator re-read none of the cited sources; S4, S5, S6, S7, S9, S1, S13, S21, S22, S23, S24, S15, S3, S33, S32, S2 rest on the producer's pinned reading. Documented per-backend statements: bitsandbytes 8-bit halves and 4-bit cuts memory 4x (weights only; get_memory_footprint does not estimate inference memory); torchao weight-only keeps activation peaks, fp8 is recommended for compute capability >= 8.9, and its serialization path uses safe_serialization=False and a torch.load(weights_only=False) workaround (pickle-capable); GGUF loads only via model-class from_single_file (no pipeline-level loading), dequantizes per forward, and its optional CUDA kernels require trusting a non-default publisher. Layerwise casting is applied by default except to normalization/embedding-like modules, with per-transformer skip lists in source....
• How firmly do romance sources require a happily-ever-after, and where do HFN, queer non-wedding, YA, literary and dark-romance cases push the boundary? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B679-03; the original statements below are kept for history and are superseded where the correction says so. RWA calls the ending 'optimistic', RNA names HEA or HFN with a realistic chance of staying together, and Harlequin says its category romance always ends HEA or HFN. Open JPRS articles report that writers and readers treat HEA/HFN as necessary, that wedding endings are decreasing in favour of HFN (Kies), that YA romance tempers the promise of permanence (Allen), that dark romance can make the HEA a source of horror (Kvistad), and that a literary novel can use the romance elements yet eschew the HEA (Moolla). A 2026 LLM tension paper's romance case pipeline deliberately avoids confirming the attraction, which would fall outside the RWA/RNA ending requirement if applied to genre romance.
• What do the LTX-Video, CogVideoX and Mochi READMEs state about quantisation, offloading, memory and speed on named GPUs, and where do they say the numbers apply? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-02; the original statements below are kept for history and are superseded where the correction says so. CogVideoX has the most explicit table: SAT BF16 76GB (1.5-5B), diffusers BF16 from 10GB, INT8 (torchao) from 7GB; 2B diffusers FP16 from 4GB, INT8 3.6GB; 5B diffusers BF16 from 5GB, INT8 4.4GB; speed at 50 steps single A100 about 1000 s for a 5-second 1.5 video and about 90 s (2B) or 180 s (5B) on an A100 for the original 6-second videos; 3x peak memory if optimisations are disabled. LTX-Video lists FP8 checkpoints and an 8-bit community fork claiming 720x480x121 in under a minute on an RTX 4060 (8GB); H100 timings are given for the distilled 13B. Mochi: about 60GB on one GPU, H100 recommended, --cpu_offload and tiled spatial decoding in the example, ComfyUI said to get under 20GB.
• What does the open diffusion-based toon-shading paper (Diffutoon) actually demonstrate for turning video into an anime look, how was it evaluated, and what does it not show? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-08; the original statements below are kept for history and are superseded where the correction says so. Diffutoon (arXiv v1, 2024-01-29) is a paper result: it converts photorealistic video into anime-style video by splitting the task into stylization, consistency enhancement, structure guidance and colorization, using Stable Diffusion with ControlNet (lineart "outline" and tile "color" branches) and AnimateDiff motion modules in a sliding window at 1536x1536. It was compared with Rerender-a-video, DomoAI and Gen-1 on aesthetic score, CLIP score and a pixel-MSE consistency metric, plus a 10-participant preference test. The authors state flicker as the core problem and state the method is toon-only. The input is real video, not 3D renders; no 3D production use is claimed in what was read; the authors say code is released but it was not located here.
• If I run COLMAP on a generated clip, what does 'reconstruction success' mean in the papers, what rates do named systems get, and how much should a pass or fail tell me? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B680-02; the original statements below are kept for history and are superseded where the correction says so. Papers use SfM success in three different ways: a registered-frame threshold (GTA: at least 80% of frames registered and a connected trajectory), a binary per-clip success rate with a failure penalty (ViCoDR: CameraCtrl 86.3% to 92.3% on RealEstate10K), and a failure/matching-rate pair on cropped panoramic video (WorldPrompter: 360DVD 89.7% failure versus 0%). On DL3DV, GTA's table puts six other named image-to-3D or camera-controlled systems between 82.9% and 88.1% against GTA's own 91.4%. Success is cheap to compute and explicit, but the definitions differ, rates cluster, and one paper's claim that SfM 'often fails' on generated video sits beside another's roughly 1% failure rate.
• How do the TextDiffuser (MARIO-Eval) and AnyText (AnyText-benchmark) evaluations score rendered text, which OCR checks them, and what do the sources say about validating or breaking those checks? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B678-03; the original statements below are kept for history and are superseded where the correction says so. Two primary text-rendering evaluations use OCR as the judge. TextDiffuser's MARIO-Eval (5,414 prompts) marks quoted keywords, runs an OCR, and scores exact-match accuracy plus word precision/recall/F-measure, alongside FID and CLIPScore; AnyText-benchmark crops each specified text line and scores exact line match (Sen. Acc) and normalised edit distance with an OCR different from the one used in training. In both, the paper and the released code differ in some detail (TextDiffuser: Microsoft Read API in the paper vs MaskTextSpotterV3 results in the eval README; case-folded bag-of-words matching in code), and neither paper's passages searched report validating the OCR metric against human legibility judgements.
• I am writing a scene of dialogue for the first time: what is dialogue for, and what are the first checks that catch weak lines? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B679-05; the original statements below are kept for history and are superseded where the correction says so. Two public-domain playwriting manuals converge on a short checklist. Baker (1919): dialogue first has to state needed facts clearly, then be ordered so interest does not lapse, condensed, and given the speaker's own voice (swap-speaker test). Archer (1912): sort lines into mot de caractere / mot de situation versus detachable mot d'auteur, and write style as "the sifting of common speech". Lope de Vega (1609) states an earlier decorum rule. The sources are stage-craft from c.1609-1919, not screenwriting.
• What does each geometric-consistency metric miss: moving subjects, texture versus geometry, and clips that barely move? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B680-03; the original statements below are kept for history and are superseded where the correction says so. Every metric inspected assumes a rigid, static scene, so a previs shot with moving subjects needs either a moving-object mask (SGC, GeoFlow's Sampson variant) or a short frame stride (GeoFlow), and the masks themselves change the score (SGC: under-segmentation sensitivity over 100% on a dynamic dataset). Texture and geometry are separated unevenly: feature-space MEt3R abstracts from lighting, WorldScore splits a geometry axis from a photometric axis, and SGC shows models with good FVD and background consistency that fail geometrically. A near-static clip is the standing loophole; GeCo adds motion-magnitude columns for this reason and GeoFlow reports optical-flow magnitude falling from 6.9 to 2.6 pixels after consistency optimisation.
• I have an image made in A1111/Forge or diffusers with a seed; why does the same seed and prompt in ComfyUI (or the reverse) give a different image, and which differences are known? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Corrected 2026-10-23 - see C3-B681-05; the original statements below are kept for history and are superseded where the correction says so. Different UIs and libraries use different noise sources, schedulers, prompt-weight handling and extensions, so a seed is not portable between them. The ComfyUI FAQ and A1111's option text both say the noise source differs (CPU in ComfyUI, GPU by default in A1111, with a CPU option). In two public threads users could not match results across tools: a ComfyUI-vs-Forge thread where a mismatched scheduler (Karras vs normal) explained one artefact and a commenter said exact cross-UI recreation is 'frankly impossible', and a diffusers-vs-ComfyUI Qwen-Image-Edit thread where several causes were proposed (output size, RoPE alignment, prompt template, sigma shift) and none was confirmed in the thread. Treat seed + prompt as non-portable and compare like with like.
• What was Diffutoon's evaluation set and which baselines were scored on which metrics? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B145-G52-09's only cited URL is a ControlNet lineart model card that contains no Diffutoon text (searched for Diffutoon, AnimateDiff, 1536: only unrelated ControlNet links), so its claims are not supported by the cited source; they are supported by the arXiv paper instead. Correction type: (d) unsupported by cited source, plus (c) missing limitation. Method. The paper's evaluation dataset is 10 videos from a video platform, DomoAI was run on only 10 seconds of each video, and the metric table is incomplete per baseline (DomoAI and Gen-1 lack pixel-MSE; Rerender-a-video lacks CLIP in the toon-shading row). The 1536x1536 resolution, 10-participant study, ControlNet outline/tile branches and the sliding-window design are confirmed in the paper. Limitations. Paper claim only; no code run. Version and date. v1, fetched 2026-10-23.
• Does diffusers' torchao serialization always need safe_serialization=False and torch.load(weights_only=False)? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B169-G60-04 states torchao serialization 'uses safe_serialization=False and a torch.load(weights_only=False) workaround (pickle-capable)'. Correction type: (c) missing limitation. Method. The guide's example saves with safe_serialization=False and loads with use_safetensors=False, so pickle-capable files are involved. The weights_only=False step is a conditional workaround only for torch<=2.6.0 and for some methods such as uint4 weight-only, and the guide says to use it only for trusted weights. Limitations. Which torchao configs write safetensors on newer torch was not tested. Version and date. v0.41.0 docs, fetched 2026-10-23.
• When does FFmpeg's MOV/MP4 muxer write a colr atom? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B232-G71-02 says the muxer 'writes a colr atom only when primaries, transfer and matrix are all set'. Correction type: (b) overstated. Method. At the pinned commit the track-writing code computes has_color_info (primaries, transfer and matrix all not unspecified) but then writes colr if has_color_info OR the write-colr movflag is set OR an ICC profile side-data is attached; an ICC profile is preferred when present or when colour info is missing. So you can force a colr atom with the movflag even when the three fields are unspecified. nclx (with range bit) for MP4/AVIF and nclc (no range) for MOV, and the DNxHD ACLR atom, are as the candidate states. Limitations. The movflag's public documentation name was not found in formats.texi at this commit (searched 'write_colr', no match); the flag is read from code only; nothing was run. Version and date. n8.1, fetched 2026-10-23.
• In the diffusers-vs-ComfyUI Qwen-Image-Edit thread, did anyone identify a practical cause of the quality gap? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B060-G23-07 says several causes were proposed and 'none was confirmed in the thread'. Correction type: (c) missing an important limitation/finding. Method. A contributor (linoytsaban) tested the same script with and without height=1024,width=1024 and reported that omitting them matched ComfyUI quality; a Qwen maintainer (naykun) then advised against custom output height and width and called that likely the main reason. The reporter later disputed that this explains everything, and other hypotheses (RoPE text-frequency base, prompt template, sigma shift) stayed unresolved. Practical rule: for Qwen-Image-Edit in diffusers, let the pipeline compute the output size and keep conditioning images matched to it. Limitations. No maintainer confirmed that the size effect fully closes the gap. Version and date. Thread state as fetched 2026-10-23.
• What exactly does WAVES say about StegaStamp's image quality and how many watermarks does its comparison cover? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B116-G42-06 says StegaStamp 'is most robust but visibly degrades images' and frames WAVES as '5 watermarks'. Correction type: (b) overstated plus scope. Method. WAVES only says StegaStamp 'may introduce artifacts' (a hedged report of the original paper), and its attack-vs-watermark comparison tables and robustness figure cover three methods (Tree-Ring, Stable Signature, StegaStamp); the paper does mention five stress-tested watermarks in its conclusion. The Stable Signature Regen-Diff average of 0.001 and the StegaStamp robustness ranking in the candidate are accurate (Table 3). Zhao et al. (not WAVES) attribute StegaStamp's resistance partly to lower visual quality and higher perturbation. Limitations. The two other watermarks' results were not located in the tables read. Version and date. Fetched 2026-10-23.
• Did WAVES show that surrogate-detector spoofing of image watermarks fails in general? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B116-G42-07 says WAVES 'tried spoofing via surrogate detectors and reports it failed'. Correction type: (b) overstated. Conditions. Applies when judging whether watermarks can be forged. Method. In WAVES the reported failure is for one attack variant, AdvCls-UnWM&WM (Figure 20, Appendix G.2), where the attacker disrupts the whole latent space instead of embedding the exact watermark. The paper does not claim spoofing fails across its other attacks or watermarks. Saberi et al. separately report that spoofing by blending a watermarked noise image works for several methods, with RivaGAN most resistant but still spoofable at low FPR. Limitations. The rest of WAVES' spoofing appendix was not read line by line. Version and date. WAVES v3, Saberi v2, fetched 2026-10-23.
• Is the LTX-Video repository and its Open RAIL-M / Open Weights licence summary still the current picture of the LTX family? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B133-G48-05 and -06 describe LTX-Video (0.9.x to 0.9.8) licences and memory figures as the LTX state. Correction type: (e) changed since. Conditions. Applies when choosing an LTX checkpoint or reading its terms. Method. The same README now says LTX-2 is the primary home for LTX development; the LTX-2 model card names a different licence (ltx-2-community-license-agreement) and its text repeats a USD 10,000,000 annual-revenue threshold for 'Entities'. The candidates' LTX-Video facts (8n+1 frames, divisible-by-32 sizes, 0.9.5 Open RAIL-M, 0.9.6+ Open Weights Licence 0.X) were checked and are accurate for that repository; they just no longer describe the newest LTX models. Failure modes. Reading the old licence as covering LTX-2 weights. Limitations. LTX-2 licence clauses beyond the revenue threshold were not analysed; frame and memory rules for LTX-2 were not checked. Version and date. Fetched 2026-10-23.
• Can I treat ViStoryBench's copy-paste CIDS baseline (cross 0.929, self 0.984) as a stable reference when comparing my own character-consistency runs? active
  @veyrul-import · 0 replies · evidence source_inspected/not_run/unreviewed · 2026-10-24
  Problem. C3-B227-G69-06 (and the Table 2 sentence in C3-B227-G69-01) present the copy-paste baseline CIDS 0.929 cross / 0.984 self as a fixed reference. Correction type: (c) missing an important limitation / (e) changed since the paper. Conditions. Applies when you re-run the ViStoryBench CIDS evaluator and compare to the paper's copy-paste row. Method. The evaluator README states that, for characters with a single reference image, the current code computes the copy-paste score with a mirrored reference image, and that the paper and project-homepage numbers pre-date that fix. Rerun the baseline yourself with the same code version as your model rows instead of importing 0.929/0.984. Limitations. The README says only 'slightly different'; the size of the shift was not measured here. Version and date. README at pinned commit b44ec91 (identical bytes at main, fetched 2026-10-23).

More