Animation pack evaluation: Protofactor Injured Animset
Technical verdict: Usable with conditions
Evaluation completeness: partial — all source files and declared checks were reviewed; engine playback and visual quality remain untested.
Confidence: medium
Evaluation date: 2026-09-16
Current evaluator: AnimSmith 0.14.0
Report format: 3
Detailed evidence: Protofactor Injured evidence
Technical decision
Prototype decision: Try a separate style-local speed state for each injury style. Configure speed and healthy-state transitions; foot contacts and character fit remain unknown.
Camera scope: Full-body humanoid animations; appearance on your target character and camera is untested. Dedicated first-person arms/viewmodel use is unverified.
This animset is one of eight locally evaluated parts of Ultimate Animation Collection. The local asset revision is unknown; the collection report separates evaluated files from the dated vendor listing.
Content in evaluated inventory: Seven idle/walk/run injury styles. Each A–G style has an evaluated three-file candidate; filenames do not establish a severity order or recovery path. These are identified motions; gameplay behavior has not been approved.
Use as seven separate injury-style locomotion candidates. Each proposed style has idle/walk/run members; there is no evidence that A–G are interchangeable directions or severity levels. Keep the selected style coherent, let the kinematic controller own motion, and check walk/run contact timing and transitions to healthy locomotion.
observed-animsmith: the official 0.14.0 evaluator inspected and measured all 72 FBXs; 28/70 declared-contract files pass and 42 fail per output format. The findings below separate source problems, project settings and untested gameplay. Transformed outputs still need gameplay and visual checks.
Capability coverage
Content present
Filename-classified A–G styles each supply idle, walk and run candidates. Each style is a distinct three-member speed-state hypothesis.
Content gaps and unknowns
The letters do not classify severity or an ordered progression. Recovery and healthy-state handoffs have no accepted set classification here; this does not establish absence from the wider source.
Evaluation still needed
Test style choice, speed calibration, planted-foot behavior, cross-style and healthy-state transitions on the target character. Engine, retarget and visual acceptance remain open.
Runtime sets and authored motion
Use each row as a separate blend or sequence proposal. The linked appendix names its exact clips, timings and measurements; controller playback and contact quality still need testing.
| Set | Controller use | Adoption decision | Exact members |
|---|---|---|---|
| Injury style A speed | Style-local idle/walk/run; controller travel | Style-local prototype. RunInjuredA has loop seam findings; review true cycles and foot contacts before speed blending (IN-LOOP). | 3 exact members |
| Injury style B speed | Style-local idle/walk/run; controller travel | Style-local prototype. Selected injury loops mostly fail their declared checks; review this style’s loop intent and contacts before speed blending. | 3 exact members |
| Injury style C speed | Style-local idle/walk/run; controller travel | Style-local prototype. Selected injury loops mostly fail their declared checks; review this style’s loop intent and contacts before speed blending. | 3 exact members |
| Injury style D speed | Style-local idle/walk/run; controller travel | Style-local prototype. Selected injury loops mostly fail their declared checks; review this style’s loop intent and contacts before speed blending. | 3 exact members |
| Injury style E speed | Style-local idle/walk/run; controller travel | Style-local prototype. Selected injury loops mostly fail their declared checks; review this style’s loop intent and contacts before speed blending. | 3 exact members |
| Injury style F speed | Style-local idle/walk/run; controller travel | Style-local prototype. Selected injury loops mostly fail their declared checks; review this style’s loop intent and contacts before speed blending. | 3 exact members |
| Injury style G speed | Style-local idle/walk/run; controller travel | Style-local prototype. Selected injury loops mostly fail their declared checks; review this style’s loop intent and contacts before speed blending. | 3 exact members |
Integration recipe
- Members/topology: Create one idle/walk/run tree per chosen style A–G. Treat style changes as explicit state transitions; do not distribute injury letters around a direction tree.
- Timing/synchronization: Establish intended cyclic playback and support-contact timing before choosing synchronization. Walks measure 1.333 s; runs measure 0.700 or 0.800 s. Match support-foot timing between moving clips; an almost-stationary idle gait phase is not a useful contact anchor.
Humanoid@RunInjuredB.fbxpasses a non-looping declaration, which does not establish it as a clean repeating run. - State ownership: Selected in-place roots are stationary. Tune speed thresholds to stride/contact behavior, not their measured zero root speed. Keep root-motion alternatives under a separate movement policy.
- Composition constraints: Preserve the limp and compensating torso pose. Upper-body masks can erase the injury expression; validate any weapon layer separately.
- Acceptance gate: Test start/stop, velocity changes, healthy-to-injured transitions and interruptions on the target character.
Technical issue register
Severity describes impact as delivered for the stated use. Residual status is explicit; untested risks are not confirmed artist defects.
| ID | Severity | Problem and impact | Primary owner | Current action | Future AnimSmith potential | Evidence/status |
|---|---|---|---|---|---|---|
| IN-LOOP | major | Twenty of the 21 selected idle/walk/run members fail their declared contracts. Humanoid@RunInjuredA.fbx has loop-closure/seam findings; repeated playback and speed blending need review. Readiness guidance. | artist-author | Correct intended loop endpoints and support-foot timing. Fourteen gait-anchor candidates were generated; thirteen still fail declared lint. Residual adoption remains unresolved; no candidates are promoted and thirteen retain contract findings. | Current anchoring changes phase; it is not proof of a clean seam or preserved limp. | Exact selected files, Take 001, current source/output lint; observed-animsmith. |
| IN-STYLE | minor | Mixing lettered styles as if they were directions or scalar severity levels has no established semantic basis. Readiness guidance. | engine-config | Choose a style and keep idle/walk/run coherent; explicitly review any style transition. Residual: project decision, not proven vendor defect. | Explicit set declarations make the choice reproducible. | Names and timing observed; artistic relationship not-evaluated. |
| IN-TRACK | minor | Dense constant tracks may add storage without useful motion. Readiness guidance. | animsmith-current-declared | Pruning reduced tracks in WalkInjuredA, but its loop check still fails. Compare required channels and engine playback before using that output. Unresolved. | Current pruning exists; performance gain remains unmeasured. | Fresh prune output and diff; observed-animsmith. |
Engine status
| Runtime | Evidence level | Technical result | Remaining gate |
|---|---|---|---|
| Unity unspecified | not-evaluated | No current import or playback. | Blend, mask, visual, and build tests. |
| Unreal Engine unspecified | not-evaluated | No current import or playback. | Retarget, graph, and build tests. |
| Godot unspecified | not-evaluated | No current conversion or playback. | Conversion/import and graph tests. |
| Bevy unspecified | not-evaluated | No current handoff or runtime test. | glTF handoff and runtime test. |
Fit and limitations
Combine with Basic Locomotion as a healthy/injured state switch, with speed and stance changes handled deliberately. Combining with weapon packs requires checking whether a grip/aim layer conflicts with the injured torso and supported limb. No cross-pack contact, masking or artistic match is approved.
These are proposed integration decisions (inferred), with target-controller, contact and artistic acceptance still not-evaluated.
Changes between AnimSmith versions
AnimSmith 0.14.0 — reran 72 input baselines, 70 per-file declared contracts and 15 remediation trials using the official release. Added current exact-member generic scenarios and developer/artist actions. AnimSmith 0.10.0 — superseded historical evidence; none of its generated outputs or engine results is relabelled as a fresh run.
Evidence status
Current evidence uses the official release and the canonical readiness ladder; commercial artifacts remain external.
Sources
- Protofactor, Animset: Injured — product context.
- AnimSmith, CLI reference and game-ready clips — evaluation boundaries.