[kashgummaraju]

Summary

Research only. No code, asset or plan file changes.

This adds a report, reports/GeckoLib creature animation survey.md, backed by four notes in research_notes/GeckoLib creature animation survey/. Together they survey 13 creature mods, GeckoLib 4.8 and 5.5, Blockbench and AzureLib, to find ways to improve current and new dinosaur animations.

Findings about Nublar's own files, each checked in this session:

• 179 rotation values in 10 clips are between 10^6 and 10^10 degrees. The clips are Brachiosaurus drink, call, eat1 and roar; Dilophosaurus eat1 and head_shake; Gallimimus swim; and Velociraptor dead, roar and scratch.
• The clips are baked densely: 110,620 keys, and a 5.6 MB T. rex file. Most channels hold only the rest pose. A reduction removes about 40 % of keys at 0.1° error.
• IdleSway misses the last tail bone of every species (both of the Brachiosaurus's last two). It also misses the Brachiosaurus head, which is named head1.
• Head turning is off for every species: all models use the one-argument DinosaurGeoModel constructor.
• The Brachiosaurus sound keyframes have no handler, so they never play.

The report then ranks techniques in four groups: asset fixes, small GeckoLib 4 code (tail turn lag, clamped neck look-at, footstep cues, layer weights, per-clip speed and hit data), medium work, and items to hold for the GeckoLib 5 port.

Work package

None. This is standalone research for the lead. It informs B09 (appearance) and B24 (art), and the GeckoLib 5 port.

Acceptance

No acceptance lines. The PR adds documents only.

Runs performed

None. The PR changes no code. Measurements were read-only Python scripts run on the animation and geo JSON files.

Decisions for the lead

• Question: should a tools/ script rewrite the six animation files to fix the bad keys, strip rest-pose data and reduce keyframes?
  • In-game example: possible bone flicker in Velociraptor roar.
  • Choice: none. The decision is left to the lead.
• Question: should the out-of-range keys become 0?
  • In-game example: each affected clip needs a visual check in Blockbench.
  • Choice: none. 0 looks likely, but this is unconfirmed.
• Question: should head look and tail motion be procedural deltas, or use dedicated control bones (geo edits)?
  • In-game example: a T. rex looking at a player during idle.
  • Choice: none.

---

Generated by Claude Code

[graphify-labs]

Graphify reviewed this change.

No issues found by static checks — no coupling regressions or blocking issues in the code graph. That is not the same as safe to merge: see what was not checked.

Not checked: tests were not run; no formal proof of the changed code.

---

Graphify review — findings

Adds a research report for the project lead that ranks what to do about Nublar's dinosaur animations: fix the clip data first, then add tail lag and head look. It changes no code or assets. It lists nine measured or source-verified faults (e.g. 179 out-of-range rotation keys, dense 24–25 keys/s baked clips, IdleSway skipping the last tail bones and the Brachiosaurus head1, head turning disabled in every DinosaurGeoModel, and unhandled sound keyframes). It compares what 13 surveyed mods do about the same problems, with licence notes. It then orders the fixes into asset-only cleanups (A1–A4), small GeckoLib 4 code (C1–C8), medium work (M1–M5) and items that wait for the GeckoLib 5 port, and ends with four open decisions for the lead, including whether to approve an asset-rewriting tool.

Review partial — this diff was larger than one review pass covers, so later files were not reviewed; some findings may be missing.

Analysis details — impact, health, verification

Impact & health

Graphify review

Impact — 164 functions depend on the 164 functions this change touches.

Health — grade A; no new coupling hotspots.

Verification — 164 functions in the blast radius were not formally verified this run (proofs are advisory here).

Gate & verification

graphify gate

PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.

Advisory (not blocking):

• verification_scope: 164 function(s) in the blast radius were not formally verified this run