Files
mtgodot-poc/test/shear-bones-survey.md
T
shenandClaude Sonnet 5 47baf6c0c6 Metin2 game client (P0–P11) + mobile asset pipeline
Networked client on the existing Godot 4.7 + libgr2 renderer:
- net: m2dev wire protocol (libsodium KX + XChaCha20), auth/select/game
  phases, EntityStore world model, ~all GC/CG headers. char create/delete,
  private shop / mall / cube, SHOP_GC_START_EX, guild, party (+ CG_PARTY_SET_STATE),
  quests, dragon soul, refine, safebox, exchange.
- UI: in-game windows migrated 1:1 from the reference uiscript/root .py —
  char status (/stat), inventory+equipment, select-item ([SELECT_ITEM] quest
  token), system-option + game-option + ESC system menu, private-shop 39-grid,
  party info board, shop tabs, atlas, minimap, quickbar, chat, …
- EterGrnLib polish: GR2 material blend/two-sided, LOD crossfade, motion-event
  dispatch, contact shadow, ray-AABB picking, weapon grip pre-transform.

Portable asset IO (A1) — all extension/libgr2/formats/mtproto reads routed
through godot::FileAccess (res:// PCK works on iOS/Android); standalone-lib
*_path() kept for the non-Godot CTests. AssetResolver + PropertyRegistry
switched to a baked index (bake_asset_index.gd) instead of std::filesystem.

Mobile builds: build-{android,ios}.sh, export-android.sh, pack-assets.sh,
gen-debug-keystore.sh. Assets ship as a zip mounted at runtime by
project/asset_pack.gd (adb push now; HTTP download is a drop-in later).

ctest 10/10, 34 GDScript suites, macOS/iOS/Android all build.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013EJxkHiNKS4kybHS3XKyAJ
2026-08-31 20:02:12 +09:00

2.1 KiB
Raw Blame History

Shear-bone survey (Godot GPU-skinning viability)

tools/anim_probe <model.gr2> <anim.gr2> — for every bone, the parent-relative local that would be fed to Skeleton3D::set_bone_pose, its shear, and the error of a T·R(quat)·S round-trip (= what Godot's GPU vertex skinning keeps after it orthonormalizes the bone matrix before upload). High round-trip error = that bone collapses on the GPU (Skeleton3D) path once an animation moves it.

Probed with pc/warrior/action/dance_1.gr2 (retarget by bone name; unmatched bones sit at bind, so their numbers reflect the model's rest rig).

model shear bones worst TRS err worst minDet GPU-skin viable?
warrior_cheongrin 7 (bone_front_01..04, Bone_shoulder_01..04) 0.94 0.065 (6.5% squash) collapses (dance_1 animates these)
assassin 0
sura_lord 16 (Bone05/06, Bone14..27) 1.52 0.0024 (0.24% squash) severe
shaman_lord 0

Takeaways

  • The shear/non-uniform-scale is baked into the rest rig of these cloth/plate helper bones (metrics are constant across the animation clock). Static bind pose survives because skin_matrix ≈ I there; any animation that moves a shear bone breaks it on GPU skinning.
  • sura is unusable on Godot's built-in GPU skinning — 16 bones, one squashed to 0.24%. It's a core class, so this isn't a corner case.
  • assassin / shaman are clean — GPU skinning would be fine for them.
  • Done (2026-08-29): the custom skinning vertex shader is implemented — m2_material.cpp SRC_SKIN does LBS with the full per-bone 4×3 matrices from an RGBAF texture (Metin2Model::gpu_skin fills it each frame), no Skeleton3D, no orthonormalization. MTGODOT_GPUSKIN=1 now routes here (the old broken set_bone_pose path is deleted). warrior dance_1 GPU render == CPU render (test/compare-shots/*.gpu.png). CPU LBS stays the default; GPU is for crowds (Phase 2 bench).

Re-run for a full race sweep (incl. monsters, NPCs, mounts) before Phase 2 to size the shader work and catch any rig that also needs >4 bone influences.