mirror of
https://github.com/Kelsidavis/WoWee.git
synced 2026-05-08 01:53:52 +00:00
0cdefb2611
Drills into a .glb's byte composition. Pairs with --info-glb (counts)
and --info-glb-tree (structure) — three lenses on the same file:
wowee_editor --info-glb-bytes custom_zones/Z/Z.glb
GLB bytes: custom_zones/Z/Z.glb
total: 891736 bytes (0.85 MB)
Sections:
header : 12 bytes 0.00%
JSON hdr : 8 bytes 0.00%
JSON : 828 bytes 0.09%
BIN hdr : 8 bytes 0.00%
BIN : 890880 bytes 99.90%
BufferViews:
idx target bytes MB share-of-bin
0 vertex 248832 0.24 27.93%
1 vertex 248832 0.24 27.93%
2 index 393216 0.38 44.14%
By attribute:
INDICES 393216 bytes (44.14% of BIN)
NORMAL 248832 bytes (27.93% of BIN)
POSITION 248832 bytes (27.93% of BIN)
Three breakdowns:
- Section costs (header + JSON chunk + BIN chunk + their headers)
- Per-bufferView with target hints (vertex / index / other)
- Per-attribute (POSITION/NORMAL/TEXCOORD_0/INDICES/etc.) bucketed
by walking accessors referenced from primitives
Catches asymmetric BIN allocation ('why is INDICES 44% of my .glb?')
and helps tune vertex layout decisions (drop normals if synthesized,
quantize positions if precision allows). Verified on a single-tile
zone bake: header+JSON is 0.1% of total, BIN is 99.9%, attribute
breakdown shows POSITION/NORMAL each 28% and INDICES 44%.
|
||
|---|---|---|
| .. | ||
| asset_extract | ||
| auth_login_probe | ||
| auth_probe | ||
| blp_convert | ||
| dbc_to_csv | ||
| editor | ||
| asset_pipeline_gui.py | ||
| backup_assets.sh | ||
| diff_classic_turtle_opcodes.py | ||
| gen_opcode_registry.py | ||
| generate_ffx_sdk_vk_permutations.sh | ||
| m2_viewer.py | ||
| opcode_map_utils.py | ||
| validate_opcode_maps.py | ||