phaneron/Kelsidavis-WoWee
1
0
Fork 0
mirror of https://github.com/Kelsidavis/WoWee.git synced 2026-05-07 01:23:52 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

feat(wom): add WOM3 multi-batch format for material-aware models

Browse source 
WOM1/WOM2 had a single mesh with one texture, which lost the multi-submesh
structure of complex M2 models (body+hair+eyes+armor each need different
textures and blend modes).

WOM3 adds a Batch array: each batch has indexStart/indexCount + a textureIndex
into texturePaths + blendMode + flags. Loader is fully backward compatible:
WOM1/WOM2 files still load, and WOM3 with no batches block falls back to a
single full-mesh batch. fromM2 now extracts batches with materials, and toM2
emits matching M2 batches so the renderer can draw them correctly.
This commit is contained in:
Kelsi 2026-05-06 01:07:00 -07:00
parent 00c078a9af
commit b736c6b2e1
3 changed files with 124 additions and 30 deletions
Download patch file Download diff file Expand all files Collapse all files

124
src/pipeline/wowee_model.cpp
View file

@ -11,6 +11,7 @@ namespace pipeline {
static constexpr uint32_t WOM_MAGIC = 0x314D4F57; // "WOM1"
static constexpr uint32_t WOM2_MAGIC = 0x324D4F57; // "WOM2"
static constexpr uint32_t WOM3_MAGIC = 0x334D4F57; // "WOM3"
bool WoweeModelLoader::exists(const std::string& basePath) {
return std::filesystem::exists(basePath + ".wom");
@ -25,9 +26,10 @@ WoweeModel WoweeModelLoader::load(const std::string& basePath) {
uint32_t magic;
f.read(reinterpret_cast<char*>(&magic), 4);
bool isV2 = (magic == WOM2_MAGIC);
if (magic != WOM_MAGIC && magic != WOM2_MAGIC) return model;
model.version = isV2 ? 2 : 1;
bool isV2 = (magic == WOM2_MAGIC || magic == WOM3_MAGIC);
bool isV3 = (magic == WOM3_MAGIC);
if (magic != WOM_MAGIC && magic != WOM2_MAGIC && magic != WOM3_MAGIC) return model;
model.version = isV3 ? 3 : (isV2 ? 2 : 1);
uint32_t vertCount, indexCount, texCount;
f.read(reinterpret_cast<char*>(&vertCount), 4);
@ -110,8 +112,25 @@ WoweeModel WoweeModelLoader::load(const std::string& basePath) {
}
}
LOG_INFO("WOM", (isV2 ? "2" : "1"), " loaded: ", basePath, " (", vertCount, " verts, ",
model.bones.size(), " bones, ", model.animations.size(), " anims)");
// WOM3: read batches (multi-material support)
if (isV3) {
uint32_t batchCount = 0;
if (f.read(reinterpret_cast<char*>(&batchCount), 4) && batchCount > 0 && batchCount <= 4096) {
model.batches.resize(batchCount);
for (uint32_t i = 0; i < batchCount; i++) {
auto& b = model.batches[i];
f.read(reinterpret_cast<char*>(&b.indexStart), 4);
f.read(reinterpret_cast<char*>(&b.indexCount), 4);
f.read(reinterpret_cast<char*>(&b.textureIndex), 4);
f.read(reinterpret_cast<char*>(&b.blendMode), 2);
f.read(reinterpret_cast<char*>(&b.flags), 2);
}
}
}
LOG_INFO("WOM", (isV3 ? "3" : (isV2 ? "2" : "1")), " loaded: ", basePath, " (",
vertCount, " verts, ", model.bones.size(), " bones, ",
model.animations.size(), " anims, ", model.batches.size(), " batches)");
return model;
}
@ -124,7 +143,13 @@ bool WoweeModelLoader::save(const WoweeModel& model, const std::string& basePath
if (!f) return false;
bool hasAnim = model.hasAnimation();
uint32_t magic = hasAnim ? WOM2_MAGIC : WOM_MAGIC;
bool hasBatches = model.hasBatches();
// WOM3 implies WOM2 layout (vertex format with bones), so we only emit
// WOM3 if the model also has animation data — pure-batch static meshes
// still go to WOM1/WOM2 with batches written via the WOM3 trailing block
// when present alongside animation. For static-only with batches, write
// as WOM3 anyway (decoder handles missing bones).
uint32_t magic = hasBatches ? WOM3_MAGIC : (hasAnim ? WOM2_MAGIC : WOM_MAGIC);
f.write(reinterpret_cast<const char*>(&magic), 4);
uint32_t vertCount = static_cast<uint32_t>(model.vertices.size());
@ -141,8 +166,8 @@ bool WoweeModelLoader::save(const WoweeModel& model, const std::string& basePath
f.write(reinterpret_cast<const char*>(&nameLen), 2);
f.write(model.name.data(), nameLen);
// WOM2 writes full vertex with bone data; WOM1 writes 32-byte vertex
if (hasAnim) {
// WOM2/WOM3 write full vertex with bone data; WOM1 writes 32-byte vertex
if (hasAnim || hasBatches) {
f.write(reinterpret_cast<const char*>(model.vertices.data()),
vertCount * sizeof(WoweeModel::Vertex));
} else {
@ -161,8 +186,8 @@ bool WoweeModelLoader::save(const WoweeModel& model, const std::string& basePath
f.write(path.data(), pathLen);
}
// WOM2: write bones and animations
if (hasAnim) {
// WOM2/WOM3: write bones and animations (always, even if empty for WOM3)
if (hasAnim || hasBatches) {
uint32_t boneCount = static_cast<uint32_t>(model.bones.size());
f.write(reinterpret_cast<const char*>(&boneCount), 4);
for (const auto& bone : model.bones) {
@ -194,8 +219,22 @@ bool WoweeModelLoader::save(const WoweeModel& model, const std::string& basePath
}
}
LOG_INFO("WOM", (hasAnim ? "2" : "1"), " saved: ", womPath, " (", vertCount, " verts, ",
model.bones.size(), " bones, ", model.animations.size(), " anims)");
// WOM3: write batches
if (hasBatches) {
uint32_t batchCount = static_cast<uint32_t>(model.batches.size());
f.write(reinterpret_cast<const char*>(&batchCount), 4);
for (const auto& b : model.batches) {
f.write(reinterpret_cast<const char*>(&b.indexStart), 4);
f.write(reinterpret_cast<const char*>(&b.indexCount), 4);
f.write(reinterpret_cast<const char*>(&b.textureIndex), 4);
f.write(reinterpret_cast<const char*>(&b.blendMode), 2);
f.write(reinterpret_cast<const char*>(&b.flags), 2);
}
}
LOG_INFO("WOM", (hasBatches ? "3" : (hasAnim ? "2" : "1")), " saved: ", womPath,
" (", vertCount, " verts, ", model.bones.size(), " bones, ",
model.animations.size(), " anims, ", model.batches.size(), " batches)");
return true;
}
@ -260,6 +299,25 @@ WoweeModel WoweeModelLoader::fromM2(const std::string& m2Path, AssetManager* am)
model.bones.push_back(wb);
}
// Convert batches with material/blend mode info (WOM3 feature).
// Each M2 batch maps to a WOM batch — preserves multi-submesh material structure.
for (const auto& mb : m2.batches) {
WoweeModel::Batch wb;
wb.indexStart = mb.indexStart;
wb.indexCount = mb.indexCount;
// Resolve textureLookup -> texture index
uint32_t lookupIdx = mb.textureIndex;
wb.textureIndex = (lookupIdx < m2.textureLookup.size())
? static_cast<uint32_t>(std::max<int16_t>(0, m2.textureLookup[lookupIdx]))
: 0;
if (mb.materialIndex < m2.materials.size()) {
const auto& mat = m2.materials[mb.materialIndex];
wb.blendMode = mat.blendMode;
wb.flags = mat.flags;
}
model.batches.push_back(wb);
}
// Convert animations (first keyframe per bone per sequence)
for (const auto& seq : m2.sequences) {
WoweeModel::Animation anim;
@ -353,18 +411,38 @@ M2Model WoweeModelLoader::toM2(const WoweeModel& wom) {
tex.filename = tp;
m.textures.push_back(tex);
}
m.textureLookup = {0};
m.textureLookup.clear();
for (uint32_t i = 0; i < wom.texturePaths.size(); i++)
m.textureLookup.push_back(static_cast<int16_t>(i));
if (m.textureLookup.empty()) m.textureLookup.push_back(0);
M2Batch batch{};
batch.textureCount = std::min(1u, static_cast<uint32_t>(wom.texturePaths.size()));
batch.indexCount = static_cast<uint32_t>(m.indices.size());
batch.vertexCount = static_cast<uint32_t>(m.vertices.size());
m.batches.push_back(batch);
M2Material mat;
mat.flags = 0;
mat.blendMode = 0;
m.materials.push_back(mat);
if (wom.hasBatches()) {
for (const auto& wb : wom.batches) {
M2Batch batch{};
batch.indexStart = wb.indexStart;
batch.indexCount = wb.indexCount;
batch.vertexCount = static_cast<uint32_t>(m.vertices.size());
batch.textureCount = 1;
batch.textureIndex = static_cast<uint16_t>(
std::min<uint32_t>(wb.textureIndex, m.textureLookup.size() - 1));
batch.materialIndex = static_cast<uint16_t>(m.materials.size());
m.batches.push_back(batch);
M2Material mat;
mat.flags = wb.flags;
mat.blendMode = wb.blendMode;
m.materials.push_back(mat);
}
} else {
M2Batch batch{};
batch.textureCount = std::min(1u, static_cast<uint32_t>(wom.texturePaths.size()));
batch.indexCount = static_cast<uint32_t>(m.indices.size());
batch.vertexCount = static_cast<uint32_t>(m.vertices.size());
m.batches.push_back(batch);
M2Material mat;
mat.flags = 0;
mat.blendMode = 0;
m.materials.push_back(mat);
}
return m;
}