gltf: draw rigged_simple animation

This includes a slightly more correct interpretation of GLTF animation
samplers.
This commit is contained in:
Zack Buhman 2026-01-04 19:14:38 -06:00
parent 3451ca26c7
commit 473bf1965b
4 changed files with 1771 additions and 136 deletions

View File

@ -1,3 +1,6 @@
#ifndef GLTF_HPP_
#define GLTF_HPP_
struct Mesh {
const D3DXVECTOR3 * position;
const DWORD position_size;
@ -21,6 +24,7 @@ struct Mesh {
struct Skin;
struct Node {
const DWORD parent_ix;
const Skin * skin; // skin index (global)
const Mesh * mesh; // mesh index (global)
const D3DXVECTOR3 translation;
@ -30,7 +34,7 @@ struct Node {
struct Skin {
const D3DXMATRIX * inverse_bind_matrices; // accessor
const Node ** joints;
const int * joints;
DWORD joints_length;
};
@ -50,12 +54,9 @@ struct AnimationSampler {
struct AnimationChannel {
const AnimationSampler * sampler; // sampler index, this animation
struct {
const Node * node; // node index
const int node_ix;
const AnimationChannelPath path; // property to animate
} target;
};
//struct Animation {
// const AnimationChannel * channels;
// const AnimationSampler * samplers;
//};
#endif

10
gltf_instance.hpp Normal file
View File

@ -0,0 +1,10 @@
#ifndef GLTF_INSTANCE_HPP_
#define GLTF_INSTANCE_HPP_
struct NodeInstance {
D3DXVECTOR3 translation;
D3DXQUATERNION rotation;
D3DXVECTOR3 scale;
};
#endif

193
main.cpp
View File

@ -3,9 +3,13 @@
#include <d3d10.h>
#include <d3dx10.h>
#include <strsafe.h>
#include <assert.h>
#include "gltf.hpp"
#include "skin.hpp"
#include "gltf_instance.hpp"
#include "rigged_simple.hpp"
#define ROOT_MESH_NODE node_2
HINSTANCE g_hInstance = NULL;
HWND g_hWnd = NULL;
@ -42,6 +46,7 @@ LRESULT CALLBACK WndProc(HWND hWnd, UINT message, WPARAM wParam, LPARAM lParam);
HRESULT InitDirect3DDevice();
void Render();
BOOL Resize();
void InitializeNodeInstances();
struct WindowSize {
UINT Width;
@ -88,6 +93,8 @@ int WINAPI wWinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPWSTR lpCmdLi
return 0;
}
InitializeNodeInstances();
MSG msg = {};
while (msg.message != WM_QUIT) {
if (PeekMessage(&msg, NULL, 0, 0, PM_REMOVE)) {
@ -399,17 +406,18 @@ HRESULT InitDirect3DDevice()
//////////////////////////////////////////////////////////////////////
// vertex buffers
//////////////////////////////////////////////////////////////////////
#define MESH mesh_0
const Mesh * mesh = ROOT_MESH_NODE.mesh;
ID3D10Buffer * pVertexBuffers[3];
// position
bd.Usage = D3D10_USAGE_DEFAULT;
bd.ByteWidth = MESH.position_size;
bd.ByteWidth = mesh->position_size;
bd.BindFlags = D3D10_BIND_VERTEX_BUFFER;
bd.CPUAccessFlags = 0;
bd.MiscFlags = 0;
initData.pSysMem = MESH.position;
initData.pSysMem = mesh->position;
hr = g_pd3dDevice->CreateBuffer(&bd, &initData, &pVertexBuffers[0]);
if (FAILED(hr)) {
print("CreateBuffer\n");
@ -418,11 +426,11 @@ HRESULT InitDirect3DDevice()
// weights
bd.Usage = D3D10_USAGE_DEFAULT;
bd.ByteWidth = MESH.weights_0_size;
bd.ByteWidth = mesh->weights_0_size;
bd.BindFlags = D3D10_BIND_VERTEX_BUFFER;
bd.CPUAccessFlags = 0;
bd.MiscFlags = 0;
initData.pSysMem = MESH.weights_0;
initData.pSysMem = mesh->weights_0;
hr = g_pd3dDevice->CreateBuffer(&bd, &initData, &pVertexBuffers[1]);
if (FAILED(hr)) {
print("CreateBuffer\n");
@ -431,11 +439,11 @@ HRESULT InitDirect3DDevice()
// joints
bd.Usage = D3D10_USAGE_DEFAULT;
bd.ByteWidth = MESH.joints_0_size;
bd.ByteWidth = mesh->joints_0_size;
bd.BindFlags = D3D10_BIND_VERTEX_BUFFER;
bd.CPUAccessFlags = 0;
bd.MiscFlags = 0;
initData.pSysMem = MESH.joints_0;
initData.pSysMem = mesh->joints_0;
hr = g_pd3dDevice->CreateBuffer(&bd, &initData, &pVertexBuffers[2]);
if (FAILED(hr)) {
print("CreateBuffer\n");
@ -443,9 +451,9 @@ HRESULT InitDirect3DDevice()
}
UINT stride[] = {
(sizeof (MESH.position[0])),
(sizeof (MESH.weights_0[0])),
(sizeof (MESH.joints_0[0])),
(sizeof (mesh->position[0])),
(sizeof (mesh->weights_0[0])),
(sizeof (mesh->joints_0[0])),
};
UINT offset[] = { 0, 0, 0 };
g_pd3dDevice->IASetVertexBuffers(0, 3, pVertexBuffers, stride, offset);
@ -455,12 +463,12 @@ HRESULT InitDirect3DDevice()
//////////////////////////////////////////////////////////////////////
bd.Usage = D3D10_USAGE_DEFAULT;
bd.ByteWidth = MESH.indices_size;
bd.ByteWidth = mesh->indices_size;
//bd.ByteWidth = (sizeof (DWORD)) * indices_length;
bd.BindFlags = D3D10_BIND_INDEX_BUFFER;
bd.CPUAccessFlags = 0;
bd.MiscFlags = 0;
initData.pSysMem = MESH.indices;
initData.pSysMem = mesh->indices;
//initData.pSysMem = indices;
hr = g_pd3dDevice->CreateBuffer(&bd, &initData, &g_pIndexBuffer);
if (FAILED(hr))
@ -476,7 +484,7 @@ HRESULT InitDirect3DDevice()
D3DXMatrixIdentity(&g_World1);
D3DXMatrixIdentity(&g_World2);
D3DXVECTOR3 Eye(0.0f, 0.0f, -3.0f);
D3DXVECTOR3 Eye(0.0f, 0.0f, -10.0f);
D3DXVECTOR3 At(0.0f, 0.0f, 0.0f);
D3DXVECTOR3 Up(0.0f, 1.0f, 0.0f);
D3DXMatrixLookAtLH(&g_View, &Eye, &At, &Up);
@ -534,45 +542,6 @@ BOOL Resize()
return true;
}
/*
void Animate(float t)
{
const float * frames = accessor_2;
const int frames_length = accessor_2_length;
while (t >= frames[frames_length - 1]) {
t -= frames[frames_length - 1];
}
// find frame
int prev_ix = -1;
for (int i = 0; i < frames_length - 1; i++) {
if (frames[i] <= t && frames[i+1] >= t) {
prev_ix = i;
break;
}
}
if (prev_ix == -1)
return;
int next_ix = prev_ix + 1;
if (next_ix >= frames_length)
return;
float lerp = (t - frames[prev_ix]) / (frames[next_ix] - frames[prev_ix]);
print("%f prev %d next %d lerp %f\n", t, prev_ix, next_ix, lerp);
const D3DXQUATERNION * animation = (D3DXQUATERNION *)accessor_3;
D3DXQUATERNION rotation;
D3DXQuaternionSlerp(&rotation,
&animation[prev_ix],
&animation[next_ix],
lerp);
D3DXMatrixRotationQuaternion(&g_World1, &rotation);
}
*/
static inline D3DXMATRIX MatrixTRS(const D3DXVECTOR3 * translation,
const D3DXQUATERNION * rotation,
const D3DXVECTOR3 * scaling)
@ -616,35 +585,98 @@ static inline float Lerp(const float * frames, float t, int frame_ix)
D3DXMATRIX mJoints[2];
NodeInstance node_inst[nodes_length];
void InitializeNodeInstances()
{
for (int i = 0; i < nodes_length; i++) {
node_inst[i].translation = nodes[i]->translation;
node_inst[i].rotation = nodes[i]->rotation;
node_inst[i].scale = nodes[i]->scale;
}
}
void VectorLerp(D3DXVECTOR3 * output,
const D3DXVECTOR3 * a,
const D3DXVECTOR3 * b,
const float t)
{
*output = *a + t * (*b - *a);
}
void Animate(float t)
{
t = loop(t, 5.5);
const AnimationChannel * channels = animation_0__channels;
const int channels_length = (sizeof (animation_0__channels)) / (sizeof (animation_0__channels[0]));
// animation_0__sampler_0
const AnimationSampler * sampler = &animation_0__sampler_0;
const float * input = sampler->input;
const D3DXQUATERNION * output = (D3DXQUATERNION *)sampler->output;
t = loop(t, 2.0833330154418945);
int frame_ix = FindFrame(sampler->input, sampler->length, t);
float lerp = Lerp(sampler->input, t, frame_ix);
D3DXQUATERNION rotation;
D3DXQuaternionSlerp(&rotation,
&output[frame_ix],
&output[frame_ix+1],
lerp);
// find frame and lerp (same accessor for all channels)
const float * input = channels[0].sampler->input;
const int input_length = channels[0].sampler->length;
// joint 1
const Skin * skin = &skin_0;
const Node * node = skin->joints[1];
int frame_ix = FindFrame(input, input_length, t);
float lerp = Lerp(input, t, frame_ix);
// T * R * S
D3DXMATRIX global_transform = MatrixTRS(&node->translation, &rotation, &node->scale);
// sample all channels
for (int i = 0; i < channels_length; i++) {
const AnimationSampler * sampler = channels[i].sampler;
NodeInstance * instance = &node_inst[channels[i].target.node_ix];
switch (channels[i].target.path) {
case ACP__TRANSLATION:
{
const D3DXVECTOR3 * output = (const D3DXVECTOR3 *)sampler->output;
VectorLerp(&instance->translation,
&output[frame_ix],
&output[frame_ix+1],
lerp);
break;
}
case ACP__ROTATION:
{
const D3DXQUATERNION * output = (const D3DXQUATERNION *)sampler->output;
D3DXQuaternionSlerp(&instance->rotation,
&output[frame_ix],
&output[frame_ix+1],
lerp);
D3DXMatrixIdentity(&mJoints[0]);
break;
}
case ACP__SCALE:
{
const D3DXVECTOR3 * output = (const D3DXVECTOR3 *)sampler->output;
VectorLerp(&instance->scale,
&output[frame_ix],
&output[frame_ix+1],
lerp);
break;
}
default:
assert(!"invalid sampler path");
break;
}
}
const D3DXMATRIX& inverse_bind_matrix = skin->inverse_bind_matrices[1];
// transform all joints
const Skin * skin = ROOT_MESH_NODE.skin;
assert(skin->joints_length == 2);
for (DWORD i = 0; i < skin->joints_length; i++) {
const int joint_ix = skin->joints[i];
const NodeInstance * instance = &node_inst[joint_ix];
mJoints[1] = inverse_bind_matrix * global_transform;
const D3DXMATRIX& inverse_bind_matrix = skin->inverse_bind_matrices[i];
// T * R * S
//&instance->translation,
//&instance->scale
D3DXVECTOR3 translation = D3DXVECTOR3(0, 0, 0);
D3DXVECTOR3 scale = D3DXVECTOR3(1, 1, 1);
D3DXMATRIX global_transform = MatrixTRS(&instance->translation,
&instance->rotation,
&instance->scale);
mJoints[i] = inverse_bind_matrix * global_transform;
}
}
void Render()
@ -660,10 +692,19 @@ void Render()
t = (dwTimeCur - dwTimeStart) / 1000.0f;
#endif
D3DXMatrixIdentity(&mJoints[0]);
D3DXMatrixIdentity(&mJoints[1]);
Animate(t);
// first cube
//D3DXMatrixRotationZ(&g_World1, t);
D3DXMATRIX rx;
D3DXMATRIX ry;
D3DXMatrixRotationX(&ry, D3DX_PI * -0.5);
D3DXMatrixRotationZ(&rx, D3DX_PI * 0.5);
D3DXMatrixMultiply(&g_World1,
&rx,
&ry);
// lights
D3DXVECTOR4 vLightDirs[2] = {
@ -707,9 +748,11 @@ void Render()
// render first cube
D3D10_TECHNIQUE_DESC techDesc;
g_pTechniqueRender->GetDesc(&techDesc);
int indices_length = ROOT_MESH_NODE.mesh->indices_size / (sizeof (DWORD));
for (UINT p = 0; p < techDesc.Passes; p++) {
g_pTechniqueRender->GetPassByIndex(p)->Apply(0);
g_pd3dDevice->DrawIndexed(accessor_0_length, 0, 0);
g_pd3dDevice->DrawIndexed(indices_length, 0, 0);
}
// render the lights

1581
rigged_simple.hpp Normal file

File diff suppressed because it is too large Load Diff