blendo/quadrilateralcowboy
0
0
Fork 0
mirror of https://github.com/blendogames/quadrilateralcowboy.git synced 2024-11-25 13:31:22 +00:00
Code Issues Projects Releases Packages Wiki Activity
quadrilateralcowboy/renderer/Model.cpp
brendon 47a1010d23 Initial commit
2020-06-12 14:06:25 -07:00

2328 lines
62 KiB
C++
Raw Blame History

/*
===========================================================================
Doom 3 GPL Source Code
Copyright (C) 1999-2011 id Software LLC, a ZeniMax Media company.
This file is part of the Doom 3 GPL Source Code (?Doom 3 Source Code?).
Doom 3 Source Code is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Doom 3 Source Code is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with Doom 3 Source Code. If not, see <http://www.gnu.org/licenses/>.
In addition, the Doom 3 Source Code is also subject to certain additional terms. You should have received a copy of these additional terms immediately following the terms and conditions of the GNU General Public License which accompanied the Doom 3 Source Code. If not, please request a copy in writing from id Software at the address below.
If you have questions concerning this license or the applicable additional terms, you may contact in writing id Software LLC, c/o ZeniMax Media Inc., Suite 120, Rockville, Maryland 20850 USA.
===========================================================================
*/
#include "../idlib/precompiled.h"
#pragma hdrstop
#include "tr_local.h"
#include "Model_local.h"
#include "Model_ase.h"
#include "Model_lwo.h"
#include "Model_ma.h"
idCVar idRenderModelStatic::r_mergeModelSurfaces( "r_mergeModelSurfaces", "1", CVAR_BOOL|CVAR_RENDERER, "combine model surfaces with the same material" );
idCVar idRenderModelStatic::r_slopVertex( "r_slopVertex", "0.01", CVAR_RENDERER, "merge xyz coordinates this far apart" );
idCVar idRenderModelStatic::r_slopTexCoord( "r_slopTexCoord", "0.001", CVAR_RENDERER, "merge texture coordinates this far apart" );
idCVar idRenderModelStatic::r_slopNormal( "r_slopNormal", "0.02", CVAR_RENDERER, "merge normals that dot less than this" );
/*
================
idRenderModelStatic::idRenderModelStatic
================
*/
idRenderModelStatic::idRenderModelStatic() {
name = "<undefined>";
bounds.Clear();
lastModifiedFrame = 0;
lastArchivedFrame = 0;
overlaysAdded = 0;
shadowHull = NULL;
isStaticWorldModel = false;
defaulted = false;
purged = false;
fastLoad = false;
reloadable = true;
levelLoadReferenced = false;
timeStamp = 0;
}
/*
================
idRenderModelStatic::~idRenderModelStatic
================
*/
idRenderModelStatic::~idRenderModelStatic() {
PurgeModel();
}
/*
==============
idRenderModelStatic::Print
==============
*/
void idRenderModelStatic::Print() const {
common->Printf( "%s\n", name.c_str() );
common->Printf( "Static model.\n" );
common->Printf( "bounds: (%f %f %f) to (%f %f %f)\n",
bounds[0][0], bounds[0][1], bounds[0][2],
bounds[1][0], bounds[1][1], bounds[1][2] );
common->Printf( " verts tris material\n" );
for ( int i = 0 ; i < NumSurfaces() ; i++ ) {
const modelSurface_t *surf = Surface( i );
srfTriangles_t *tri = surf->geometry;
const idMaterial *material = surf->shader;
if ( !tri ) {
common->Printf( "%2i: %s, NULL surface geometry\n", i, material->GetName() );
continue;
}
common->Printf( "%2i: %5i %5i %s", i, tri->numVerts, tri->numIndexes / 3, material->GetName() );
if ( tri->generateNormals ) {
common->Printf( " (smoothed)\n" );
} else {
common->Printf( "\n" );
}
}
}
/*
==============
idRenderModelStatic::Memory
==============
*/
int idRenderModelStatic::Memory() const {
int totalBytes = 0;
totalBytes += sizeof( *this );
totalBytes += name.DynamicMemoryUsed();
totalBytes += surfaces.MemoryUsed();
if ( shadowHull ) {
totalBytes += R_TriSurfMemory( shadowHull );
}
for ( int j = 0 ; j < NumSurfaces() ; j++ ) {
const modelSurface_t *surf = Surface( j );
if ( !surf->geometry ) {
continue;
}
totalBytes += R_TriSurfMemory( surf->geometry );
}
return totalBytes;
}
/*
==============
idRenderModelStatic::List
==============
*/
void idRenderModelStatic::List() const {
int totalTris = 0;
int totalVerts = 0;
int totalBytes = 0;
totalBytes = Memory();
char closed = 'C';
for ( int j = 0 ; j < NumSurfaces() ; j++ ) {
const modelSurface_t *surf = Surface( j );
if ( !surf->geometry ) {
continue;
}
if ( !surf->geometry->perfectHull ) {
closed = ' ';
}
totalTris += surf->geometry->numIndexes / 3;
totalVerts += surf->geometry->numVerts;
}
common->Printf( "%c%4ik %3i %4i %4i %s", closed, totalBytes/1024, NumSurfaces(), totalVerts, totalTris, Name() );
if ( IsDynamicModel() == DM_CACHED ) {
common->Printf( " (DM_CACHED)" );
}
if ( IsDynamicModel() == DM_CONTINUOUS ) {
common->Printf( " (DM_CONTINUOUS)" );
}
if ( defaulted ) {
common->Printf( " (DEFAULTED)" );
}
if ( bounds[0][0] >= bounds[1][0] ) {
common->Printf( " (EMPTY BOUNDS)" );
}
if ( bounds[1][0] - bounds[0][0] > 100000 ) {
common->Printf( " (HUGE BOUNDS)" );
}
common->Printf( "\n" );
}
/*
================
idRenderModelStatic::IsDefaultModel
================
*/
bool idRenderModelStatic::IsDefaultModel() const {
return defaulted;
}
/*
================
AddCubeFace
================
*/
static void AddCubeFace( srfTriangles_t *tri, idVec3 v1, idVec3 v2, idVec3 v3, idVec3 v4 ) {
tri->verts[tri->numVerts+0].Clear();
tri->verts[tri->numVerts+0].xyz = v1 * 8;
tri->verts[tri->numVerts+0].st[0] = 0;
tri->verts[tri->numVerts+0].st[1] = 0;
tri->verts[tri->numVerts+1].Clear();
tri->verts[tri->numVerts+1].xyz = v2 * 8;
tri->verts[tri->numVerts+1].st[0] = 1;
tri->verts[tri->numVerts+1].st[1] = 0;
tri->verts[tri->numVerts+2].Clear();
tri->verts[tri->numVerts+2].xyz = v3 * 8;
tri->verts[tri->numVerts+2].st[0] = 1;
tri->verts[tri->numVerts+2].st[1] = 1;
tri->verts[tri->numVerts+3].Clear();
tri->verts[tri->numVerts+3].xyz = v4 * 8;
tri->verts[tri->numVerts+3].st[0] = 0;
tri->verts[tri->numVerts+3].st[1] = 1;
tri->indexes[tri->numIndexes+0] = tri->numVerts + 0;
tri->indexes[tri->numIndexes+1] = tri->numVerts + 1;
tri->indexes[tri->numIndexes+2] = tri->numVerts + 2;
tri->indexes[tri->numIndexes+3] = tri->numVerts + 0;
tri->indexes[tri->numIndexes+4] = tri->numVerts + 2;
tri->indexes[tri->numIndexes+5] = tri->numVerts + 3;
tri->numVerts += 4;
tri->numIndexes += 6;
}
/*
================
idRenderModelStatic::MakeDefaultModel
================
*/
void idRenderModelStatic::MakeDefaultModel() {
defaulted = true;
// throw out any surfaces we already have
PurgeModel();
// create one new surface
modelSurface_t surf;
srfTriangles_t *tri = R_AllocStaticTriSurf();
surf.shader = tr.defaultMaterial;
surf.geometry = tri;
R_AllocStaticTriSurfVerts( tri, 24 );
R_AllocStaticTriSurfIndexes( tri, 36 );
AddCubeFace( tri, idVec3(-1, 1, 1), idVec3(1, 1, 1), idVec3(1, -1, 1), idVec3(-1, -1, 1) );
AddCubeFace( tri, idVec3(-1, 1, -1), idVec3(-1, -1, -1), idVec3(1, -1, -1), idVec3(1, 1, -1) );
AddCubeFace( tri, idVec3(1, -1, 1), idVec3(1, 1, 1), idVec3(1, 1, -1), idVec3(1, -1, -1) );
AddCubeFace( tri, idVec3(-1, -1, 1), idVec3(-1, -1, -1), idVec3(-1, 1, -1), idVec3(-1, 1, 1) );
AddCubeFace( tri, idVec3(-1, -1, 1), idVec3(1, -1, 1), idVec3(1, -1, -1), idVec3(-1, -1, -1) );
AddCubeFace( tri, idVec3(-1, 1, 1), idVec3(-1, 1, -1), idVec3(1, 1, -1), idVec3(1, 1, 1) );
tri->generateNormals = true;
AddSurface( surf );
FinishSurfaces();
}
/*
================
idRenderModelStatic::PartialInitFromFile
================
*/
void idRenderModelStatic::PartialInitFromFile( const char *fileName ) {
fastLoad = true;
InitFromFile( fileName );
}
/*
================
idRenderModelStatic::InitFromFile
================
*/
void idRenderModelStatic::InitFromFile( const char *fileName ) {
bool loaded;
idStr extension;
InitEmpty( fileName );
// FIXME: load new .proc map format
name.ExtractFileExtension( extension );
if ( extension.Icmp( "ase" ) == 0 ) {
loaded = LoadASE( name );
reloadable = true;
} else if ( extension.Icmp( "lwo" ) == 0 ) {
loaded = LoadLWO( name );
reloadable = true;
} else if ( extension.Icmp( "flt" ) == 0 ) {
loaded = LoadFLT( name );
reloadable = true;
} else if ( extension.Icmp( "ma" ) == 0 ) {
loaded = LoadMA( name );
reloadable = true;
} else {
common->Warning( "idRenderModelStatic::InitFromFile: unknown type for model: \'%s\'", name.c_str() );
loaded = false;
}
if ( !loaded ) {
common->Warning( "Couldn't load model: '%s'", name.c_str() );
MakeDefaultModel();
return;
}
// it is now available for use
purged = false;
// create the bounds for culling and dynamic surface creation
FinishSurfaces();
}
/*
================
idRenderModelStatic::LoadModel
================
*/
void idRenderModelStatic::LoadModel() {
PurgeModel();
InitFromFile( name );
}
/*
================
idRenderModelStatic::InitEmpty
================
*/
void idRenderModelStatic::InitEmpty( const char *fileName ) {
// model names of the form _area* are static parts of the
// world, and have already been considered for optimized shadows
// other model names are inline entity models, and need to be
// shadowed normally
if ( !idStr::Cmpn( fileName, "_area", 5 ) ) {
isStaticWorldModel = true;
} else {
isStaticWorldModel = false;
}
name = fileName;
reloadable = false; // if it didn't come from a file, we can't reload it
PurgeModel();
purged = false;
bounds.Zero();
}
/*
================
idRenderModelStatic::AddSurface
================
*/
void idRenderModelStatic::AddSurface( modelSurface_t surface ) {
surfaces.Append( surface );
if ( surface.geometry ) {
bounds += surface.geometry->bounds;
}
}
/*
================
idRenderModelStatic::Name
================
*/
const char *idRenderModelStatic::Name() const {
return name;
}
/*
================
idRenderModelStatic::Timestamp
================
*/
ID_TIME_T idRenderModelStatic::Timestamp() const {
return timeStamp;
}
/*
================
idRenderModelStatic::NumSurfaces
================
*/
int idRenderModelStatic::NumSurfaces() const {
return surfaces.Num();
}
/*
================
idRenderModelStatic::NumBaseSurfaces
================
*/
int idRenderModelStatic::NumBaseSurfaces() const {
return surfaces.Num() - overlaysAdded;
}
/*
================
idRenderModelStatic::Surface
================
*/
const modelSurface_t *idRenderModelStatic::Surface( int surfaceNum ) const {
return &surfaces[surfaceNum];
}
/*
================
idRenderModelStatic::AllocSurfaceTriangles
================
*/
srfTriangles_t *idRenderModelStatic::AllocSurfaceTriangles( int numVerts, int numIndexes ) const {
srfTriangles_t *tri = R_AllocStaticTriSurf();
R_AllocStaticTriSurfVerts( tri, numVerts );
R_AllocStaticTriSurfIndexes( tri, numIndexes );
return tri;
}
/*
================
idRenderModelStatic::FreeSurfaceTriangles
================
*/
void idRenderModelStatic::FreeSurfaceTriangles( srfTriangles_t *tris ) const {
R_FreeStaticTriSurf( tris );
}
/*
================
idRenderModelStatic::ShadowHull
================
*/
srfTriangles_t *idRenderModelStatic::ShadowHull() const {
return shadowHull;
}
/*
================
idRenderModelStatic::IsStaticWorldModel
================
*/
bool idRenderModelStatic::IsStaticWorldModel() const {
return isStaticWorldModel;
}
/*
================
idRenderModelStatic::IsDynamicModel
================
*/
dynamicModel_t idRenderModelStatic::IsDynamicModel() const {
// dynamic subclasses will override this
return DM_STATIC;
}
/*
================
idRenderModelStatic::IsReloadable
================
*/
bool idRenderModelStatic::IsReloadable() const {
return reloadable;
}
/*
================
idRenderModelStatic::Bounds
================
*/
idBounds idRenderModelStatic::Bounds( const struct renderEntity_s *mdef ) const {
return bounds;
}
/*
================
idRenderModelStatic::DepthHack
================
*/
float idRenderModelStatic::DepthHack() const {
return 0.0f;
}
/*
================
idRenderModelStatic::InstantiateDynamicModel
================
*/
idRenderModel *idRenderModelStatic::InstantiateDynamicModel( const struct renderEntity_s *ent, const struct viewDef_s *view, idRenderModel *cachedModel ) {
if ( cachedModel ) {
delete cachedModel;
cachedModel = NULL;
}
common->Error( "InstantiateDynamicModel called on static model '%s'", name.c_str() );
return NULL;
}
/*
================
idRenderModelStatic::NumJoints
================
*/
int idRenderModelStatic::NumJoints( void ) const {
return 0;
}
/*
================
idRenderModelStatic::GetJoints
================
*/
const idMD5Joint *idRenderModelStatic::GetJoints( void ) const {
return NULL;
}
/*
================
idRenderModelStatic::GetJointHandle
================
*/
jointHandle_t idRenderModelStatic::GetJointHandle( const char *name ) const {
return INVALID_JOINT;
}
/*
================
idRenderModelStatic::GetJointName
================
*/
const char * idRenderModelStatic::GetJointName( jointHandle_t handle ) const {
return "";
}
/*
================
idRenderModelStatic::GetDefaultPose
================
*/
const idJointQuat *idRenderModelStatic::GetDefaultPose( void ) const {
return NULL;
}
/*
================
idRenderModelStatic::NearestJoint
================
*/
int idRenderModelStatic::NearestJoint( int surfaceNum, int a, int b, int c ) const {
return INVALID_JOINT;
}
//=====================================================================
/*
================
idRenderModelStatic::FinishSurfaces
The mergeShadows option allows surfaces with different textures to share
silhouette edges for shadow calculation, instead of leaving shared edges
hanging.
If any of the original shaders have the noSelfShadow flag set, the surfaces
can't be merged, because they will need to be drawn in different order.
If there is only one surface, a separate merged surface won't be generated.
A model with multiple surfaces can't later have a skinned shader change the
state of the noSelfShadow flag.
-----------------
Creates mirrored copies of two sided surfaces with normal maps, which would
otherwise light funny.
Extends the bounds of deformed surfaces so they don't cull incorrectly at screen edges.
================
*/
void idRenderModelStatic::FinishSurfaces() {
int i;
int totalVerts, totalIndexes;
purged = false;
// make sure we don't have a huge bounds even if we don't finish everything
bounds.Zero();
if ( surfaces.Num() == 0 ) {
return;
}
// renderBump doesn't care about most of this
if ( fastLoad ) {
bounds.Zero();
for ( i = 0 ; i < surfaces.Num() ; i++ ) {
const modelSurface_t *surf = &surfaces[i];
R_BoundTriSurf( surf->geometry );
bounds.AddBounds( surf->geometry->bounds );
}
return;
}
// cleanup all the final surfaces, but don't create sil edges
totalVerts = 0;
totalIndexes = 0;
// decide if we are going to merge all the surfaces into one shadower
int numOriginalSurfaces = surfaces.Num();
// make sure there aren't any NULL shaders or geometry
for ( i = 0 ; i < numOriginalSurfaces ; i++ ) {
const modelSurface_t *surf = &surfaces[i];
if ( surf->geometry == NULL || surf->shader == NULL ) {
MakeDefaultModel();
common->Error( "Model %s, surface %i had NULL geometry", name.c_str(), i );
}
if ( surf->shader == NULL ) {
MakeDefaultModel();
common->Error( "Model %s, surface %i had NULL shader", name.c_str(), i );
}
}
// duplicate and reverse triangles for two sided bump mapped surfaces
// note that this won't catch surfaces that have their shaders dynamically
// changed, and won't work with animated models.
// It is better to create completely separate surfaces, rather than
// add vertexes and indexes to the existing surface, because the
// tangent generation wouldn't like the acute shared edges
for ( i = 0 ; i < numOriginalSurfaces ; i++ ) {
const modelSurface_t *surf = &surfaces[i];
if ( surf->shader->ShouldCreateBackSides() ) {
srfTriangles_t *newTri;
newTri = R_CopyStaticTriSurf( surf->geometry );
R_ReverseTriangles( newTri );
modelSurface_t newSurf;
newSurf.shader = surf->shader;
newSurf.geometry = newTri;
AddSurface( newSurf );
}
}
// clean the surfaces
for ( i = 0 ; i < surfaces.Num() ; i++ ) {
const modelSurface_t *surf = &surfaces[i];
R_CleanupTriangles( surf->geometry, surf->geometry->generateNormals, true, surf->shader->UseUnsmoothedTangents() );
if ( surf->shader->SurfaceCastsShadow() ) {
totalVerts += surf->geometry->numVerts;
totalIndexes += surf->geometry->numIndexes;
}
}
// add up the total surface area for development information
for ( i = 0 ; i < surfaces.Num() ; i++ ) {
const modelSurface_t *surf = &surfaces[i];
srfTriangles_t *tri = surf->geometry;
for ( int j = 0 ; j < tri->numIndexes ; j += 3 ) {
float area = idWinding::TriangleArea( tri->verts[tri->indexes[j]].xyz,
tri->verts[tri->indexes[j+1]].xyz, tri->verts[tri->indexes[j+2]].xyz );
const_cast<idMaterial *>(surf->shader)->AddToSurfaceArea( area );
}
}
// calculate the bounds
if ( surfaces.Num() == 0 ) {
bounds.Zero();
} else {
bounds.Clear();
for ( i = 0 ; i < surfaces.Num() ; i++ ) {
modelSurface_t *surf = &surfaces[i];
// if the surface has a deformation, increase the bounds
// the amount here is somewhat arbitrary, designed to handle
// autosprites and flares, but could be done better with exact
// deformation information.
// Note that this doesn't handle deformations that are skinned in
// at run time...
if ( surf->shader->Deform() != DFRM_NONE ) {
srfTriangles_t *tri = surf->geometry;
idVec3 mid = ( tri->bounds[1] + tri->bounds[0] ) * 0.5f;
float radius = ( tri->bounds[0] - mid ).Length();
radius += 20.0f;
tri->bounds[0][0] = mid[0] - radius;
tri->bounds[0][1] = mid[1] - radius;
tri->bounds[0][2] = mid[2] - radius;
tri->bounds[1][0] = mid[0] + radius;
tri->bounds[1][1] = mid[1] + radius;
tri->bounds[1][2] = mid[2] + radius;
}
// add to the model bounds
bounds.AddBounds( surf->geometry->bounds );
}
}
}
/*
=================
idRenderModelStatic::ConvertASEToModelSurfaces
=================
*/
typedef struct matchVert_s {
struct matchVert_s *next;
int v, tv;
byte color[4];
idVec3 normal;
} matchVert_t;
bool idRenderModelStatic::ConvertASEToModelSurfaces( const struct aseModel_s *ase ) {
aseObject_t * object;
aseMesh_t * mesh;
aseMaterial_t * material;
const idMaterial *im1, *im2;
srfTriangles_t *tri;
int objectNum;
int i, j, k;
int v, tv;
int * vRemap;
int * tvRemap;
matchVert_t * mvTable; // all of the match verts
matchVert_t ** mvHash; // points inside mvTable for each xyz index
matchVert_t * lastmv;
matchVert_t * mv;
idVec3 normal;
float uOffset, vOffset, textureSin, textureCos;
float uTiling, vTiling;
int * mergeTo;
byte * color;
static byte identityColor[4] = { 255, 255, 255, 255 };
modelSurface_t surf, *modelSurf;
if ( !ase ) {
return false;
}
if ( ase->objects.Num() < 1 ) {
return false;
}
timeStamp = ase->timeStamp;
// the modeling programs can save out multiple surfaces with a common
// material, but we would like to mege them together where possible
// meaning that this->NumSurfaces() <= ase->objects.currentElements
mergeTo = (int *)_alloca( ase->objects.Num() * sizeof( *mergeTo ) );
surf.geometry = NULL;
if ( ase->materials.Num() == 0 ) {
// if we don't have any materials, dump everything into a single surface
surf.shader = tr.defaultMaterial;
surf.id = 0;
this->AddSurface( surf );
for ( i = 0 ; i < ase->objects.Num() ; i++ ) {
mergeTo[i] = 0;
}
} else if ( !r_mergeModelSurfaces.GetBool() ) {
// don't merge any
for ( i = 0 ; i < ase->objects.Num() ; i++ ) {
mergeTo[i] = i;
object = ase->objects[i];
material = ase->materials[object->materialRef];
surf.shader = declManager->FindMaterial( material->name );
surf.id = this->NumSurfaces();
this->AddSurface( surf );
}
} else {
// search for material matches
for ( i = 0 ; i < ase->objects.Num() ; i++ ) {
object = ase->objects[i];
material = ase->materials[object->materialRef];
im1 = declManager->FindMaterial( material->name );
if ( im1->IsDiscrete() ) {
// flares, autosprites, etc
j = this->NumSurfaces();
} else {
for ( j = 0 ; j < this->NumSurfaces() ; j++ ) {
modelSurf = &this->surfaces[j];
im2 = modelSurf->shader;
if ( im1 == im2 ) {
// merge this
mergeTo[i] = j;
break;
}
}
}
if ( j == this->NumSurfaces() ) {
// didn't merge
mergeTo[i] = j;
surf.shader = im1;
surf.id = this->NumSurfaces();
this->AddSurface( surf );
}
}
}
idVectorSubset<idVec3, 3> vertexSubset;
idVectorSubset<idVec2, 2> texCoordSubset;
// build the surfaces
for ( objectNum = 0 ; objectNum < ase->objects.Num() ; objectNum++ ) {
object = ase->objects[objectNum];
mesh = &object->mesh;
material = ase->materials[object->materialRef];
im1 = declManager->FindMaterial( material->name );
bool normalsParsed = mesh->normalsParsed;
// completely ignore any explict normals on surfaces with a renderbump command
// which will guarantee the best contours and least vertexes.
const char *rb = im1->GetRenderBump();
if ( rb && rb[0] ) {
normalsParsed = false;
}
// It seems like the tools our artists are using often generate
// verts and texcoords slightly separated that should be merged
// note that we really should combine the surfaces with common materials
// before doing this operation, because we can miss a slop combination
// if they are in different surfaces
vRemap = (int *)R_StaticAlloc( mesh->numVertexes * sizeof( vRemap[0] ) );
if ( fastLoad ) {
// renderbump doesn't care about vertex count
for ( j = 0; j < mesh->numVertexes; j++ ) {
vRemap[j] = j;
}
} else {
float vertexEpsilon = r_slopVertex.GetFloat();
float expand = 2 * 32 * vertexEpsilon;
idVec3 mins, maxs;
SIMDProcessor->MinMax( mins, maxs, mesh->vertexes, mesh->numVertexes );
mins -= idVec3( expand, expand, expand );
maxs += idVec3( expand, expand, expand );
vertexSubset.Init( mins, maxs, 32, 1024 );
for ( j = 0; j < mesh->numVertexes; j++ ) {
vRemap[j] = vertexSubset.FindVector( mesh->vertexes, j, vertexEpsilon );
}
}
tvRemap = (int *)R_StaticAlloc( mesh->numTVertexes * sizeof( tvRemap[0] ) );
if ( fastLoad ) {
// renderbump doesn't care about vertex count
for ( j = 0; j < mesh->numTVertexes; j++ ) {
tvRemap[j] = j;
}
} else {
float texCoordEpsilon = r_slopTexCoord.GetFloat();
float expand = 2 * 32 * texCoordEpsilon;
idVec2 mins, maxs;
SIMDProcessor->MinMax( mins, maxs, mesh->tvertexes, mesh->numTVertexes );
mins -= idVec2( expand, expand );
maxs += idVec2( expand, expand );
texCoordSubset.Init( mins, maxs, 32, 1024 );
for ( j = 0; j < mesh->numTVertexes; j++ ) {
tvRemap[j] = texCoordSubset.FindVector( mesh->tvertexes, j, texCoordEpsilon );
}
}
// we need to find out how many unique vertex / texcoord combinations
// there are, because ASE tracks them separately but we need them unified
// the maximum possible number of combined vertexes is the number of indexes
mvTable = (matchVert_t *)R_ClearedStaticAlloc( mesh->numFaces * 3 * sizeof( mvTable[0] ) );
// we will have a hash chain based on the xyz values
mvHash = (matchVert_t **)R_ClearedStaticAlloc( mesh->numVertexes * sizeof( mvHash[0] ) );
// allocate triangle surface
tri = R_AllocStaticTriSurf();
tri->numVerts = 0;
tri->numIndexes = 0;
R_AllocStaticTriSurfIndexes( tri, mesh->numFaces * 3 );
tri->generateNormals = !normalsParsed;
// init default normal, color and tex coord index
normal.Zero();
color = identityColor;
tv = 0;
// find all the unique combinations
float normalEpsilon = 1.0f - r_slopNormal.GetFloat();
for ( j = 0; j < mesh->numFaces; j++ ) {
for ( k = 0; k < 3; k++ ) {
v = mesh->faces[j].vertexNum[k];
if ( v < 0 || v >= mesh->numVertexes ) {
common->Error( "ConvertASEToModelSurfaces: bad vertex index in ASE file %s", name.c_str() );
}
// collapse the position if it was slightly offset
v = vRemap[v];
// we may or may not have texcoords to compare
if ( mesh->numTVFaces == mesh->numFaces && mesh->numTVertexes != 0 ) {
tv = mesh->faces[j].tVertexNum[k];
if ( tv < 0 || tv >= mesh->numTVertexes ) {
common->Error( "ConvertASEToModelSurfaces: bad tex coord index in ASE file %s", name.c_str() );
}
// collapse the tex coord if it was slightly offset
tv = tvRemap[tv];
}
// we may or may not have normals to compare
if ( normalsParsed ) {
normal = mesh->faces[j].vertexNormals[k];
}
// we may or may not have colors to compare
if ( mesh->colorsParsed ) {
color = mesh->faces[j].vertexColors[k];
}
// find a matching vert
for ( lastmv = NULL, mv = mvHash[v]; mv != NULL; lastmv = mv, mv = mv->next ) {
if ( mv->tv != tv ) {
continue;
}
if ( *(unsigned *)mv->color != *(unsigned *)color ) {
continue;
}
if ( !normalsParsed ) {
// if we are going to create the normals, just
// matching texcoords is enough
break;
}
if ( mv->normal * normal > normalEpsilon ) {
break; // we already have this one
}
}
if ( !mv ) {
// allocate a new match vert and link to hash chain
mv = &mvTable[ tri->numVerts ];
mv->v = v;
mv->tv = tv;
mv->normal = normal;
*(unsigned *)mv->color = *(unsigned *)color;
mv->next = NULL;
if ( lastmv ) {
lastmv->next = mv;
} else {
mvHash[v] = mv;
}
tri->numVerts++;
}
tri->indexes[tri->numIndexes] = mv - mvTable;
tri->numIndexes++;
}
}
// allocate space for the indexes and copy them
if ( tri->numIndexes > mesh->numFaces * 3 ) {
common->FatalError( "ConvertASEToModelSurfaces: index miscount in ASE file %s", name.c_str() );
}
if ( tri->numVerts > mesh->numFaces * 3 ) {
common->FatalError( "ConvertASEToModelSurfaces: vertex miscount in ASE file %s", name.c_str() );
}
// an ASE allows the texture coordinates to be scaled, translated, and rotated
if ( ase->materials.Num() == 0 ) {
uOffset = vOffset = 0.0f;
uTiling = vTiling = 1.0f;
textureSin = 0.0f;
textureCos = 1.0f;
} else {
material = ase->materials[object->materialRef];
uOffset = -material->uOffset;
vOffset = material->vOffset;
uTiling = material->uTiling;
vTiling = material->vTiling;
textureSin = idMath::Sin( material->angle );
textureCos = idMath::Cos( material->angle );
}
// now allocate and generate the combined vertexes
R_AllocStaticTriSurfVerts( tri, tri->numVerts );
for ( j = 0; j < tri->numVerts; j++ ) {
mv = &mvTable[j];
tri->verts[ j ].Clear();
tri->verts[ j ].xyz = mesh->vertexes[ mv->v ];
tri->verts[ j ].normal = mv->normal;
*(unsigned *)tri->verts[j].color = *(unsigned *)mv->color;
if ( mesh->numTVFaces == mesh->numFaces && mesh->numTVertexes != 0 ) {
const idVec2 &tv = mesh->tvertexes[ mv->tv ];
float u = tv.x * uTiling + uOffset;
float v = tv.y * vTiling + vOffset;
tri->verts[ j ].st[0] = u * textureCos + v * textureSin;
tri->verts[ j ].st[1] = u * -textureSin + v * textureCos;
}
}
R_StaticFree( mvTable );
R_StaticFree( mvHash );
R_StaticFree( tvRemap );
R_StaticFree( vRemap );
// see if we need to merge with a previous surface of the same material
modelSurf = &this->surfaces[mergeTo[ objectNum ]];
srfTriangles_t *mergeTri = modelSurf->geometry;
if ( !mergeTri ) {
modelSurf->geometry = tri;
} else {
modelSurf->geometry = R_MergeTriangles( mergeTri, tri );
R_FreeStaticTriSurf( tri );
R_FreeStaticTriSurf( mergeTri );
}
}
return true;
}
/*
=================
idRenderModelStatic::ConvertLWOToModelSurfaces
=================
*/
bool idRenderModelStatic::ConvertLWOToModelSurfaces( const struct st_lwObject *lwo ) {
const idMaterial *im1, *im2;
srfTriangles_t *tri;
lwSurface * lwoSurf;
int numTVertexes;
int i, j, k;
int v, tv;
idVec3 * vList;
int * vRemap;
idVec2 * tvList;
int * tvRemap;
matchVert_t * mvTable; // all of the match verts
matchVert_t ** mvHash; // points inside mvTable for each xyz index
matchVert_t * lastmv;
matchVert_t * mv;
idVec3 normal;
int * mergeTo;
byte color[4];
modelSurface_t surf, *modelSurf;
if ( !lwo ) {
return false;
}
if ( lwo->surf == NULL ) {
return false;
}
timeStamp = lwo->timeStamp;
// count the number of surfaces
i = 0;
for ( lwoSurf = lwo->surf; lwoSurf; lwoSurf = lwoSurf->next ) {
i++;
}
// the modeling programs can save out multiple surfaces with a common
// material, but we would like to merge them together where possible
mergeTo = (int *)_alloca( i * sizeof( mergeTo[0] ) );
memset( &surf, 0, sizeof( surf ) );
if ( !r_mergeModelSurfaces.GetBool() ) {
// don't merge any
for ( lwoSurf = lwo->surf, i = 0; lwoSurf; lwoSurf = lwoSurf->next, i++ ) {
mergeTo[i] = i;
surf.shader = declManager->FindMaterial( lwoSurf->name );
surf.id = this->NumSurfaces();
this->AddSurface( surf );
}
} else {
// search for material matches
for ( lwoSurf = lwo->surf, i = 0; lwoSurf; lwoSurf = lwoSurf->next, i++ ) {
im1 = declManager->FindMaterial( lwoSurf->name );
if ( im1->IsDiscrete() ) {
// flares, autosprites, etc
j = this->NumSurfaces();
} else {
for ( j = 0 ; j < this->NumSurfaces() ; j++ ) {
modelSurf = &this->surfaces[j];
im2 = modelSurf->shader;
if ( im1 == im2 ) {
// merge this
mergeTo[i] = j;
break;
}
}
}
if ( j == this->NumSurfaces() ) {
// didn't merge
mergeTo[i] = j;
surf.shader = im1;
surf.id = this->NumSurfaces();
this->AddSurface( surf );
}
}
}
idVectorSubset<idVec3, 3> vertexSubset;
idVectorSubset<idVec2, 2> texCoordSubset;
// we only ever use the first layer
lwLayer *layer = lwo->layer;
// vertex positions
if ( layer->point.count <= 0 ) {
common->Warning( "ConvertLWOToModelSurfaces: model \'%s\' has bad or missing vertex data", name.c_str() );
return false;
}
vList = (idVec3 *)R_StaticAlloc( layer->point.count * sizeof( vList[0] ) );
for ( j = 0; j < layer->point.count; j++ ) {
vList[j].x = layer->point.pt[j].pos[0];
vList[j].y = layer->point.pt[j].pos[2];
vList[j].z = layer->point.pt[j].pos[1];
}
// vertex texture coords
numTVertexes = 0;
if ( layer->nvmaps ) {
for( lwVMap *vm = layer->vmap; vm; vm = vm->next ) {
if ( vm->type == LWID_('T','X','U','V') ) {
numTVertexes += vm->nverts;
}
}
}
if ( numTVertexes ) {
tvList = (idVec2 *)Mem_Alloc( numTVertexes * sizeof( tvList[0] ) );
int offset = 0;
for( lwVMap *vm = layer->vmap; vm; vm = vm->next ) {
if ( vm->type == LWID_('T','X','U','V') ) {
vm->offset = offset;
for ( k = 0; k < vm->nverts; k++ ) {
tvList[k + offset].x = vm->val[k][0];
tvList[k + offset].y = 1.0f - vm->val[k][1]; // invert the t
}
offset += vm->nverts;
}
}
} else {
common->Warning( "ConvertLWOToModelSurfaces: model \'%s\' has bad or missing uv data", name.c_str() );
numTVertexes = 1;
tvList = (idVec2 *)Mem_ClearedAlloc( numTVertexes * sizeof( tvList[0] ) );
}
// It seems like the tools our artists are using often generate
// verts and texcoords slightly separated that should be merged
// note that we really should combine the surfaces with common materials
// before doing this operation, because we can miss a slop combination
// if they are in different surfaces
vRemap = (int *)R_StaticAlloc( layer->point.count * sizeof( vRemap[0] ) );
if ( fastLoad ) {
// renderbump doesn't care about vertex count
for ( j = 0; j < layer->point.count; j++ ) {
vRemap[j] = j;
}
} else {
float vertexEpsilon = r_slopVertex.GetFloat();
float expand = 2 * 32 * vertexEpsilon;
idVec3 mins, maxs;
SIMDProcessor->MinMax( mins, maxs, vList, layer->point.count );
mins -= idVec3( expand, expand, expand );
maxs += idVec3( expand, expand, expand );
vertexSubset.Init( mins, maxs, 32, 1024 );
for ( j = 0; j < layer->point.count; j++ ) {
vRemap[j] = vertexSubset.FindVector( vList, j, vertexEpsilon );
}
}
tvRemap = (int *)R_StaticAlloc( numTVertexes * sizeof( tvRemap[0] ) );
if ( fastLoad ) {
// renderbump doesn't care about vertex count
for ( j = 0; j < numTVertexes; j++ ) {
tvRemap[j] = j;
}
} else {
float texCoordEpsilon = r_slopTexCoord.GetFloat();
float expand = 2 * 32 * texCoordEpsilon;
idVec2 mins, maxs;
SIMDProcessor->MinMax( mins, maxs, tvList, numTVertexes );
mins -= idVec2( expand, expand );
maxs += idVec2( expand, expand );
texCoordSubset.Init( mins, maxs, 32, 1024 );
for ( j = 0; j < numTVertexes; j++ ) {
tvRemap[j] = texCoordSubset.FindVector( tvList, j, texCoordEpsilon );
}
}
// build the surfaces
for ( lwoSurf = lwo->surf, i = 0; lwoSurf; lwoSurf = lwoSurf->next, i++ ) {
im1 = declManager->FindMaterial( lwoSurf->name );
bool normalsParsed = true;
// completely ignore any explict normals on surfaces with a renderbump command
// which will guarantee the best contours and least vertexes.
const char *rb = im1->GetRenderBump();
if ( rb && rb[0] ) {
normalsParsed = false;
}
// we need to find out how many unique vertex / texcoord combinations there are
// the maximum possible number of combined vertexes is the number of indexes
mvTable = (matchVert_t *)R_ClearedStaticAlloc( layer->polygon.count * 3 * sizeof( mvTable[0] ) );
// we will have a hash chain based on the xyz values
mvHash = (matchVert_t **)R_ClearedStaticAlloc( layer->point.count * sizeof( mvHash[0] ) );
// allocate triangle surface
tri = R_AllocStaticTriSurf();
tri->numVerts = 0;
tri->numIndexes = 0;
R_AllocStaticTriSurfIndexes( tri, layer->polygon.count * 3 );
tri->generateNormals = !normalsParsed;
// find all the unique combinations
float normalEpsilon;
if ( fastLoad ) {
normalEpsilon = 1.0f; // don't merge unless completely exact
} else {
normalEpsilon = 1.0f - r_slopNormal.GetFloat();
}
for ( j = 0; j < layer->polygon.count; j++ ) {
lwPolygon *poly = &layer->polygon.pol[j];
if ( poly->surf != lwoSurf ) {
continue;
}
if ( poly->nverts != 3 ) {
common->Warning( "ConvertLWOToModelSurfaces: model %s has too many verts for a poly! Make sure you triplet it down", name.c_str() );
continue;
}
for ( k = 0; k < 3; k++ ) {
v = vRemap[poly->v[k].index];
normal.x = poly->v[k].norm[0];
normal.y = poly->v[k].norm[2];
normal.z = poly->v[k].norm[1];
// LWO models aren't all that pretty when it comes down to the floating point values they store
normal.FixDegenerateNormal();
tv = 0;
color[0] = lwoSurf->color.rgb[0] * 255;
color[1] = lwoSurf->color.rgb[1] * 255;
color[2] = lwoSurf->color.rgb[2] * 255;
color[3] = 255;
// first set attributes from the vertex
lwPoint *pt = &layer->point.pt[poly->v[k].index];
int nvm;
for ( nvm = 0; nvm < pt->nvmaps; nvm++ ) {
lwVMapPt *vm = &pt->vm[nvm];
if ( vm->vmap->type == LWID_('T','X','U','V') ) {
tv = tvRemap[vm->index + vm->vmap->offset];
}
if ( vm->vmap->type == LWID_('R','G','B','A') ) {
for ( int chan = 0; chan < 4; chan++ ) {
color[chan] = 255 * vm->vmap->val[vm->index][chan];
}
}
}
// then override with polygon attributes
for ( nvm = 0; nvm < poly->v[k].nvmaps; nvm++ ) {
lwVMapPt *vm = &poly->v[k].vm[nvm];
if ( vm->vmap->type == LWID_('T','X','U','V') ) {
tv = tvRemap[vm->index + vm->vmap->offset];
}
if ( vm->vmap->type == LWID_('R','G','B','A') ) {
for ( int chan = 0; chan < 4; chan++ ) {
color[chan] = 255 * vm->vmap->val[vm->index][chan];
}
}
}
// find a matching vert
for ( lastmv = NULL, mv = mvHash[v]; mv != NULL; lastmv = mv, mv = mv->next ) {
if ( mv->tv != tv ) {
continue;
}
if ( *(unsigned *)mv->color != *(unsigned *)color ) {
continue;
}
if ( !normalsParsed ) {
// if we are going to create the normals, just
// matching texcoords is enough
break;
}
if ( mv->normal * normal > normalEpsilon ) {
break; // we already have this one
}
}
if ( !mv ) {
// allocate a new match vert and link to hash chain
mv = &mvTable[ tri->numVerts ];
mv->v = v;
mv->tv = tv;
mv->normal = normal;
*(unsigned *)mv->color = *(unsigned *)color;
mv->next = NULL;
if ( lastmv ) {
lastmv->next = mv;
} else {
mvHash[v] = mv;
}
tri->numVerts++;
}
tri->indexes[tri->numIndexes] = mv - mvTable;
tri->numIndexes++;
}
}
// allocate space for the indexes and copy them
if ( tri->numIndexes > layer->polygon.count * 3 ) {
common->FatalError( "ConvertLWOToModelSurfaces: index miscount in LWO file %s", name.c_str() );
}
if ( tri->numVerts > layer->polygon.count * 3 ) {
common->FatalError( "ConvertLWOToModelSurfaces: vertex miscount in LWO file %s", name.c_str() );
}
// now allocate and generate the combined vertexes
R_AllocStaticTriSurfVerts( tri, tri->numVerts );
for ( j = 0; j < tri->numVerts; j++ ) {
mv = &mvTable[j];
tri->verts[ j ].Clear();
tri->verts[ j ].xyz = vList[ mv->v ];
tri->verts[ j ].st = tvList[ mv->tv ];
tri->verts[ j ].normal = mv->normal;
*(unsigned *)tri->verts[j].color = *(unsigned *)mv->color;
}
R_StaticFree( mvTable );
R_StaticFree( mvHash );
// see if we need to merge with a previous surface of the same material
modelSurf = &this->surfaces[mergeTo[ i ]];
srfTriangles_t *mergeTri = modelSurf->geometry;
if ( !mergeTri ) {
modelSurf->geometry = tri;
} else {
modelSurf->geometry = R_MergeTriangles( mergeTri, tri );
R_FreeStaticTriSurf( tri );
R_FreeStaticTriSurf( mergeTri );
}
}
R_StaticFree( tvRemap );
R_StaticFree( vRemap );
R_StaticFree( tvList );
R_StaticFree( vList );
return true;
}
/*
=================
idRenderModelStatic::ConvertLWOToASE
=================
*/
struct aseModel_s *idRenderModelStatic::ConvertLWOToASE( const struct st_lwObject *obj, const char *fileName ) {
int j, k;
aseModel_t *ase;
if ( !obj ) {
return NULL;
}
// NOTE: using new operator because aseModel_t contains idList class objects
ase = new aseModel_t;
ase->timeStamp = obj->timeStamp;
ase->objects.Resize( obj->nlayers, obj->nlayers );
int materialRef = 0;
for ( lwSurface *surf = obj->surf; surf; surf = surf->next ) {
aseMaterial_t *mat = (aseMaterial_t *)Mem_ClearedAlloc( sizeof( *mat ) );
strcpy( mat->name, surf->name );
mat->uTiling = mat->vTiling = 1;
mat->angle = mat->uOffset = mat->vOffset = 0;
ase->materials.Append( mat );
lwLayer *layer = obj->layer;
aseObject_t *object = (aseObject_t *)Mem_ClearedAlloc( sizeof( *object ) );
object->materialRef = materialRef++;
aseMesh_t *mesh = &object->mesh;
ase->objects.Append( object );
mesh->numFaces = layer->polygon.count;
mesh->numTVFaces = mesh->numFaces;
mesh->faces = (aseFace_t *)Mem_Alloc( mesh->numFaces * sizeof( mesh->faces[0] ) );
mesh->numVertexes = layer->point.count;
mesh->vertexes = (idVec3 *)Mem_Alloc( mesh->numVertexes * sizeof( mesh->vertexes[0] ) );
// vertex positions
if ( layer->point.count <= 0 ) {
common->Warning( "ConvertLWOToASE: model \'%s\' has bad or missing vertex data", name.c_str() );
}
for ( j = 0; j < layer->point.count; j++ ) {
mesh->vertexes[j].x = layer->point.pt[j].pos[0];
mesh->vertexes[j].y = layer->point.pt[j].pos[2];
mesh->vertexes[j].z = layer->point.pt[j].pos[1];
}
// vertex texture coords
mesh->numTVertexes = 0;
if ( layer->nvmaps ) {
for( lwVMap *vm = layer->vmap; vm; vm = vm->next ) {
if ( vm->type == LWID_('T','X','U','V') ) {
mesh->numTVertexes += vm->nverts;
}
}
}
if ( mesh->numTVertexes ) {
mesh->tvertexes = (idVec2 *)Mem_Alloc( mesh->numTVertexes * sizeof( mesh->tvertexes[0] ) );
int offset = 0;
for( lwVMap *vm = layer->vmap; vm; vm = vm->next ) {
if ( vm->type == LWID_('T','X','U','V') ) {
vm->offset = offset;
for ( k = 0; k < vm->nverts; k++ ) {
mesh->tvertexes[k + offset].x = vm->val[k][0];
mesh->tvertexes[k + offset].y = 1.0f - vm->val[k][1]; // invert the t
}
offset += vm->nverts;
}
}
} else {
common->Warning( "ConvertLWOToASE: model \'%s\' has bad or missing uv data", fileName );
mesh->numTVertexes = 1;
mesh->tvertexes = (idVec2 *)Mem_ClearedAlloc( mesh->numTVertexes * sizeof( mesh->tvertexes[0] ) );
}
mesh->normalsParsed = true;
mesh->colorsParsed = true; // because we are falling back to the surface color
// triangles
int faceIndex = 0;
for ( j = 0; j < layer->polygon.count; j++ ) {
lwPolygon *poly = &layer->polygon.pol[j];
if ( poly->surf != surf ) {
continue;
}
if ( poly->nverts != 3 ) {
common->Warning( "ConvertLWOToASE: model %s has too many verts for a poly! Make sure you triplet it down", fileName );
continue;
}
mesh->faces[faceIndex].faceNormal.x = poly->norm[0];
mesh->faces[faceIndex].faceNormal.y = poly->norm[2];
mesh->faces[faceIndex].faceNormal.z = poly->norm[1];
for ( k = 0; k < 3; k++ ) {
mesh->faces[faceIndex].vertexNum[k] = poly->v[k].index;
mesh->faces[faceIndex].vertexNormals[k].x = poly->v[k].norm[0];
mesh->faces[faceIndex].vertexNormals[k].y = poly->v[k].norm[2];
mesh->faces[faceIndex].vertexNormals[k].z = poly->v[k].norm[1];
// complete fallbacks
mesh->faces[faceIndex].tVertexNum[k] = 0;
mesh->faces[faceIndex].vertexColors[k][0] = surf->color.rgb[0] * 255;
mesh->faces[faceIndex].vertexColors[k][1] = surf->color.rgb[1] * 255;
mesh->faces[faceIndex].vertexColors[k][2] = surf->color.rgb[2] * 255;
mesh->faces[faceIndex].vertexColors[k][3] = 255;
// first set attributes from the vertex
lwPoint *pt = &layer->point.pt[poly->v[k].index];
int nvm;
for ( nvm = 0; nvm < pt->nvmaps; nvm++ ) {
lwVMapPt *vm = &pt->vm[nvm];
if ( vm->vmap->type == LWID_('T','X','U','V') ) {
mesh->faces[faceIndex].tVertexNum[k] = vm->index + vm->vmap->offset;
}
if ( vm->vmap->type == LWID_('R','G','B','A') ) {
for ( int chan = 0; chan < 4; chan++ ) {
mesh->faces[faceIndex].vertexColors[k][chan] = 255 * vm->vmap->val[vm->index][chan];
}
}
}
// then override with polygon attributes
for ( nvm = 0; nvm < poly->v[k].nvmaps; nvm++ ) {
lwVMapPt *vm = &poly->v[k].vm[nvm];
if ( vm->vmap->type == LWID_('T','X','U','V') ) {
mesh->faces[faceIndex].tVertexNum[k] = vm->index + vm->vmap->offset;
}
if ( vm->vmap->type == LWID_('R','G','B','A') ) {
for ( int chan = 0; chan < 4; chan++ ) {
mesh->faces[faceIndex].vertexColors[k][chan] = 255 * vm->vmap->val[vm->index][chan];
}
}
}
}
faceIndex++;
}
mesh->numFaces = faceIndex;
mesh->numTVFaces = faceIndex;
aseFace_t *newFaces = ( aseFace_t* )Mem_Alloc( mesh->numFaces * sizeof ( mesh->faces[0] ) );
memcpy( newFaces, mesh->faces, sizeof( mesh->faces[0] ) * mesh->numFaces );
Mem_Free( mesh->faces );
mesh->faces = newFaces;
}
return ase;
}
/*
=================
idRenderModelStatic::ConvertMAToModelSurfaces
=================
*/
bool idRenderModelStatic::ConvertMAToModelSurfaces (const struct maModel_s *ma ) {
maObject_t * object;
maMesh_t * mesh;
maMaterial_t * material;
const idMaterial *im1, *im2;
srfTriangles_t *tri;
int objectNum;
int i, j, k;
int v, tv;
int * vRemap;
int * tvRemap;
matchVert_t * mvTable; // all of the match verts
matchVert_t ** mvHash; // points inside mvTable for each xyz index
matchVert_t * lastmv;
matchVert_t * mv;
idVec3 normal;
float uOffset, vOffset, textureSin, textureCos;
float uTiling, vTiling;
int * mergeTo;
byte * color;
static byte identityColor[4] = { 255, 255, 255, 255 };
modelSurface_t surf, *modelSurf;
if ( !ma ) {
return false;
}
if ( ma->objects.Num() < 1 ) {
return false;
}
timeStamp = ma->timeStamp;
// the modeling programs can save out multiple surfaces with a common
// material, but we would like to mege them together where possible
// meaning that this->NumSurfaces() <= ma->objects.currentElements
mergeTo = (int *)_alloca( ma->objects.Num() * sizeof( *mergeTo ) );
surf.geometry = NULL;
if ( ma->materials.Num() == 0 ) {
// if we don't have any materials, dump everything into a single surface
surf.shader = tr.defaultMaterial;
surf.id = 0;
this->AddSurface( surf );
for ( i = 0 ; i < ma->objects.Num() ; i++ ) {
mergeTo[i] = 0;
}
} else if ( !r_mergeModelSurfaces.GetBool() ) {
// don't merge any
for ( i = 0 ; i < ma->objects.Num() ; i++ ) {
mergeTo[i] = i;
object = ma->objects[i];
if(object->materialRef >= 0) {
material = ma->materials[object->materialRef];
surf.shader = declManager->FindMaterial( material->name );
} else {
surf.shader = tr.defaultMaterial;
}
surf.id = this->NumSurfaces();
this->AddSurface( surf );
}
} else {
// search for material matches
for ( i = 0 ; i < ma->objects.Num() ; i++ ) {
object = ma->objects[i];
if(object->materialRef >= 0) {
material = ma->materials[object->materialRef];
im1 = declManager->FindMaterial( material->name );
} else {
im1 = tr.defaultMaterial;
}
if ( im1->IsDiscrete() ) {
// flares, autosprites, etc
j = this->NumSurfaces();
} else {
for ( j = 0 ; j < this->NumSurfaces() ; j++ ) {
modelSurf = &this->surfaces[j];
im2 = modelSurf->shader;
if ( im1 == im2 ) {
// merge this
mergeTo[i] = j;
break;
}
}
}
if ( j == this->NumSurfaces() ) {
// didn't merge
mergeTo[i] = j;
surf.shader = im1;
surf.id = this->NumSurfaces();
this->AddSurface( surf );
}
}
}
idVectorSubset<idVec3, 3> vertexSubset;
idVectorSubset<idVec2, 2> texCoordSubset;
// build the surfaces
for ( objectNum = 0 ; objectNum < ma->objects.Num() ; objectNum++ ) {
object = ma->objects[objectNum];
mesh = &object->mesh;
if(object->materialRef >= 0) {
material = ma->materials[object->materialRef];
im1 = declManager->FindMaterial( material->name );
} else {
im1 = tr.defaultMaterial;
}
bool normalsParsed = mesh->normalsParsed;
// completely ignore any explict normals on surfaces with a renderbump command
// which will guarantee the best contours and least vertexes.
const char *rb = im1->GetRenderBump();
if ( rb && rb[0] ) {
normalsParsed = false;
}
// It seems like the tools our artists are using often generate
// verts and texcoords slightly separated that should be merged
// note that we really should combine the surfaces with common materials
// before doing this operation, because we can miss a slop combination
// if they are in different surfaces
vRemap = (int *)R_StaticAlloc( mesh->numVertexes * sizeof( vRemap[0] ) );
if ( fastLoad ) {
// renderbump doesn't care about vertex count
for ( j = 0; j < mesh->numVertexes; j++ ) {
vRemap[j] = j;
}
} else {
float vertexEpsilon = r_slopVertex.GetFloat();
float expand = 2 * 32 * vertexEpsilon;
idVec3 mins, maxs;
SIMDProcessor->MinMax( mins, maxs, mesh->vertexes, mesh->numVertexes );
mins -= idVec3( expand, expand, expand );
maxs += idVec3( expand, expand, expand );
vertexSubset.Init( mins, maxs, 32, 1024 );
for ( j = 0; j < mesh->numVertexes; j++ ) {
vRemap[j] = vertexSubset.FindVector( mesh->vertexes, j, vertexEpsilon );
}
}
tvRemap = (int *)R_StaticAlloc( mesh->numTVertexes * sizeof( tvRemap[0] ) );
if ( fastLoad ) {
// renderbump doesn't care about vertex count
for ( j = 0; j < mesh->numTVertexes; j++ ) {
tvRemap[j] = j;
}
} else {
float texCoordEpsilon = r_slopTexCoord.GetFloat();
float expand = 2 * 32 * texCoordEpsilon;
idVec2 mins, maxs;
SIMDProcessor->MinMax( mins, maxs, mesh->tvertexes, mesh->numTVertexes );
mins -= idVec2( expand, expand );
maxs += idVec2( expand, expand );
texCoordSubset.Init( mins, maxs, 32, 1024 );
for ( j = 0; j < mesh->numTVertexes; j++ ) {
tvRemap[j] = texCoordSubset.FindVector( mesh->tvertexes, j, texCoordEpsilon );
}
}
// we need to find out how many unique vertex / texcoord / color combinations
// there are, because MA tracks them separately but we need them unified
// the maximum possible number of combined vertexes is the number of indexes
mvTable = (matchVert_t *)R_ClearedStaticAlloc( mesh->numFaces * 3 * sizeof( mvTable[0] ) );
// we will have a hash chain based on the xyz values
mvHash = (matchVert_t **)R_ClearedStaticAlloc( mesh->numVertexes * sizeof( mvHash[0] ) );
// allocate triangle surface
tri = R_AllocStaticTriSurf();
tri->numVerts = 0;
tri->numIndexes = 0;
R_AllocStaticTriSurfIndexes( tri, mesh->numFaces * 3 );
tri->generateNormals = !normalsParsed;
// init default normal, color and tex coord index
normal.Zero();
color = identityColor;
tv = 0;
// find all the unique combinations
float normalEpsilon = 1.0f - r_slopNormal.GetFloat();
for ( j = 0; j < mesh->numFaces; j++ ) {
for ( k = 0; k < 3; k++ ) {
v = mesh->faces[j].vertexNum[k];
if ( v < 0 || v >= mesh->numVertexes ) {
common->Error( "ConvertMAToModelSurfaces: bad vertex index in MA file %s", name.c_str() );
}
// collapse the position if it was slightly offset
v = vRemap[v];
// we may or may not have texcoords to compare
if ( mesh->numTVertexes != 0 ) {
tv = mesh->faces[j].tVertexNum[k];
if ( tv < 0 || tv >= mesh->numTVertexes ) {
common->Error( "ConvertMAToModelSurfaces: bad tex coord index in MA file %s", name.c_str() );
}
// collapse the tex coord if it was slightly offset
tv = tvRemap[tv];
}
// we may or may not have normals to compare
if ( normalsParsed ) {
normal = mesh->faces[j].vertexNormals[k];
}
//BSM: Todo: Fix the vertex colors
// we may or may not have colors to compare
if ( mesh->faces[j].vertexColors[k] != -1 && mesh->faces[j].vertexColors[k] != -999 ) {
color = &mesh->colors[mesh->faces[j].vertexColors[k]*4];
}
// find a matching vert
for ( lastmv = NULL, mv = mvHash[v]; mv != NULL; lastmv = mv, mv = mv->next ) {
if ( mv->tv != tv ) {
continue;
}
if ( *(unsigned *)mv->color != *(unsigned *)color ) {
continue;
}
if ( !normalsParsed ) {
// if we are going to create the normals, just
// matching texcoords is enough
break;
}
if ( mv->normal * normal > normalEpsilon ) {
break; // we already have this one
}
}
if ( !mv ) {
// allocate a new match vert and link to hash chain
mv = &mvTable[ tri->numVerts ];
mv->v = v;
mv->tv = tv;
mv->normal = normal;
*(unsigned *)mv->color = *(unsigned *)color;
mv->next = NULL;
if ( lastmv ) {
lastmv->next = mv;
} else {
mvHash[v] = mv;
}
tri->numVerts++;
}
tri->indexes[tri->numIndexes] = mv - mvTable;
tri->numIndexes++;
}
}
// allocate space for the indexes and copy them
if ( tri->numIndexes > mesh->numFaces * 3 ) {
common->FatalError( "ConvertMAToModelSurfaces: index miscount in MA file %s", name.c_str() );
}
if ( tri->numVerts > mesh->numFaces * 3 ) {
common->FatalError( "ConvertMAToModelSurfaces: vertex miscount in MA file %s", name.c_str() );
}
// an MA allows the texture coordinates to be scaled, translated, and rotated
//BSM: Todo: Does Maya support this and if so how
//if ( ase->materials.Num() == 0 ) {
uOffset = vOffset = 0.0f;
uTiling = vTiling = 1.0f;
textureSin = 0.0f;
textureCos = 1.0f;
//} else {
// material = ase->materials[object->materialRef];
// uOffset = -material->uOffset;
// vOffset = material->vOffset;
// uTiling = material->uTiling;
// vTiling = material->vTiling;
// textureSin = idMath::Sin( material->angle );
// textureCos = idMath::Cos( material->angle );
//}
// now allocate and generate the combined vertexes
R_AllocStaticTriSurfVerts( tri, tri->numVerts );
for ( j = 0; j < tri->numVerts; j++ ) {
mv = &mvTable[j];
tri->verts[ j ].Clear();
tri->verts[ j ].xyz = mesh->vertexes[ mv->v ];
tri->verts[ j ].normal = mv->normal;
*(unsigned *)tri->verts[j].color = *(unsigned *)mv->color;
if ( mesh->numTVertexes != 0 ) {
const idVec2 &tv = mesh->tvertexes[ mv->tv ];
float u = tv.x * uTiling + uOffset;
float v = tv.y * vTiling + vOffset;
tri->verts[ j ].st[0] = u * textureCos + v * textureSin;
tri->verts[ j ].st[1] = u * -textureSin + v * textureCos;
}
}
R_StaticFree( mvTable );
R_StaticFree( mvHash );
R_StaticFree( tvRemap );
R_StaticFree( vRemap );
// see if we need to merge with a previous surface of the same material
modelSurf = &this->surfaces[mergeTo[ objectNum ]];
srfTriangles_t *mergeTri = modelSurf->geometry;
if ( !mergeTri ) {
modelSurf->geometry = tri;
} else {
modelSurf->geometry = R_MergeTriangles( mergeTri, tri );
R_FreeStaticTriSurf( tri );
R_FreeStaticTriSurf( mergeTri );
}
}
return true;
}
/*
=================
idRenderModelStatic::LoadASE
=================
*/
bool idRenderModelStatic::LoadASE( const char *fileName ) {
aseModel_t *ase;
ase = ASE_Load( fileName );
if ( ase == NULL ) {
return false;
}
ConvertASEToModelSurfaces( ase );
ASE_Free( ase );
return true;
}
/*
=================
idRenderModelStatic::LoadLWO
=================
*/
bool idRenderModelStatic::LoadLWO( const char *fileName ) {
unsigned int failID;
int failPos;
lwObject *lwo;
lwo = lwGetObject( fileName, &failID, &failPos );
if ( lwo == NULL ) {
return false;
}
ConvertLWOToModelSurfaces( lwo );
lwFreeObject( lwo );
return true;
}
/*
=================
idRenderModelStatic::LoadMA
=================
*/
bool idRenderModelStatic::LoadMA( const char *fileName ) {
maModel_t *ma;
ma = MA_Load( fileName );
if ( ma == NULL ) {
return false;
}
ConvertMAToModelSurfaces( ma );
MA_Free( ma );
return true;
}
/*
=================
idRenderModelStatic::LoadFLT
USGS height map data for megaTexture experiments
=================
*/
bool idRenderModelStatic::LoadFLT( const char *fileName ) {
float *data;
int len;
len = fileSystem->ReadFile( fileName, (void **)&data );
if ( len <= 0 ) {
return false;
}
int size = sqrt( len / 4.0f );
// bound the altitudes
float min = 9999999;
float max = -9999999;
for ( int i = 0 ; i < len/4 ; i++ ) {
data[i] = BigFloat( data[i] );
if ( data[i] == -9999 ) {
data[i] = 0; // unscanned areas
}
if ( data[i] < min ) {
min = data[i];
}
if ( data[i] > max ) {
max = data[i];
}
}
#if 1
// write out a gray scale height map
byte *image = (byte *)R_StaticAlloc( len );
byte *image_p = image;
for ( int i = 0 ; i < len/4 ; i++ ) {
float v = ( data[i] - min ) / ( max - min );
image_p[0] =
image_p[1] =
image_p[2] = v * 255;
image_p[3] = 255;
image_p += 4;
}
idStr tgaName = fileName;
tgaName.StripFileExtension();
tgaName += ".tga";
R_WriteTGA( tgaName.c_str(), image, size, size, false );
R_StaticFree( image );
//return false;
#endif
// find the island above sea level
int minX, maxX, minY, maxY;
{
int i;
for ( minX = 0 ; minX < size ; minX++ ) {
for ( i = 0 ; i < size ; i++ ) {
if ( data[i*size + minX] > 1.0 ) {
break;
}
}
if ( i != size ) {
break;
}
}
for ( maxX = size-1 ; maxX > 0 ; maxX-- ) {
for ( i = 0 ; i < size ; i++ ) {
if ( data[i*size + maxX] > 1.0 ) {
break;
}
}
if ( i != size ) {
break;
}
}
for ( minY = 0 ; minY < size ; minY++ ) {
for ( i = 0 ; i < size ; i++ ) {
if ( data[minY*size + i] > 1.0 ) {
break;
}
}
if ( i != size ) {
break;
}
}
for ( maxY = size-1 ; maxY < size ; maxY-- ) {
for ( i = 0 ; i < size ; i++ ) {
if ( data[maxY*size + i] > 1.0 ) {
break;
}
}
if ( i != size ) {
break;
}
}
}
int width = maxX - minX + 1;
int height = maxY - minY + 1;
//width /= 2;
// allocate triangle surface
srfTriangles_t *tri = R_AllocStaticTriSurf();
tri->numVerts = width * height;
tri->numIndexes = (width-1) * (height-1) * 6;
fastLoad = true; // don't do all the sil processing
R_AllocStaticTriSurfIndexes( tri, tri->numIndexes );
R_AllocStaticTriSurfVerts( tri, tri->numVerts );
for ( int i = 0 ; i < height ; i++ ) {
for ( int j = 0; j < width ; j++ ) {
int v = i * width + j;
tri->verts[ v ].Clear();
tri->verts[ v ].xyz[0] = j * 10; // each sample is 10 meters
tri->verts[ v ].xyz[1] = -i * 10;
tri->verts[ v ].xyz[2] = data[(minY+i)*size+minX+j]; // height is in meters
tri->verts[ v ].st[0] = (float) j / (width-1);
tri->verts[ v ].st[1] = 1.0 - ( (float) i / (height-1) );
}
}
for ( int i = 0 ; i < height-1 ; i++ ) {
for ( int j = 0; j < width-1 ; j++ ) {
int v = ( i * (width-1) + j ) * 6;
#if 0
tri->indexes[ v + 0 ] = i * width + j;
tri->indexes[ v + 1 ] = (i+1) * width + j;
tri->indexes[ v + 2 ] = (i+1) * width + j + 1;
tri->indexes[ v + 3 ] = i * width + j;
tri->indexes[ v + 4 ] = (i+1) * width + j + 1;
tri->indexes[ v + 5 ] = i * width + j + 1;
#else
tri->indexes[ v + 0 ] = i * width + j;
tri->indexes[ v + 1 ] = i * width + j + 1;
tri->indexes[ v + 2 ] = (i+1) * width + j + 1;
tri->indexes[ v + 3 ] = i * width + j;
tri->indexes[ v + 4 ] = (i+1) * width + j + 1;
tri->indexes[ v + 5 ] = (i+1) * width + j;
#endif
}
}
fileSystem->FreeFile( data );
modelSurface_t surface;
surface.geometry = tri;
surface.id = 0;
surface.shader = tr.defaultMaterial; // declManager->FindMaterial( "shaderDemos/megaTexture" );
this->AddSurface( surface );
return true;
}
//=============================================================================
/*
================
idRenderModelStatic::PurgeModel
================
*/
void idRenderModelStatic::PurgeModel() {
int i;
modelSurface_t *surf;
for ( i = 0 ; i < surfaces.Num() ; i++ ) {
surf = &surfaces[i];
if ( surf->geometry ) {
R_FreeStaticTriSurf( surf->geometry );
}
}
surfaces.Clear();
purged = true;
}
/*
==============
idRenderModelStatic::FreeVertexCache
We are about to restart the vertex cache, so dump everything
==============
*/
void idRenderModelStatic::FreeVertexCache( void ) {
for ( int j = 0 ; j < surfaces.Num() ; j++ ) {
srfTriangles_t *tri = surfaces[j].geometry;
if ( !tri ) {
continue;
}
if ( tri->ambientCache ) {
vertexCache.Free( tri->ambientCache );
tri->ambientCache = NULL;
}
// static shadows may be present
if ( tri->shadowCache ) {
vertexCache.Free( tri->shadowCache );
tri->shadowCache = NULL;
}
}
}
/*
================
idRenderModelStatic::ReadFromDemoFile
================
*/
void idRenderModelStatic::ReadFromDemoFile( class idDemoFile *f ) {
PurgeModel();
InitEmpty( f->ReadHashString() );
int i, j, numSurfaces;
f->ReadInt( numSurfaces );
for ( i = 0 ; i < numSurfaces ; i++ ) {
modelSurface_t surf;
surf.shader = declManager->FindMaterial( f->ReadHashString() );
srfTriangles_t *tri = R_AllocStaticTriSurf();
f->ReadInt( tri->numIndexes );
R_AllocStaticTriSurfIndexes( tri, tri->numIndexes );
for ( j = 0; j < tri->numIndexes; ++j )
f->ReadInt( (int&)tri->indexes[j] );
f->ReadInt( tri->numVerts );
R_AllocStaticTriSurfVerts( tri, tri->numVerts );
for ( j = 0; j < tri->numVerts; ++j ) {
f->ReadVec3( tri->verts[j].xyz );
f->ReadVec2( tri->verts[j].st );
f->ReadVec3( tri->verts[j].normal );
f->ReadVec3( tri->verts[j].tangents[0] );
f->ReadVec3( tri->verts[j].tangents[1] );
f->ReadUnsignedChar( tri->verts[j].color[0] );
f->ReadUnsignedChar( tri->verts[j].color[1] );
f->ReadUnsignedChar( tri->verts[j].color[2] );
f->ReadUnsignedChar( tri->verts[j].color[3] );
}
surf.geometry = tri;
this->AddSurface( surf );
}
this->FinishSurfaces();
}
/*
================
idRenderModelStatic::WriteToDemoFile
================
*/
void idRenderModelStatic::WriteToDemoFile( class idDemoFile *f ) {
int data[1];
// note that it has been updated
lastArchivedFrame = tr.frameCount;
data[0] = DC_DEFINE_MODEL;
f->WriteInt( data[0] );
f->WriteHashString( this->Name() );
int i, j, iData = surfaces.Num();
f->WriteInt( iData );
for ( i = 0 ; i < surfaces.Num() ; i++ ) {
const modelSurface_t *surf = &surfaces[i];
f->WriteHashString( surf->shader->GetName() );
srfTriangles_t *tri = surf->geometry;
f->WriteInt( tri->numIndexes );
for ( j = 0; j < tri->numIndexes; ++j )
f->WriteInt( (int&)tri->indexes[j] );
f->WriteInt( tri->numVerts );
for ( j = 0; j < tri->numVerts; ++j ) {
f->WriteVec3( tri->verts[j].xyz );
f->WriteVec2( tri->verts[j].st );
f->WriteVec3( tri->verts[j].normal );
f->WriteVec3( tri->verts[j].tangents[0] );
f->WriteVec3( tri->verts[j].tangents[1] );
f->WriteUnsignedChar( tri->verts[j].color[0] );
f->WriteUnsignedChar( tri->verts[j].color[1] );
f->WriteUnsignedChar( tri->verts[j].color[2] );
f->WriteUnsignedChar( tri->verts[j].color[3] );
}
}
}
/*
================
idRenderModelStatic::IsLoaded
================
*/
bool idRenderModelStatic::IsLoaded( void ) {
return !purged;
}
/*
================
idRenderModelStatic::SetLevelLoadReferenced
================
*/
void idRenderModelStatic::SetLevelLoadReferenced( bool referenced ) {
levelLoadReferenced = referenced;
}
/*
================
idRenderModelStatic::IsLevelLoadReferenced
================
*/
bool idRenderModelStatic::IsLevelLoadReferenced( void ) {
return levelLoadReferenced;
}
/*
=================
idRenderModelStatic::TouchData
=================
*/
void idRenderModelStatic::TouchData( void ) {
for ( int i = 0 ; i < surfaces.Num() ; i++ ) {
const modelSurface_t *surf = &surfaces[i];
// re-find the material to make sure it gets added to the
// level keep list
declManager->FindMaterial( surf->shader->GetName() );
}
}
/*
=================
idRenderModelStatic::DeleteSurfaceWithId
=================
*/
bool idRenderModelStatic::DeleteSurfaceWithId( int id ) {
int i;
for ( i = 0; i < surfaces.Num(); i++ ) {
if ( surfaces[i].id == id ) {
R_FreeStaticTriSurf( surfaces[i].geometry );
surfaces.RemoveIndex( i );
return true;
}
}
return false;
}
/*
=================
idRenderModelStatic::DeleteSurfacesWithNegativeId
=================
*/
void idRenderModelStatic::DeleteSurfacesWithNegativeId( void ) {
int i;
for ( i = 0; i < surfaces.Num(); i++ ) {
if ( surfaces[i].id < 0 ) {
R_FreeStaticTriSurf( surfaces[i].geometry );
surfaces.RemoveIndex( i );
i--;
}
}
}
/*
=================
idRenderModelStatic::FindSurfaceWithId
=================
*/
bool idRenderModelStatic::FindSurfaceWithId( int id, int &surfaceNum ) {
int i;
for ( i = 0; i < surfaces.Num(); i++ ) {
if ( surfaces[i].id == id ) {
surfaceNum = i;
return true;
}
}
return false;
}
Reference in a new issue View git blame Copy permalink