ref: bea7b7bf8ccbc2bc41906517079e76fcfb31cb5a
dir: /code/renderer/tr_main.c/
/* =========================================================================== Copyright (C) 1999-2005 Id Software, Inc. This file is part of Quake III Arena source code. Quake III Arena 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 2 of the License, or (at your option) any later version. Quake III Arena 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 Foobar; if not, write to the Free Software Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA =========================================================================== */ // tr_main.c -- main control flow for each frame #include "tr_local.h" trGlobals_t tr; static float s_flipMatrix[16] = { // convert from our coordinate system (looking down X) // to OpenGL's coordinate system (looking down -Z) 0, 0, -1, 0, -1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1 }; refimport_t ri; // entities that will have procedurally generated surfaces will just // point at this for their sorting surface surfaceType_t entitySurface = SF_ENTITY; /* ================= R_CullLocalBox Returns CULL_IN, CULL_CLIP, or CULL_OUT ================= */ int R_CullLocalBox (vec3_t bounds[2]) { int i, j; vec3_t transformed[8]; float dists[8]; vec3_t v; cplane_t *frust; int anyBack; int front, back; if ( r_nocull->integer ) { return CULL_CLIP; } // transform into world space for (i = 0 ; i < 8 ; i++) { v[0] = bounds[i&1][0]; v[1] = bounds[(i>>1)&1][1]; v[2] = bounds[(i>>2)&1][2]; VectorCopy( tr.or.origin, transformed[i] ); VectorMA( transformed[i], v[0], tr.or.axis[0], transformed[i] ); VectorMA( transformed[i], v[1], tr.or.axis[1], transformed[i] ); VectorMA( transformed[i], v[2], tr.or.axis[2], transformed[i] ); } // check against frustum planes anyBack = 0; for (i = 0 ; i < 4 ; i++) { frust = &tr.viewParms.frustum[i]; front = back = 0; for (j = 0 ; j < 8 ; j++) { dists[j] = DotProduct(transformed[j], frust->normal); if ( dists[j] > frust->dist ) { front = 1; if ( back ) { break; // a point is in front } } else { back = 1; } } if ( !front ) { // all points were behind one of the planes return CULL_OUT; } anyBack |= back; } if ( !anyBack ) { return CULL_IN; // completely inside frustum } return CULL_CLIP; // partially clipped } /* ** R_CullLocalPointAndRadius */ int R_CullLocalPointAndRadius( vec3_t pt, float radius ) { vec3_t transformed; R_LocalPointToWorld( pt, transformed ); return R_CullPointAndRadius( transformed, radius ); } /* ** R_CullPointAndRadius */ int R_CullPointAndRadius( vec3_t pt, float radius ) { int i; float dist; cplane_t *frust; qboolean mightBeClipped = qfalse; if ( r_nocull->integer ) { return CULL_CLIP; } // check against frustum planes for (i = 0 ; i < 4 ; i++) { frust = &tr.viewParms.frustum[i]; dist = DotProduct( pt, frust->normal) - frust->dist; if ( dist < -radius ) { return CULL_OUT; } else if ( dist <= radius ) { mightBeClipped = qtrue; } } if ( mightBeClipped ) { return CULL_CLIP; } return CULL_IN; // completely inside frustum } /* ================= R_LocalNormalToWorld ================= */ void R_LocalNormalToWorld (vec3_t local, vec3_t world) { world[0] = local[0] * tr.or.axis[0][0] + local[1] * tr.or.axis[1][0] + local[2] * tr.or.axis[2][0]; world[1] = local[0] * tr.or.axis[0][1] + local[1] * tr.or.axis[1][1] + local[2] * tr.or.axis[2][1]; world[2] = local[0] * tr.or.axis[0][2] + local[1] * tr.or.axis[1][2] + local[2] * tr.or.axis[2][2]; } /* ================= R_LocalPointToWorld ================= */ void R_LocalPointToWorld (vec3_t local, vec3_t world) { world[0] = local[0] * tr.or.axis[0][0] + local[1] * tr.or.axis[1][0] + local[2] * tr.or.axis[2][0] + tr.or.origin[0]; world[1] = local[0] * tr.or.axis[0][1] + local[1] * tr.or.axis[1][1] + local[2] * tr.or.axis[2][1] + tr.or.origin[1]; world[2] = local[0] * tr.or.axis[0][2] + local[1] * tr.or.axis[1][2] + local[2] * tr.or.axis[2][2] + tr.or.origin[2]; } /* ================= R_WorldToLocal ================= */ void R_WorldToLocal (vec3_t world, vec3_t local) { local[0] = DotProduct(world, tr.or.axis[0]); local[1] = DotProduct(world, tr.or.axis[1]); local[2] = DotProduct(world, tr.or.axis[2]); } /* ========================== R_TransformModelToClip ========================== */ void R_TransformModelToClip( const vec3_t src, const float *modelMatrix, const float *projectionMatrix, vec4_t eye, vec4_t dst ) { int i; for ( i = 0 ; i < 4 ; i++ ) { eye[i] = src[0] * modelMatrix[ i + 0 * 4 ] + src[1] * modelMatrix[ i + 1 * 4 ] + src[2] * modelMatrix[ i + 2 * 4 ] + 1 * modelMatrix[ i + 3 * 4 ]; } for ( i = 0 ; i < 4 ; i++ ) { dst[i] = eye[0] * projectionMatrix[ i + 0 * 4 ] + eye[1] * projectionMatrix[ i + 1 * 4 ] + eye[2] * projectionMatrix[ i + 2 * 4 ] + eye[3] * projectionMatrix[ i + 3 * 4 ]; } } /* ========================== R_TransformClipToWindow ========================== */ void R_TransformClipToWindow( const vec4_t clip, const viewParms_t *view, vec4_t normalized, vec4_t window ) { normalized[0] = clip[0] / clip[3]; normalized[1] = clip[1] / clip[3]; normalized[2] = ( clip[2] + clip[3] ) / ( 2 * clip[3] ); window[0] = 0.5f * ( 1.0f + normalized[0] ) * view->viewportWidth; window[1] = 0.5f * ( 1.0f + normalized[1] ) * view->viewportHeight; window[2] = normalized[2]; window[0] = (int) ( window[0] + 0.5 ); window[1] = (int) ( window[1] + 0.5 ); } /* ========================== myGlMultMatrix ========================== */ void myGlMultMatrix( const float *a, const float *b, float *out ) { int i, j; for ( i = 0 ; i < 4 ; i++ ) { for ( j = 0 ; j < 4 ; j++ ) { out[ i * 4 + j ] = a [ i * 4 + 0 ] * b [ 0 * 4 + j ] + a [ i * 4 + 1 ] * b [ 1 * 4 + j ] + a [ i * 4 + 2 ] * b [ 2 * 4 + j ] + a [ i * 4 + 3 ] * b [ 3 * 4 + j ]; } } } /* ================= R_RotateForEntity Generates an orientation for an entity and viewParms Does NOT produce any GL calls Called by both the front end and the back end ================= */ void R_RotateForEntity( const trRefEntity_t *ent, const viewParms_t *viewParms, orientationr_t *or ) { float glMatrix[16]; vec3_t delta; float axisLength; if ( ent->e.reType != RT_MODEL ) { *or = viewParms->world; return; } VectorCopy( ent->e.origin, or->origin ); VectorCopy( ent->e.axis[0], or->axis[0] ); VectorCopy( ent->e.axis[1], or->axis[1] ); VectorCopy( ent->e.axis[2], or->axis[2] ); glMatrix[0] = or->axis[0][0]; glMatrix[4] = or->axis[1][0]; glMatrix[8] = or->axis[2][0]; glMatrix[12] = or->origin[0]; glMatrix[1] = or->axis[0][1]; glMatrix[5] = or->axis[1][1]; glMatrix[9] = or->axis[2][1]; glMatrix[13] = or->origin[1]; glMatrix[2] = or->axis[0][2]; glMatrix[6] = or->axis[1][2]; glMatrix[10] = or->axis[2][2]; glMatrix[14] = or->origin[2]; glMatrix[3] = 0; glMatrix[7] = 0; glMatrix[11] = 0; glMatrix[15] = 1; myGlMultMatrix( glMatrix, viewParms->world.modelMatrix, or->modelMatrix ); // calculate the viewer origin in the model's space // needed for fog, specular, and environment mapping VectorSubtract( viewParms->or.origin, or->origin, delta ); // compensate for scale in the axes if necessary if ( ent->e.nonNormalizedAxes ) { axisLength = VectorLength( ent->e.axis[0] ); if ( !axisLength ) { axisLength = 0; } else { axisLength = 1.0f / axisLength; } } else { axisLength = 1.0f; } or->viewOrigin[0] = DotProduct( delta, or->axis[0] ) * axisLength; or->viewOrigin[1] = DotProduct( delta, or->axis[1] ) * axisLength; or->viewOrigin[2] = DotProduct( delta, or->axis[2] ) * axisLength; } /* ================= R_RotateForViewer Sets up the modelview matrix for a given viewParm ================= */ void R_RotateForViewer (void) { float viewerMatrix[16]; vec3_t origin; Com_Memset (&tr.or, 0, sizeof(tr.or)); tr.or.axis[0][0] = 1; tr.or.axis[1][1] = 1; tr.or.axis[2][2] = 1; VectorCopy (tr.viewParms.or.origin, tr.or.viewOrigin); // transform by the camera placement VectorCopy( tr.viewParms.or.origin, origin ); viewerMatrix[0] = tr.viewParms.or.axis[0][0]; viewerMatrix[4] = tr.viewParms.or.axis[0][1]; viewerMatrix[8] = tr.viewParms.or.axis[0][2]; viewerMatrix[12] = -origin[0] * viewerMatrix[0] + -origin[1] * viewerMatrix[4] + -origin[2] * viewerMatrix[8]; viewerMatrix[1] = tr.viewParms.or.axis[1][0]; viewerMatrix[5] = tr.viewParms.or.axis[1][1]; viewerMatrix[9] = tr.viewParms.or.axis[1][2]; viewerMatrix[13] = -origin[0] * viewerMatrix[1] + -origin[1] * viewerMatrix[5] + -origin[2] * viewerMatrix[9]; viewerMatrix[2] = tr.viewParms.or.axis[2][0]; viewerMatrix[6] = tr.viewParms.or.axis[2][1]; viewerMatrix[10] = tr.viewParms.or.axis[2][2]; viewerMatrix[14] = -origin[0] * viewerMatrix[2] + -origin[1] * viewerMatrix[6] + -origin[2] * viewerMatrix[10]; viewerMatrix[3] = 0; viewerMatrix[7] = 0; viewerMatrix[11] = 0; viewerMatrix[15] = 1; // convert from our coordinate system (looking down X) // to OpenGL's coordinate system (looking down -Z) myGlMultMatrix( viewerMatrix, s_flipMatrix, tr.or.modelMatrix ); tr.viewParms.world = tr.or; } /* ** SetFarClip */ static void SetFarClip( void ) { float farthestCornerDistance = 0; int i; // if not rendering the world (icons, menus, etc) // set a 2k far clip plane if ( tr.refdef.rdflags & RDF_NOWORLDMODEL ) { tr.viewParms.zFar = 2048; return; } // // set far clipping planes dynamically // farthestCornerDistance = 0; for ( i = 0; i < 8; i++ ) { vec3_t v; vec3_t vecTo; float distance; if ( i & 1 ) { v[0] = tr.viewParms.visBounds[0][0]; } else { v[0] = tr.viewParms.visBounds[1][0]; } if ( i & 2 ) { v[1] = tr.viewParms.visBounds[0][1]; } else { v[1] = tr.viewParms.visBounds[1][1]; } if ( i & 4 ) { v[2] = tr.viewParms.visBounds[0][2]; } else { v[2] = tr.viewParms.visBounds[1][2]; } VectorSubtract( v, tr.viewParms.or.origin, vecTo ); distance = vecTo[0] * vecTo[0] + vecTo[1] * vecTo[1] + vecTo[2] * vecTo[2]; if ( distance > farthestCornerDistance ) { farthestCornerDistance = distance; } } tr.viewParms.zFar = sqrt( farthestCornerDistance ); } /* =============== R_SetupProjection =============== */ void R_SetupProjection( void ) { float xmin, xmax, ymin, ymax; float width, height, depth; float zNear, zFar; // dynamically compute far clip plane distance SetFarClip(); // // set up projection matrix // zNear = r_znear->value; zFar = tr.viewParms.zFar; ymax = zNear * tan( tr.refdef.fov_y * M_PI / 360.0f ); ymin = -ymax; xmax = zNear * tan( tr.refdef.fov_x * M_PI / 360.0f ); xmin = -xmax; width = xmax - xmin; height = ymax - ymin; depth = zFar - zNear; tr.viewParms.projectionMatrix[0] = 2 * zNear / width; tr.viewParms.projectionMatrix[4] = 0; tr.viewParms.projectionMatrix[8] = ( xmax + xmin ) / width; // normally 0 tr.viewParms.projectionMatrix[12] = 0; tr.viewParms.projectionMatrix[1] = 0; tr.viewParms.projectionMatrix[5] = 2 * zNear / height; tr.viewParms.projectionMatrix[9] = ( ymax + ymin ) / height; // normally 0 tr.viewParms.projectionMatrix[13] = 0; tr.viewParms.projectionMatrix[2] = 0; tr.viewParms.projectionMatrix[6] = 0; tr.viewParms.projectionMatrix[10] = -( zFar + zNear ) / depth; tr.viewParms.projectionMatrix[14] = -2 * zFar * zNear / depth; tr.viewParms.projectionMatrix[3] = 0; tr.viewParms.projectionMatrix[7] = 0; tr.viewParms.projectionMatrix[11] = -1; tr.viewParms.projectionMatrix[15] = 0; } /* ================= R_SetupFrustum Setup that culling frustum planes for the current view ================= */ void R_SetupFrustum (void) { int i; float xs, xc; float ang; ang = tr.viewParms.fovX / 180 * M_PI * 0.5f; xs = sin( ang ); xc = cos( ang ); VectorScale( tr.viewParms.or.axis[0], xs, tr.viewParms.frustum[0].normal ); VectorMA( tr.viewParms.frustum[0].normal, xc, tr.viewParms.or.axis[1], tr.viewParms.frustum[0].normal ); VectorScale( tr.viewParms.or.axis[0], xs, tr.viewParms.frustum[1].normal ); VectorMA( tr.viewParms.frustum[1].normal, -xc, tr.viewParms.or.axis[1], tr.viewParms.frustum[1].normal ); ang = tr.viewParms.fovY / 180 * M_PI * 0.5f; xs = sin( ang ); xc = cos( ang ); VectorScale( tr.viewParms.or.axis[0], xs, tr.viewParms.frustum[2].normal ); VectorMA( tr.viewParms.frustum[2].normal, xc, tr.viewParms.or.axis[2], tr.viewParms.frustum[2].normal ); VectorScale( tr.viewParms.or.axis[0], xs, tr.viewParms.frustum[3].normal ); VectorMA( tr.viewParms.frustum[3].normal, -xc, tr.viewParms.or.axis[2], tr.viewParms.frustum[3].normal ); for (i=0 ; i<4 ; i++) { tr.viewParms.frustum[i].type = PLANE_NON_AXIAL; tr.viewParms.frustum[i].dist = DotProduct (tr.viewParms.or.origin, tr.viewParms.frustum[i].normal); SetPlaneSignbits( &tr.viewParms.frustum[i] ); } } /* ================= R_MirrorPoint ================= */ void R_MirrorPoint (vec3_t in, orientation_t *surface, orientation_t *camera, vec3_t out) { int i; vec3_t local; vec3_t transformed; float d; VectorSubtract( in, surface->origin, local ); VectorClear( transformed ); for ( i = 0 ; i < 3 ; i++ ) { d = DotProduct(local, surface->axis[i]); VectorMA( transformed, d, camera->axis[i], transformed ); } VectorAdd( transformed, camera->origin, out ); } void R_MirrorVector (vec3_t in, orientation_t *surface, orientation_t *camera, vec3_t out) { int i; float d; VectorClear( out ); for ( i = 0 ; i < 3 ; i++ ) { d = DotProduct(in, surface->axis[i]); VectorMA( out, d, camera->axis[i], out ); } } /* ============= R_PlaneForSurface ============= */ void R_PlaneForSurface (surfaceType_t *surfType, cplane_t *plane) { srfTriangles_t *tri; srfPoly_t *poly; drawVert_t *v1, *v2, *v3; vec4_t plane4; if (!surfType) { Com_Memset (plane, 0, sizeof(*plane)); plane->normal[0] = 1; return; } switch (*surfType) { case SF_FACE: *plane = ((srfSurfaceFace_t *)surfType)->plane; return; case SF_TRIANGLES: tri = (srfTriangles_t *)surfType; v1 = tri->verts + tri->indexes[0]; v2 = tri->verts + tri->indexes[1]; v3 = tri->verts + tri->indexes[2]; PlaneFromPoints( plane4, v1->xyz, v2->xyz, v3->xyz ); VectorCopy( plane4, plane->normal ); plane->dist = plane4[3]; return; case SF_POLY: poly = (srfPoly_t *)surfType; PlaneFromPoints( plane4, poly->verts[0].xyz, poly->verts[1].xyz, poly->verts[2].xyz ); VectorCopy( plane4, plane->normal ); plane->dist = plane4[3]; return; default: Com_Memset (plane, 0, sizeof(*plane)); plane->normal[0] = 1; return; } } /* ================= R_GetPortalOrientation entityNum is the entity that the portal surface is a part of, which may be moving and rotating. Returns qtrue if it should be mirrored ================= */ qboolean R_GetPortalOrientations( drawSurf_t *drawSurf, int entityNum, orientation_t *surface, orientation_t *camera, vec3_t pvsOrigin, qboolean *mirror ) { int i; cplane_t originalPlane, plane; trRefEntity_t *e; float d; vec3_t transformed; // create plane axis for the portal we are seeing R_PlaneForSurface( drawSurf->surface, &originalPlane ); // rotate the plane if necessary if ( entityNum != ENTITYNUM_WORLD ) { tr.currentEntityNum = entityNum; tr.currentEntity = &tr.refdef.entities[entityNum]; // get the orientation of the entity R_RotateForEntity( tr.currentEntity, &tr.viewParms, &tr.or ); // rotate the plane, but keep the non-rotated version for matching // against the portalSurface entities R_LocalNormalToWorld( originalPlane.normal, plane.normal ); plane.dist = originalPlane.dist + DotProduct( plane.normal, tr.or.origin ); // translate the original plane originalPlane.dist = originalPlane.dist + DotProduct( originalPlane.normal, tr.or.origin ); } else { plane = originalPlane; } VectorCopy( plane.normal, surface->axis[0] ); PerpendicularVector( surface->axis[1], surface->axis[0] ); CrossProduct( surface->axis[0], surface->axis[1], surface->axis[2] ); // locate the portal entity closest to this plane. // origin will be the origin of the portal, origin2 will be // the origin of the camera for ( i = 0 ; i < tr.refdef.num_entities ; i++ ) { e = &tr.refdef.entities[i]; if ( e->e.reType != RT_PORTALSURFACE ) { continue; } d = DotProduct( e->e.origin, originalPlane.normal ) - originalPlane.dist; if ( d > 64 || d < -64) { continue; } // get the pvsOrigin from the entity VectorCopy( e->e.oldorigin, pvsOrigin ); // if the entity is just a mirror, don't use as a camera point if ( e->e.oldorigin[0] == e->e.origin[0] && e->e.oldorigin[1] == e->e.origin[1] && e->e.oldorigin[2] == e->e.origin[2] ) { VectorScale( plane.normal, plane.dist, surface->origin ); VectorCopy( surface->origin, camera->origin ); VectorSubtract( vec3_origin, surface->axis[0], camera->axis[0] ); VectorCopy( surface->axis[1], camera->axis[1] ); VectorCopy( surface->axis[2], camera->axis[2] ); *mirror = qtrue; return qtrue; } // project the origin onto the surface plane to get // an origin point we can rotate around d = DotProduct( e->e.origin, plane.normal ) - plane.dist; VectorMA( e->e.origin, -d, surface->axis[0], surface->origin ); // now get the camera origin and orientation VectorCopy( e->e.oldorigin, camera->origin ); AxisCopy( e->e.axis, camera->axis ); VectorSubtract( vec3_origin, camera->axis[0], camera->axis[0] ); VectorSubtract( vec3_origin, camera->axis[1], camera->axis[1] ); // optionally rotate if ( e->e.oldframe ) { // if a speed is specified if ( e->e.frame ) { // continuous rotate d = (tr.refdef.time/1000.0f) * e->e.frame; VectorCopy( camera->axis[1], transformed ); RotatePointAroundVector( camera->axis[1], camera->axis[0], transformed, d ); CrossProduct( camera->axis[0], camera->axis[1], camera->axis[2] ); } else { // bobbing rotate, with skinNum being the rotation offset d = sin( tr.refdef.time * 0.003f ); d = e->e.skinNum + d * 4; VectorCopy( camera->axis[1], transformed ); RotatePointAroundVector( camera->axis[1], camera->axis[0], transformed, d ); CrossProduct( camera->axis[0], camera->axis[1], camera->axis[2] ); } } else if ( e->e.skinNum ) { d = e->e.skinNum; VectorCopy( camera->axis[1], transformed ); RotatePointAroundVector( camera->axis[1], camera->axis[0], transformed, d ); CrossProduct( camera->axis[0], camera->axis[1], camera->axis[2] ); } *mirror = qfalse; return qtrue; } // if we didn't locate a portal entity, don't render anything. // We don't want to just treat it as a mirror, because without a // portal entity the server won't have communicated a proper entity set // in the snapshot // unfortunately, with local movement prediction it is easily possible // to see a surface before the server has communicated the matching // portal surface entity, so we don't want to print anything here... //ri.Printf( PRINT_ALL, "Portal surface without a portal entity\n" ); return qfalse; } static qboolean IsMirror( const drawSurf_t *drawSurf, int entityNum ) { int i; cplane_t originalPlane, plane; trRefEntity_t *e; float d; // create plane axis for the portal we are seeing R_PlaneForSurface( drawSurf->surface, &originalPlane ); // rotate the plane if necessary if ( entityNum != ENTITYNUM_WORLD ) { tr.currentEntityNum = entityNum; tr.currentEntity = &tr.refdef.entities[entityNum]; // get the orientation of the entity R_RotateForEntity( tr.currentEntity, &tr.viewParms, &tr.or ); // rotate the plane, but keep the non-rotated version for matching // against the portalSurface entities R_LocalNormalToWorld( originalPlane.normal, plane.normal ); plane.dist = originalPlane.dist + DotProduct( plane.normal, tr.or.origin ); // translate the original plane originalPlane.dist = originalPlane.dist + DotProduct( originalPlane.normal, tr.or.origin ); } else { plane = originalPlane; } // locate the portal entity closest to this plane. // origin will be the origin of the portal, origin2 will be // the origin of the camera for ( i = 0 ; i < tr.refdef.num_entities ; i++ ) { e = &tr.refdef.entities[i]; if ( e->e.reType != RT_PORTALSURFACE ) { continue; } d = DotProduct( e->e.origin, originalPlane.normal ) - originalPlane.dist; if ( d > 64 || d < -64) { continue; } // if the entity is just a mirror, don't use as a camera point if ( e->e.oldorigin[0] == e->e.origin[0] && e->e.oldorigin[1] == e->e.origin[1] && e->e.oldorigin[2] == e->e.origin[2] ) { return qtrue; } return qfalse; } return qfalse; } /* ** SurfIsOffscreen ** ** Determines if a surface is completely offscreen. */ static qboolean SurfIsOffscreen( const drawSurf_t *drawSurf, vec4_t clipDest[128] ) { float shortest = 100000000; int entityNum; int numTriangles; shader_t *shader; int fogNum; int dlighted; vec4_t clip, eye; int i; unsigned int pointOr = 0; unsigned int pointAnd = (unsigned int)~0; if ( glConfig.smpActive ) { // FIXME! we can't do RB_BeginSurface/RB_EndSurface stuff with smp! return qfalse; } R_RotateForViewer(); R_DecomposeSort( drawSurf->sort, &entityNum, &shader, &fogNum, &dlighted ); RB_BeginSurface( shader, fogNum ); rb_surfaceTable[ *drawSurf->surface ]( drawSurf->surface ); assert( tess.numVertexes < 128 ); for ( i = 0; i < tess.numVertexes; i++ ) { int j; unsigned int pointFlags = 0; R_TransformModelToClip( tess.xyz[i], tr.or.modelMatrix, tr.viewParms.projectionMatrix, eye, clip ); for ( j = 0; j < 3; j++ ) { if ( clip[j] >= clip[3] ) { pointFlags |= (1 << (j*2)); } else if ( clip[j] <= -clip[3] ) { pointFlags |= ( 1 << (j*2+1)); } } pointAnd &= pointFlags; pointOr |= pointFlags; } // trivially reject if ( pointAnd ) { return qtrue; } // determine if this surface is backfaced and also determine the distance // to the nearest vertex so we can cull based on portal range. Culling // based on vertex distance isn't 100% correct (we should be checking for // range to the surface), but it's good enough for the types of portals // we have in the game right now. numTriangles = tess.numIndexes / 3; for ( i = 0; i < tess.numIndexes; i += 3 ) { vec3_t normal; float dot; float len; VectorSubtract( tess.xyz[tess.indexes[i]], tr.viewParms.or.origin, normal ); len = VectorLengthSquared( normal ); // lose the sqrt if ( len < shortest ) { shortest = len; } if ( ( dot = DotProduct( normal, tess.normal[tess.indexes[i]] ) ) >= 0 ) { numTriangles--; } } if ( !numTriangles ) { return qtrue; } // mirrors can early out at this point, since we don't do a fade over distance // with them (although we could) if ( IsMirror( drawSurf, entityNum ) ) { return qfalse; } if ( shortest > (tess.shader->portalRange*tess.shader->portalRange) ) { return qtrue; } return qfalse; } /* ======================== R_MirrorViewBySurface Returns qtrue if another view has been rendered ======================== */ qboolean R_MirrorViewBySurface (drawSurf_t *drawSurf, int entityNum) { vec4_t clipDest[128]; viewParms_t newParms; viewParms_t oldParms; orientation_t surface, camera; // don't recursively mirror if (tr.viewParms.isPortal) { ri.Printf( PRINT_DEVELOPER, "WARNING: recursive mirror/portal found\n" ); return qfalse; } if ( r_noportals->integer || (r_fastsky->integer == 1) ) { return qfalse; } // trivially reject portal/mirror if ( SurfIsOffscreen( drawSurf, clipDest ) ) { return qfalse; } // save old viewParms so we can return to it after the mirror view oldParms = tr.viewParms; newParms = tr.viewParms; newParms.isPortal = qtrue; if ( !R_GetPortalOrientations( drawSurf, entityNum, &surface, &camera, newParms.pvsOrigin, &newParms.isMirror ) ) { return qfalse; // bad portal, no portalentity } R_MirrorPoint (oldParms.or.origin, &surface, &camera, newParms.or.origin ); VectorSubtract( vec3_origin, camera.axis[0], newParms.portalPlane.normal ); newParms.portalPlane.dist = DotProduct( camera.origin, newParms.portalPlane.normal ); R_MirrorVector (oldParms.or.axis[0], &surface, &camera, newParms.or.axis[0]); R_MirrorVector (oldParms.or.axis[1], &surface, &camera, newParms.or.axis[1]); R_MirrorVector (oldParms.or.axis[2], &surface, &camera, newParms.or.axis[2]); // OPTIMIZE: restrict the viewport on the mirrored view // render the mirror view R_RenderView (&newParms); tr.viewParms = oldParms; return qtrue; } /* ================= R_SpriteFogNum See if a sprite is inside a fog volume ================= */ int R_SpriteFogNum( trRefEntity_t *ent ) { int i, j; fog_t *fog; if ( tr.refdef.rdflags & RDF_NOWORLDMODEL ) { return 0; } for ( i = 1 ; i < tr.world->numfogs ; i++ ) { fog = &tr.world->fogs[i]; for ( j = 0 ; j < 3 ; j++ ) { if ( ent->e.origin[j] - ent->e.radius >= fog->bounds[1][j] ) { break; } if ( ent->e.origin[j] + ent->e.radius <= fog->bounds[0][j] ) { break; } } if ( j == 3 ) { return i; } } return 0; } /* ========================================================================================== DRAWSURF SORTING ========================================================================================== */ /* ================= qsort replacement ================= */ #define SWAP_DRAW_SURF(a,b) temp=((int *)a)[0];((int *)a)[0]=((int *)b)[0];((int *)b)[0]=temp; temp=((int *)a)[1];((int *)a)[1]=((int *)b)[1];((int *)b)[1]=temp; /* this parameter defines the cutoff between using quick sort and insertion sort for arrays; arrays with lengths shorter or equal to the below value use insertion sort */ #define CUTOFF 8 /* testing shows that this is good value */ static void shortsort( drawSurf_t *lo, drawSurf_t *hi ) { drawSurf_t *p, *max; int temp; while (hi > lo) { max = lo; for (p = lo + 1; p <= hi; p++ ) { if ( p->sort > max->sort ) { max = p; } } SWAP_DRAW_SURF(max, hi); hi--; } } /* sort the array between lo and hi (inclusive) FIXME: this was lifted and modified from the microsoft lib source... */ void qsortFast ( void *base, unsigned num, unsigned width ) { char *lo, *hi; /* ends of sub-array currently sorting */ char *mid; /* points to middle of subarray */ char *loguy, *higuy; /* traveling pointers for partition step */ unsigned size; /* size of the sub-array */ char *lostk[30], *histk[30]; int stkptr; /* stack for saving sub-array to be processed */ int temp; if ( sizeof(drawSurf_t) != 8 ) { ri.Error( ERR_DROP, "change SWAP_DRAW_SURF macro" ); } /* Note: the number of stack entries required is no more than 1 + log2(size), so 30 is sufficient for any array */ if (num < 2 || width == 0) return; /* nothing to do */ stkptr = 0; /* initialize stack */ lo = base; hi = (char *)base + width * (num-1); /* initialize limits */ /* this entry point is for pseudo-recursion calling: setting lo and hi and jumping to here is like recursion, but stkptr is prserved, locals aren't, so we preserve stuff on the stack */ recurse: size = (hi - lo) / width + 1; /* number of el's to sort */ /* below a certain size, it is faster to use a O(n^2) sorting method */ if (size <= CUTOFF) { shortsort((drawSurf_t *)lo, (drawSurf_t *)hi); } else { /* First we pick a partititioning element. The efficiency of the algorithm demands that we find one that is approximately the median of the values, but also that we select one fast. Using the first one produces bad performace if the array is already sorted, so we use the middle one, which would require a very wierdly arranged array for worst case performance. Testing shows that a median-of-three algorithm does not, in general, increase performance. */ mid = lo + (size / 2) * width; /* find middle element */ SWAP_DRAW_SURF(mid, lo); /* swap it to beginning of array */ /* We now wish to partition the array into three pieces, one consisiting of elements <= partition element, one of elements equal to the parition element, and one of element >= to it. This is done below; comments indicate conditions established at every step. */ loguy = lo; higuy = hi + width; /* Note that higuy decreases and loguy increases on every iteration, so loop must terminate. */ for (;;) { /* lo <= loguy < hi, lo < higuy <= hi + 1, A[i] <= A[lo] for lo <= i <= loguy, A[i] >= A[lo] for higuy <= i <= hi */ do { loguy += width; } while (loguy <= hi && ( ((drawSurf_t *)loguy)->sort <= ((drawSurf_t *)lo)->sort ) ); /* lo < loguy <= hi+1, A[i] <= A[lo] for lo <= i < loguy, either loguy > hi or A[loguy] > A[lo] */ do { higuy -= width; } while (higuy > lo && ( ((drawSurf_t *)higuy)->sort >= ((drawSurf_t *)lo)->sort ) ); /* lo-1 <= higuy <= hi, A[i] >= A[lo] for higuy < i <= hi, either higuy <= lo or A[higuy] < A[lo] */ if (higuy < loguy) break; /* if loguy > hi or higuy <= lo, then we would have exited, so A[loguy] > A[lo], A[higuy] < A[lo], loguy < hi, highy > lo */ SWAP_DRAW_SURF(loguy, higuy); /* A[loguy] < A[lo], A[higuy] > A[lo]; so condition at top of loop is re-established */ } /* A[i] >= A[lo] for higuy < i <= hi, A[i] <= A[lo] for lo <= i < loguy, higuy < loguy, lo <= higuy <= hi implying: A[i] >= A[lo] for loguy <= i <= hi, A[i] <= A[lo] for lo <= i <= higuy, A[i] = A[lo] for higuy < i < loguy */ SWAP_DRAW_SURF(lo, higuy); /* put partition element in place */ /* OK, now we have the following: A[i] >= A[higuy] for loguy <= i <= hi, A[i] <= A[higuy] for lo <= i < higuy A[i] = A[lo] for higuy <= i < loguy */ /* We've finished the partition, now we want to sort the subarrays [lo, higuy-1] and [loguy, hi]. We do the smaller one first to minimize stack usage. We only sort arrays of length 2 or more.*/ if ( higuy - 1 - lo >= hi - loguy ) { if (lo + width < higuy) { lostk[stkptr] = lo; histk[stkptr] = higuy - width; ++stkptr; } /* save big recursion for later */ if (loguy < hi) { lo = loguy; goto recurse; /* do small recursion */ } } else { if (loguy < hi) { lostk[stkptr] = loguy; histk[stkptr] = hi; ++stkptr; /* save big recursion for later */ } if (lo + width < higuy) { hi = higuy - width; goto recurse; /* do small recursion */ } } } /* We have sorted the array, except for any pending sorts on the stack. Check if there are any, and do them. */ --stkptr; if (stkptr >= 0) { lo = lostk[stkptr]; hi = histk[stkptr]; goto recurse; /* pop subarray from stack */ } else return; /* all subarrays done */ } //========================================================================================== /* ================= R_AddDrawSurf ================= */ void R_AddDrawSurf( surfaceType_t *surface, shader_t *shader, int fogIndex, int dlightMap ) { int index; // instead of checking for overflow, we just mask the index // so it wraps around index = tr.refdef.numDrawSurfs & DRAWSURF_MASK; // the sort data is packed into a single 32 bit value so it can be // compared quickly during the qsorting process tr.refdef.drawSurfs[index].sort = (shader->sortedIndex << QSORT_SHADERNUM_SHIFT) | tr.shiftedEntityNum | ( fogIndex << QSORT_FOGNUM_SHIFT ) | (int)dlightMap; tr.refdef.drawSurfs[index].surface = surface; tr.refdef.numDrawSurfs++; } /* ================= R_DecomposeSort ================= */ void R_DecomposeSort( unsigned sort, int *entityNum, shader_t **shader, int *fogNum, int *dlightMap ) { *fogNum = ( sort >> QSORT_FOGNUM_SHIFT ) & 31; *shader = tr.sortedShaders[ ( sort >> QSORT_SHADERNUM_SHIFT ) & (MAX_SHADERS-1) ]; *entityNum = ( sort >> QSORT_ENTITYNUM_SHIFT ) & 1023; *dlightMap = sort & 3; } /* ================= R_SortDrawSurfs ================= */ void R_SortDrawSurfs( drawSurf_t *drawSurfs, int numDrawSurfs ) { shader_t *shader; int fogNum; int entityNum; int dlighted; int i; // it is possible for some views to not have any surfaces if ( numDrawSurfs < 1 ) { // we still need to add it for hyperspace cases R_AddDrawSurfCmd( drawSurfs, numDrawSurfs ); return; } // if we overflowed MAX_DRAWSURFS, the drawsurfs // wrapped around in the buffer and we will be missing // the first surfaces, not the last ones if ( numDrawSurfs > MAX_DRAWSURFS ) { numDrawSurfs = MAX_DRAWSURFS; } // sort the drawsurfs by sort type, then orientation, then shader qsortFast (drawSurfs, numDrawSurfs, sizeof(drawSurf_t) ); // check for any pass through drawing, which // may cause another view to be rendered first for ( i = 0 ; i < numDrawSurfs ; i++ ) { R_DecomposeSort( (drawSurfs+i)->sort, &entityNum, &shader, &fogNum, &dlighted ); if ( shader->sort > SS_PORTAL ) { break; } // no shader should ever have this sort type if ( shader->sort == SS_BAD ) { ri.Error (ERR_DROP, "Shader '%s'with sort == SS_BAD", shader->name ); } // if the mirror was completely clipped away, we may need to check another surface if ( R_MirrorViewBySurface( (drawSurfs+i), entityNum) ) { // this is a debug option to see exactly what is being mirrored if ( r_portalOnly->integer ) { return; } break; // only one mirror view at a time } } R_AddDrawSurfCmd( drawSurfs, numDrawSurfs ); } /* ============= R_AddEntitySurfaces ============= */ void R_AddEntitySurfaces (void) { trRefEntity_t *ent; shader_t *shader; if ( !r_drawentities->integer ) { return; } for ( tr.currentEntityNum = 0; tr.currentEntityNum < tr.refdef.num_entities; tr.currentEntityNum++ ) { ent = tr.currentEntity = &tr.refdef.entities[tr.currentEntityNum]; ent->needDlights = qfalse; // preshift the value we are going to OR into the drawsurf sort tr.shiftedEntityNum = tr.currentEntityNum << QSORT_ENTITYNUM_SHIFT; // // the weapon model must be handled special -- // we don't want the hacked weapon position showing in // mirrors, because the true body position will already be drawn // if ( (ent->e.renderfx & RF_FIRST_PERSON) && tr.viewParms.isPortal) { continue; } // simple generated models, like sprites and beams, are not culled switch ( ent->e.reType ) { case RT_PORTALSURFACE: break; // don't draw anything case RT_SPRITE: case RT_BEAM: case RT_LIGHTNING: case RT_RAIL_CORE: case RT_RAIL_RINGS: // self blood sprites, talk balloons, etc should not be drawn in the primary // view. We can't just do this check for all entities, because md3 // entities may still want to cast shadows from them if ( (ent->e.renderfx & RF_THIRD_PERSON) && !tr.viewParms.isPortal) { continue; } shader = R_GetShaderByHandle( ent->e.customShader ); R_AddDrawSurf( &entitySurface, shader, R_SpriteFogNum( ent ), 0 ); break; case RT_MODEL: // we must set up parts of tr.or for model culling R_RotateForEntity( ent, &tr.viewParms, &tr.or ); tr.currentModel = R_GetModelByHandle( ent->e.hModel ); if (!tr.currentModel) { R_AddDrawSurf( &entitySurface, tr.defaultShader, 0, 0 ); } else { switch ( tr.currentModel->type ) { case MOD_MESH: R_AddMD3Surfaces( ent ); break; case MOD_MD4: R_AddAnimSurfaces( ent ); break; case MOD_BRUSH: R_AddBrushModelSurfaces( ent ); break; case MOD_BAD: // null model axis if ( (ent->e.renderfx & RF_THIRD_PERSON) && !tr.viewParms.isPortal) { break; } shader = R_GetShaderByHandle( ent->e.customShader ); R_AddDrawSurf( &entitySurface, tr.defaultShader, 0, 0 ); break; default: ri.Error( ERR_DROP, "R_AddEntitySurfaces: Bad modeltype" ); break; } } break; default: ri.Error( ERR_DROP, "R_AddEntitySurfaces: Bad reType" ); } } } /* ==================== R_GenerateDrawSurfs ==================== */ void R_GenerateDrawSurfs( void ) { R_AddWorldSurfaces (); R_AddPolygonSurfaces(); // set the projection matrix with the minimum zfar // now that we have the world bounded // this needs to be done before entities are // added, because they use the projection // matrix for lod calculation R_SetupProjection (); R_AddEntitySurfaces (); } /* ================ R_DebugPolygon ================ */ void R_DebugPolygon( int color, int numPoints, float *points ) { int i; GL_State( GLS_DEPTHMASK_TRUE | GLS_SRCBLEND_ONE | GLS_DSTBLEND_ONE ); // draw solid shade qglColor3f( color&1, (color>>1)&1, (color>>2)&1 ); qglBegin( GL_POLYGON ); for ( i = 0 ; i < numPoints ; i++ ) { qglVertex3fv( points + i * 3 ); } qglEnd(); // draw wireframe outline GL_State( GLS_POLYMODE_LINE | GLS_DEPTHMASK_TRUE | GLS_SRCBLEND_ONE | GLS_DSTBLEND_ONE ); qglDepthRange( 0, 0 ); qglColor3f( 1, 1, 1 ); qglBegin( GL_POLYGON ); for ( i = 0 ; i < numPoints ; i++ ) { qglVertex3fv( points + i * 3 ); } qglEnd(); qglDepthRange( 0, 1 ); } /* ==================== R_DebugGraphics Visualization aid for movement clipping debugging ==================== */ void R_DebugGraphics( void ) { if ( !r_debugSurface->integer ) { return; } // the render thread can't make callbacks to the main thread R_SyncRenderThread(); GL_Bind( tr.whiteImage); GL_Cull( CT_FRONT_SIDED ); ri.CM_DrawDebugSurface( R_DebugPolygon ); } /* ================ R_RenderView A view may be either the actual camera view, or a mirror / remote location ================ */ void R_RenderView (viewParms_t *parms) { int firstDrawSurf; if ( parms->viewportWidth <= 0 || parms->viewportHeight <= 0 ) { return; } tr.viewCount++; tr.viewParms = *parms; tr.viewParms.frameSceneNum = tr.frameSceneNum; tr.viewParms.frameCount = tr.frameCount; firstDrawSurf = tr.refdef.numDrawSurfs; tr.viewCount++; // set viewParms.world R_RotateForViewer (); R_SetupFrustum (); R_GenerateDrawSurfs(); R_SortDrawSurfs( tr.refdef.drawSurfs + firstDrawSurf, tr.refdef.numDrawSurfs - firstDrawSurf ); // draw main system development information (surface outlines, etc) R_DebugGraphics(); }