This repository has been archived on 2023-10-09. You can view files and clone it, but cannot push or open issues or pull requests.
Files
blender-archive/source/blender/src/previewrender.c

1248 lines
28 KiB
C
Raw Normal View History

2002-10-12 11:37:38 +00:00
/* previewrender.c GRAPHICS
*
* maart 95
*
* $Id$
*
* ***** BEGIN GPL/BL DUAL LICENSE BLOCK *****
*
* This program 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. The Blender
* Foundation also sells licenses for use in proprietary software under
* the Blender License. See http://www.blender.org/BL/ for information
* about this.
*
* This program 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 this program; if not, write to the Free Software Foundation,
* Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*
* The Original Code is Copyright (C) 2001-2002 by NaN Holding BV.
* All rights reserved.
*
* The Original Code is: all of this file.
*
* Contributor(s): none yet.
*
* ***** END GPL/BL DUAL LICENSE BLOCK *****
*/
/* global includes */
#include <stdlib.h>
#include <math.h>
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
2002-10-12 11:37:38 +00:00
#ifndef WIN32
#include <unistd.h>
#else
#include <io.h>
#endif
#include "MEM_guardedalloc.h"
#include "BLI_arithb.h"
#include "BKE_utildefines.h"
#include "MTC_matrixops.h"
#include "render.h"
#include "mydevice.h"
2002-10-12 11:37:38 +00:00
#include "DNA_texture_types.h"
#include "DNA_world_types.h"
#include "DNA_camera_types.h"
#include "DNA_image_types.h"
#include "DNA_object_types.h"
#include "DNA_lamp_types.h"
#include "DNA_space_types.h"
#include "DNA_scene_types.h"
#include "DNA_screen_types.h"
#include "BKE_global.h"
#include "BKE_image.h"
#include "BKE_texture.h"
#include "BKE_material.h"
#include "BKE_world.h"
#include "BKE_texture.h"
#include "BSE_headerbuttons.h"
2002-10-12 11:37:38 +00:00
#include "BIF_gl.h"
#include "BIF_screen.h"
#include "BIF_space.h" /* allqueue */
#include "BIF_butspace.h"
2002-10-12 11:37:38 +00:00
#include "BIF_drawimage.h" /* rectwrite_part */
#include "BIF_mywindow.h"
#include "BIF_interface.h"
#include "PIL_time.h"
2002-10-12 11:37:38 +00:00
#include "RE_renderconverter.h"
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
#define PR_RECTX 141
#define PR_RECTY 141
#define PR_XMIN 10
#define PR_YMIN 5
#define PR_XMAX 200
#define PR_YMAX 195
2002-10-12 11:37:38 +00:00
#define PR_FACY (PR_YMAX-PR_YMIN-4)/(PR_RECTY)
static rcti prerect;
static int pr_sizex, pr_sizey;
static float pr_facx, pr_facy;
/* implementation */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
static short snijpunt(float *v1, float *v2, float *v3, float *rtlabda, float *ray1, float *ray2)
2002-10-12 11:37:38 +00:00
{
float x0,x1,x2,t00,t01,t02,t10,t11,t12,t20,t21,t22;
float m0,m1,m2,deeldet,det1,det2,det3;
float rtu, rtv;
t00= v3[0]-v1[0];
t01= v3[1]-v1[1];
t02= v3[2]-v1[2];
t10= v3[0]-v2[0];
t11= v3[1]-v2[1];
t12= v3[2]-v2[2];
t20= ray1[0]-ray2[0];
t21= ray1[1]-ray2[1];
t22= ray1[2]-ray2[2];
x0= t11*t22-t12*t21;
x1= t12*t20-t10*t22;
x2= t10*t21-t11*t20;
deeldet= t00*x0+t01*x1+t02*x2;
if(deeldet!=0.0) {
m0= ray1[0]-v3[0];
m1= ray1[1]-v3[1];
m2= ray1[2]-v3[2];
det1= m0*x0+m1*x1+m2*x2;
rtu= det1/deeldet;
if(rtu<=0.0) {
det2= t00*(m1*t22-m2*t21);
det2+= t01*(m2*t20-m0*t22);
det2+= t02*(m0*t21-m1*t20);
rtv= det2/deeldet;
if(rtv<=0.0) {
if(rtu+rtv>= -1.0) {
det3= m0*(t12*t01-t11*t02);
det3+= m1*(t10*t02-t12*t00);
det3+= m2*(t11*t00-t10*t01);
*rtlabda= det3/deeldet;
if(*rtlabda>=0.0 && *rtlabda<=1.0) {
return 1;
}
}
}
}
}
return 0;
}
static float rcubev[7][3]= {
{-0.002055, 6.627364, -3.369742},
{-6.031684, -3.750204, -1.992980},
{-6.049086, 3.817431, 1.969788},
{ 6.031685, 3.833064, 1.992979},
{ 6.049086, -3.734571, -1.969787},
{ 0.002054, -6.544502, 3.369744},
{-0.015348, 1.023131, 7.332510} };
static int rcubi[3][4]= {
{3, 6, 5, 4},
{1, 5, 6, 2},
{3, 0, 2, 6} };
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
static int ray_previewrender(int x, int y, float *vec, float *vn)
2002-10-12 11:37:38 +00:00
{
float scalef= 10.0/100.0;
2002-10-12 11:37:38 +00:00
float ray1[3], ray2[3];
float minlabda, labda;
int totface= 3, hitface= -1;
int a;
ray1[0]= ray2[0]= x*scalef;
ray1[1]= ray2[1]= y*scalef;
ray1[2]= -10.0;
ray2[2]= 10.0;
minlabda= 1.0;
for(a=0; a<totface; a++) {
if(snijpunt( rcubev[rcubi[a][0]], rcubev[rcubi[a][1]], rcubev[rcubi[a][2]], &labda, ray1, ray2)) {
if( labda < minlabda) {
minlabda= labda;
hitface= a;
}
}
if(snijpunt( rcubev[rcubi[a][0]], rcubev[rcubi[a][2]], rcubev[rcubi[a][3]], &labda, ray1, ray2)) {
if( labda < minlabda) {
minlabda= labda;
hitface= a;
}
}
}
if(hitface > -1) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
CalcNormFloat(rcubev[rcubi[hitface][0]], rcubev[rcubi[hitface][1]], rcubev[rcubi[hitface][2]], vn);
2002-10-12 11:37:38 +00:00
vec[0]= (minlabda*(ray1[0]-ray2[0])+ray2[0])/3.7;
vec[1]= (minlabda*(ray1[1]-ray2[1])+ray2[1])/3.7;
vec[2]= (minlabda*(ray1[2]-ray2[2])+ray2[2])/3.7;
return 1;
}
return 0;
}
static unsigned int previewback(int type, int x, int y)
{
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
/* checkerboard, for later
x+= PR_RECTX/2;
y+= PR_RECTX/2;
if( ((x/24) + (y/24)) & 1) return 0x40404040;
else return 0xa0a0a0a0;
*/
2002-10-12 11:37:38 +00:00
if(type & MA_DARK) {
if(abs(x)>abs(y)) return 0;
else return 0x40404040;
}
else {
if(abs(x)>abs(y)) return 0x40404040;
else return 0xa0a0a0a0;
}
}
static void view2d_to_window(int win, int *x_r, int *y_r)
{
int x= *x_r, y= *y_r;
int size[2], origin[2];
float winmat[4][4];
bwin_getsinglematrix(win, winmat);
bwin_getsize(win, &size[0], &size[1]);
bwin_getsuborigin(win, &origin[0], &origin[1]);
*x_r= origin[0] + (size[0]*(0.5 + 0.5*(x*winmat[0][0] + y*winmat[1][0] + winmat[3][0])));
*y_r= origin[1] + (size[1]*(0.5 + 0.5*(x*winmat[0][1] + y*winmat[1][1] + winmat[3][1])));
}
static void set_previewrect(int win, int xmin, int ymin, int xmax, int ymax)
{
prerect.xmin= xmin;
prerect.ymin= ymin;
prerect.xmax= xmax;
prerect.ymax= ymax;
view2d_to_window(win, &prerect.xmin, &prerect.ymin);
view2d_to_window(win, &prerect.xmax, &prerect.ymax);
pr_sizex= (prerect.xmax-prerect.xmin);
pr_sizey= (prerect.ymax-prerect.ymin);
pr_facx= ( (float)pr_sizex-1)/PR_RECTX;
pr_facy= ( (float)pr_sizey-1)/PR_RECTY;
}
static void display_pr_scanline(unsigned int *rect, int recty)
{
static double lasttime= 0;
/* we display 3 new scanlines, one old, the overlap is for wacky 3d cards that cant handle zoom proper */
2002-10-12 11:37:38 +00:00
if(recty % 2) return;
if(recty<2) return;
rect+= (recty-2)*PR_RECTX;
/* enlarge a bit in the y direction, to avoid GL/mesa bug */
2002-10-12 11:37:38 +00:00
glPixelZoom(pr_facx, pr_facy);
glRasterPos2f( (float)PR_XMIN+0.5, 1.0+(float)PR_YMIN + (recty*PR_FACY) );
glDrawPixels(PR_RECTX, 3, GL_RGBA, GL_UNSIGNED_BYTE, rect);
glPixelZoom(1.0, 1.0);
/* flush opengl for cards with frontbuffer slowness */
if(recty==PR_RECTY-1 || (PIL_check_seconds_timer() - lasttime > 0.05)) {
lasttime= PIL_check_seconds_timer();
glFinish();
}
2002-10-12 11:37:38 +00:00
}
static void draw_tex_crop(Tex *tex)
{
rcti rct;
int ret= 0;
if(tex==0) return;
if(tex->type==TEX_IMAGE) {
if(tex->cropxmin==0.0) ret++;
if(tex->cropymin==0.0) ret++;
if(tex->cropxmax==1.0) ret++;
if(tex->cropymax==1.0) ret++;
if(ret==4) return;
rct.xmin= PR_XMIN+2+tex->cropxmin*(PR_XMAX-PR_XMIN-4);
rct.xmax= PR_XMIN+2+tex->cropxmax*(PR_XMAX-PR_XMIN-4);
rct.ymin= PR_YMIN+2+tex->cropymin*(PR_YMAX-PR_YMIN-4);
rct.ymax= PR_YMIN+2+tex->cropymax*(PR_YMAX-PR_YMIN-4);
glPolygonMode(GL_FRONT_AND_BACK, GL_LINE);
glColor3ub(0, 0, 0);
glRecti(rct.xmin+1, rct.ymin-1, rct.xmax+1, rct.ymax-1);
glColor3ub(255, 255, 255);
glRecti(rct.xmin, rct.ymin, rct.xmax, rct.ymax);
glPolygonMode(GL_FRONT_AND_BACK, GL_FILL);
}
}
void BIF_all_preview_changed(void)
{
ScrArea *sa;
SpaceButs *sbuts;
sa= G.curscreen->areabase.first;
while(sa) {
if(sa->spacetype==SPACE_BUTS) {
sbuts= sa->spacedata.first;
sbuts->cury= 0;
addafterqueue(sa->win, RENDERPREVIEW, 1);
}
sa= sa->next;
}
}
2002-10-12 11:37:38 +00:00
void BIF_preview_changed(SpaceButs *sbuts)
{
/* can be called when no buttonswindow visible */
if(sbuts) {
sbuts->cury= 0;
addafterqueue(sbuts->area->win, RENDERPREVIEW, 1);
}
2002-10-12 11:37:38 +00:00
}
/* is supposed to be called with correct panel offset matrix */
void BIF_previewdraw(void)
2002-10-12 11:37:38 +00:00
{
SpaceButs *sbuts= curarea->spacedata.first;
2002-10-12 11:37:38 +00:00
set_previewrect(sbuts->area->win, PR_XMIN, PR_YMIN, PR_XMAX, PR_YMAX);
if (sbuts->rect==0) BIF_preview_changed(sbuts);
else {
2002-10-12 11:37:38 +00:00
int y;
for (y=0; y<PR_RECTY; y++) {
display_pr_scanline(sbuts->rect, y);
}
if (sbuts->mainb==CONTEXT_SHADING && sbuts->tab[CONTEXT_SHADING]==TAB_SHADING_TEX) {
2002-10-12 11:37:38 +00:00
draw_tex_crop(sbuts->lockpoin);
}
}
if(sbuts->cury==0) BIF_preview_changed(sbuts);
2002-10-12 11:37:38 +00:00
}
static void sky_preview_pixel(float lens, int x, int y, char *rect)
{
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
float view[3];
2002-10-12 11:37:38 +00:00
if(R.wrld.skytype & WO_SKYPAPER) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
view[0]= (2*x)/(float)PR_RECTX;
view[1]= (2*y)/(float)PR_RECTY;
view[2]= 0.0;
2002-10-12 11:37:38 +00:00
}
else {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
view[0]= x;
view[1]= y;
view[2]= -lens*PR_RECTX/32.0;
Normalise(view);
2002-10-12 11:37:38 +00:00
}
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
RE_sky(view, rect);
2002-10-12 11:37:38 +00:00
}
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
static void lamp_preview_pixel(ShadeInput *shi, LampRen *la, int x, int y, char *rect)
2002-10-12 11:37:38 +00:00
{
float inpr, i, t, dist, distkw, vec[3];
int col;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi->co[0]= (float)x/(PR_RECTX/4);
shi->co[1]= (float)y/(PR_RECTX/4);
shi->co[2]= 0;
2002-10-12 11:37:38 +00:00
vec[0]= 0.02*x;
vec[1]= 0.02*y;
vec[2]= 0.005*PR_RECTX;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->view, vec);
dist= Normalise(shi->view);
2002-10-12 11:37:38 +00:00
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
if(la->mode & LA_TEXTURE) do_lamp_tex(la, vec, shi);
2002-10-12 11:37:38 +00:00
if(la->type==LA_SUN || la->type==LA_HEMI) {
dist= 1.0;
}
else {
if(la->mode & LA_QUAD) {
t= 1.0;
if(la->ld1>0.0)
t= la->dist/(la->dist+la->ld1*dist);
if(la->ld2>0.0) {
distkw= la->dist*la->dist;
t= t*distkw/(t*distkw+la->ld2*dist*dist);
}
dist= t;
}
else {
dist= (la->dist/(la->dist+dist));
}
}
if(la->type==LA_SPOT) {
if(la->mode & LA_SQUARE) {
/* slightly smaller... */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
inpr= 1.7*cos(MAX2(fabs(shi->view[0]/shi->view[2]) , fabs(shi->view[1]/shi->view[2]) ));
2002-10-12 11:37:38 +00:00
}
else {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
inpr= shi->view[2];
2002-10-12 11:37:38 +00:00
}
t= la->spotsi;
if(inpr<t) dist= 0.0;
else {
t= inpr-t;
if(t<la->spotbl && la->spotbl!=0.0) {
/* soft area */
2002-10-12 11:37:38 +00:00
i= t/la->spotbl;
t= i*i;
i= t*i;
inpr*=(3.0*t-2.0*i);
}
}
dist*=inpr;
}
Area lights and more... - New lamp type added "Area". This uses the radiosity formula (Stoke) to calculate the amount of energy which is received from a plane. Result is very nice local light, which nicely spreads out. - Area lamps have a 'gamma' option to control the light spread - Area lamp builtin sizes: square, rect, cube & box. Only first 2 are implemented. Set a type, and define area size - Button area size won't affect the amount of energy. But scaling the lamp in 3d window will do. This is to cover the case when you scale an entire scene, the light then will remain identical If you just want to change area lamp size, use buttons when you dont want to make the scene too bright or too dark - Since area lights realistically are sensitive for distance (quadratic), the effect it has is quickly too much, or too less. For this the "Dist" value in Lamp can be used. Set it at Dist=10 to have reasonable light on distance 10 Blender units (assumed you didnt scale lamp object). - I tried square sized specularity, but this looked totally weird. Not committed - Plan is to extend area light with 3d dimensions, boxes and cubes. - Note that area light is one-sided, towards negative Z. I need to design a nice drawing method for it. Area Shadow - Since there are a lot of variables associated with soft shadow, they now only are available for Area lights. Allowing spot & normal lamp to have soft shadow is possible though, but will require a reorganisation of the Lamp buttons. Is a point of research & feedback still. - Apart from area size, you now can individually set amount of samples in X and Y direction (for area lamp type 'Rect'). For box type area lamp, this will become 3 dimensions - Area shadows have four options: "Clip circle" : only uses a circular shape of samples, gives smoother results "Dither" : use a 2x2 dither mask "Jitter" : applys a pseudo-random offset to samples "Umbra" : extra emphasis on area that's fully in shadow. Raytrace speedup - improved filling in faces in Octree. Large faces occupied too many nodes - added a coherence check; rays fired sequentially that begin and end in same octree nodes, and that don't intersect, are quickly rejected - rendering shadow scenes benefits from this 20-40%. My statue test monkey file now renders in 19 seconds (was 30). Plus: - adjusted specular max to 511, and made sure Blinn spec has again this incredible small spec size - for UI rounded theme: the color "button" displayed RGB color too dark - fixed countall() function, to also include Subsurf totals - removed setting the 'near' clipping for pressing dot-key numpad - when you press the buttons-window icon for 'Shading Context' the context automaticilly switches as with F5 hotkey Please be warned that this is not a release... settings in files might not work as it did, nor guaranteed to work when we do a release. :)
2003-12-29 16:52:51 +00:00
else if ELEM(la->type, LA_LOCAL, LA_AREA) dist*= shi->view[2];
2002-10-12 11:37:38 +00:00
col= 255.0*dist*la->r;
if(col<=0) rect[0]= 0; else if(col>=255) rect[0]= 255; else rect[0]= col;
col= 255.0*dist*la->g;
if(col<=0) rect[1]= 0; else if(col>=255) rect[1]= 255; else rect[1]= col;
col= 255.0*dist*la->b;
if(col<=0) rect[2]= 0; else if(col>=255) rect[2]= 255; else rect[2]= col;
}
static void init_previewhalo(HaloRen *har, Material *mat)
{
har->type= 0;
if(mat->mode & MA_HALO_XALPHA) har->type |= HA_XALPHA;
har->mat= mat;
har->hard= mat->har;
har->rad= PR_RECTX/2.0;
har->radsq= PR_RECTX*PR_RECTX/4.0;
har->alfa= mat->alpha;
har->add= 255.0*mat->add;
har->r= 255.0*mat->r;
har->g= 255.0*mat->g;
har->b= 255.0*mat->b;
har->xs= PR_RECTX/2.0;
har->ys= PR_RECTX/2.0;
har->zs= har->zd= 0;
har->seed= (mat->seed1 % 256);
if( (mat->mode & MA_HALOTEX) && mat->mtex[0] ) har->tex= 1; else har->tex=0;
if(mat->mode & MA_STAR) har->starpoints= mat->starc; else har->starpoints= 0;
if(mat->mode & MA_HALO_LINES) har->linec= mat->linec; else har->linec= 0;
if(mat->mode & MA_HALO_RINGS) har->ringc= mat->ringc; else har->ringc= 0;
if(mat->mode & MA_HALO_FLARE) har->flarec= mat->flarec; else har->flarec= 0;
if(har->flarec) {
har->xs-= PR_RECTX/3;
har->ys+= PR_RECTX/3;
har->rad*= 0.3;
har->radsq= har->rad*har->rad;
har->pixels= har->rad*har->rad*har->rad;
}
}
static void halo_preview_pixel(HaloRen *har, int startx, int endx, int y, char *rect)
{
float dist, xn, yn, xsq, ysq;
int x;
char front[4];
if(har->flarec) yn= y-PR_RECTX/3;
else yn= y;
ysq= yn*yn;
for(x=startx; x<endx; x++) {
if(har->flarec) xn= x+PR_RECTX/3;
else xn= x;
xsq= xn*xn;
dist= xsq+ysq;
if(dist<har->radsq) {
RE_shadehalo(har, front, 0, dist, xn, yn, har->flarec);
RE_addalphaAddfac(rect, front, har->add);
}
rect+= 4;
}
}
static void previewflare(SpaceButs *sbuts, HaloRen *har, unsigned int *rect)
{
uiBlock *block;
2002-10-12 11:37:38 +00:00
float ycor;
unsigned int *rectot;
int afmx, afmy, rectx, recty;
block= uiFindOpenPanelBlockName(&curarea->uiblocks, "Preview");
if(block==NULL) return;
2002-10-12 11:37:38 +00:00
/* temps */
ycor= R.ycor;
rectx= R.rectx;
recty= R.recty;
afmx= R.afmx;
afmy= R.afmy;
rectot= R.rectot;
R.ycor= 1.0;
R.rectx= PR_RECTX;
R.recty= PR_RECTY;
R.afmx= PR_RECTX/2;
R.afmy= PR_RECTY/2;
R.rectot= rect;
waitcursor(1);
RE_renderflare(har);
waitcursor(0);
// not sure why, either waitcursor or renderflare screws up
areawinset(curarea->win);
uiPanelPush(block);
BIF_previewdraw();
uiPanelPop(block);
2002-10-12 11:37:38 +00:00
/* temps */
R.ycor= ycor;
R.rectx= rectx;
R.recty= recty;
R.afmx= afmx;
R.afmy= afmy;
R.rectot= rectot;
}
extern float Tin, Tr, Tg, Tb, Ta; /* texture.c */
static void texture_preview_pixel(Tex *tex, int x, int y, char *rect)
{
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
float i, v1, xsq, ysq, texvec[3], dummy[3];
2002-10-12 11:37:38 +00:00
int rgbnor, tracol, skip=0;
if(tex->type==TEX_IMAGE) {
v1= 1.0/PR_RECTX;
texvec[0]= 0.5+v1*x;
texvec[1]= 0.5+v1*y;
/* no coordinate mapping, exception: repeat */
2002-10-12 11:37:38 +00:00
if(tex->xrepeat>1) {
texvec[0] *= tex->xrepeat;
if(texvec[0]>1.0) texvec[0] -= (int)(texvec[0]);
}
if(tex->yrepeat>1) {
texvec[1] *= tex->yrepeat;
if(texvec[1]>1.0) texvec[1] -= (int)(texvec[1]);
}
}
else if(tex->type==TEX_ENVMAP) {
if(tex->env) {
ysq= y*y;
xsq= x*x;
if(xsq+ysq < (PR_RECTX/2)*(PR_RECTY/2)) {
texvec[2]= sqrt( (float)((PR_RECTX/2)*(PR_RECTY/2)-xsq-ysq) );
texvec[0]= -x;
texvec[1]= -y;
Normalise(texvec);
i= 2.0*(texvec[2]);
texvec[0]= (i*texvec[0]);
texvec[1]= (i*texvec[1]);
texvec[2]= (-1.0+i*texvec[2]);
}
else {
skip= 1;
Ta= 0.0;
}
}
else {
skip= 1;
Ta= 0.0;
}
}
else {
v1= 2.0/PR_RECTX;
texvec[0]= v1*x;
texvec[1]= v1*y;
texvec[2]= 0.0;
}
/* does not return Tin */
2002-10-12 11:37:38 +00:00
if(tex->type==TEX_STUCCI) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
tex->nor= dummy;
dummy[0]= 1.0;
dummy[1]= dummy[2]= 0.0;
2002-10-12 11:37:38 +00:00
}
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
if(skip==0) rgbnor= multitex(tex, texvec, NULL, NULL, 0);
2002-10-12 11:37:38 +00:00
else rgbnor= 1;
if(rgbnor & 1) {
rect[0]= 255.0*Tr;
rect[1]= 255.0*Tg;
rect[2]= 255.0*Tb;
if(Ta!=1.0) {
tracol= 64+100*(abs(x)>abs(y));
tracol= (1.0-Ta)*tracol;
rect[0]= tracol+ (rect[0]*Ta) ;
rect[1]= tracol+ (rect[1]*Ta) ;
rect[2]= tracol+ (rect[2]*Ta) ;
}
}
else {
if(tex->type==TEX_STUCCI) {
Tin= 0.5 + 0.7*tex->nor[0];
CLAMP(Tin, 0.0, 1.0);
}
rect[0]= 255.0*Tin;
rect[1]= 255.0*Tin;
rect[2]= 255.0*Tin;
}
}
static float pr1_lamp[3]= {2.3, -2.4, -4.6};
static float pr2_lamp[3]= {-8.8, -5.6, -1.5};
static float pr1_col[3]= {0.8, 0.8, 0.8};
static float pr2_col[3]= {0.5, 0.6, 0.7};
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
static void refraction_prv(int *x, int *y, float *n, float index)
{
float dot, fac, view[3], len;
index= 1.0/index;
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
view[0]= index*(float)*x;
view[1]= ((float)*y)/index;
view[2]= 20.0;
len= Normalise(view);
dot= view[0]*n[0] + view[1]*n[1] + view[2]*n[2];
if(dot>0.0) {
fac= 1.0 - (1.0 - dot*dot)*index*index;
if(fac<= 0.0) return;
fac= -dot*index + sqrt(fac);
}
else {
index = 1.0/index;
fac= 1.0 - (1.0 - dot*dot)*index*index;
if(fac<= 0.0) return;
fac= -dot*index - sqrt(fac);
}
*x= (int)(len*(index*view[0] + fac*n[0]));
*y= (int)(len*(index*view[1] + fac*n[1]));
}
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
static void shade_preview_pixel(ShadeInput *shi, float *vec, int x, int y,char *rect, int smooth)
2002-10-12 11:37:38 +00:00
{
extern float fresnel_fac(float *view, float *vn, float ior, float fac);
2002-10-12 11:37:38 +00:00
Material *mat;
float v1,inp, inprspec=0, isr=0.0, isb=0.0, isg=0.0;
float ir=0.0, ib=0.0, ig=0.0;
float view[3], lv[3], *la, alpha;
float eul[3], tmat[3][3], imat[3][3];
int temp, a;
char tracol;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
mat= shi->matren;
2002-10-12 11:37:38 +00:00
v1= 1.0/PR_RECTX;
view[0]= v1*x;
view[1]= v1*y;
view[2]= 1.0;
Normalise(view);
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi->refcol[0]= shi->refcol[1]= shi->refcol[2]= shi->refcol[3]= 0.0;
2002-10-12 11:37:38 +00:00
/* texture handling */
2002-10-12 11:37:38 +00:00
if(mat->texco) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->lo, vec);
2002-10-12 11:37:38 +00:00
if(mat->pr_type==MA_CUBE) {
eul[0]= (297)*M_PI/180.0;
eul[1]= 0.0;
eul[2]= (45)*M_PI/180.0;
EulToMat3(eul, tmat);
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
MTC_Mat3MulVecfl(tmat, shi->lo);
MTC_Mat3MulVecfl(tmat, shi->vn);
2002-10-12 11:37:38 +00:00
/* hack for cubemap, why!!! */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
SWAP(float, shi->vn[0], shi->vn[1]);
2002-10-12 11:37:38 +00:00
}
/* textures otherwise upside down */
if(mat->pr_type==MA_CUBE || mat->pr_type==MA_SPHERE)
shi->lo[2]= -shi->lo[2];
2002-10-12 11:37:38 +00:00
if(mat->texco & TEXCO_GLOB) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->gl, shi->lo);
2002-10-12 11:37:38 +00:00
}
if(mat->texco & TEXCO_WINDOW) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->winco, shi->lo);
2002-10-12 11:37:38 +00:00
}
if(mat->texco & TEXCO_STICKY) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->sticky, shi->lo);
2002-10-12 11:37:38 +00:00
}
if(mat->texco & TEXCO_UV) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->uv, shi->lo);
2002-10-12 11:37:38 +00:00
}
if(mat->texco & TEXCO_OBJECT) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->co, shi->lo);
2002-10-12 11:37:38 +00:00
}
if(mat->texco & TEXCO_NORM) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi->orn[0]= shi->vn[0];
shi->orn[1]= shi->vn[1];
shi->orn[2]= shi->vn[2];
2002-10-12 11:37:38 +00:00
}
if(mat->texco & TEXCO_REFL) {
/* for bump texture */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
VECCOPY(shi->view, view);
2002-10-12 11:37:38 +00:00
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
inp= -2.0*(shi->vn[0]*view[0]+shi->vn[1]*view[1]+shi->vn[2]*view[2]);
shi->ref[0]= (view[0]+inp*shi->vn[0]);
shi->ref[1]= (view[1]+inp*shi->vn[1]);
shi->ref[2]= (view[2]+inp*shi->vn[2]);
2002-10-12 11:37:38 +00:00
}
/* Clear displase vec for preview */
shi->displace[0]= shi->displace[1]= shi->displace[2]= 0.0;
/* normals flipped in render... */
if(mat->mapto & MAP_NORM) VecMulf(shi->vn, -1.0);
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
do_material_tex(shi);
/* normals flipped in render... */
if(mat->mapto & MAP_NORM) VecMulf(shi->vn, -1.0);
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
if(mat->texco & TEXCO_REFL) {
/* normals in render are pointing different... rhm */
if(smooth) shi->ref[1]= -shi->ref[1];
}
2002-10-12 11:37:38 +00:00
if(mat->pr_type==MA_CUBE) {
/* rotate normal back for normals texture */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
SWAP(float, shi->vn[0], shi->vn[1]);
2002-10-12 11:37:38 +00:00
MTC_Mat3Inv(imat, tmat);
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
MTC_Mat3MulVecfl(imat, shi->vn);
2002-10-12 11:37:38 +00:00
}
}
/* set it here, because ray_mirror will affect it */
alpha= mat->alpha;
if(mat->mapto & MAP_DISPLACE) { /* Quick hack of fake displacement preview */
shi->vn[0]-=2.0*shi->displace[2];
shi->vn[1]-=2.0*shi->displace[0];
shi->vn[2]+=2.0*shi->displace[1];
Normalise(shi->vn);
}
if(mat->mode & (MA_ZTRA|MA_RAYTRANSP))
if(mat->fresnel_tra!=0.0)
alpha*= fresnel_fac(view, shi->vn, mat->fresnel_tra_i, mat->fresnel_tra);
2002-10-12 11:37:38 +00:00
if(mat->mode & MA_SHLESS) {
temp= 255.0*(mat->r);
if(temp>255) rect[0]= 255; else if(temp<0) rect[0]= 0; else rect[0]= temp;
temp= 255.0*(mat->g);
if(temp>255) rect[1]= 255; else if(temp<0) rect[1]= 0; else rect[1]= temp;
temp= 255.0*(mat->b);
if(temp>255) rect[2]= 255; else if(temp<0) rect[2]= 0; else rect[2]= temp;
}
else {
for(a=0; a<2; a++) {
if(a==0) la= pr1_lamp;
else la= pr2_lamp;
lv[0]= vec[0]-la[0];
lv[1]= vec[1]-la[1];
lv[2]= vec[2]-la[2];
Normalise(lv);
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
inp= shi->vn[0]*lv[0]+shi->vn[1]*lv[1]+shi->vn[2]*lv[2];
2002-10-12 11:37:38 +00:00
if(inp<0.0) inp= 0.0;
if(mat->spec) {
if(inp>0.0) {
/* specular shaders */
float specfac;
if(mat->spec_shader==MA_SPEC_PHONG)
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
specfac= Phong_Spec(shi->vn, lv, view, mat->har);
else if(mat->spec_shader==MA_SPEC_COOKTORR)
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
specfac= CookTorr_Spec(shi->vn, lv, view, mat->har);
else if(mat->spec_shader==MA_SPEC_BLINN)
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
specfac= Blinn_Spec(shi->vn, lv, view, mat->refrac, (float)mat->har);
else
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
specfac= Toon_Spec(shi->vn, lv, view, mat->param[2], mat->param[3]);
inprspec= specfac*mat->spec;
isr+= inprspec*mat->specr;
isg+= inprspec*mat->specg;
isb+= inprspec*mat->specb;
2002-10-12 11:37:38 +00:00
}
}
/* diffuse shaders */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
if(mat->diff_shader==MA_DIFF_ORENNAYAR) inp= OrenNayar_Diff(shi->vn, lv, view, mat->roughness);
else if(mat->diff_shader==MA_DIFF_TOON) inp= Toon_Diff(shi->vn, lv, view, mat->param[0], mat->param[1]);
// else Lambert
2002-10-12 11:37:38 +00:00
inp= (mat->ref*inp + mat->emit);
if(a==0) la= pr1_col;
else la= pr2_col;
ir+= inp*la[0];
ig+= inp*la[1];
ib+= inp*la[2];
}
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
/* drawing checkerboard and sky */
if(mat->mode & MA_RAYMIRROR) {
float col, div, y, z;
int fac;
/* rotate a bit in x */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
y= shi->ref[1]; z= shi->ref[2];
shi->ref[1]= 0.98*y - 0.17*z;
shi->ref[2]= 0.17*y + 0.98*z;
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
/* scale */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
div= (0.85*shi->ref[1]);
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
shi->refcol[0]= mat->ray_mirror*fresnel_fac(view, shi->vn, mat->fresnel_mir_i, mat->fresnel_mir);
/* not real 'alpha', but mirror overriding transparency */
if(mat->mode & MA_RAYTRANSP) {
float fac= sqrt(shi->refcol[0]);
alpha= alpha*(1.0-fac) + fac;
}
else alpha= alpha*(1.0-shi->refcol[0]) + shi->refcol[0];
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
if(div<0.0) {
/* minus 0.5 prevents too many small tiles in distance */
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
fac= (int)(shi->ref[0]/(div-0.1) ) + (int)(shi->ref[2]/(div-0.1) );
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
if(fac & 1) col= 0.8;
else col= 0.3;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi->refcol[1]= shi->refcol[0]*col;
shi->refcol[2]= shi->refcol[1];
shi->refcol[3]= shi->refcol[2];
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
}
else {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi->refcol[1]= 0.0;
shi->refcol[2]= shi->refcol[0]*0.3*div;
shi->refcol[3]= shi->refcol[0]*0.8*div;
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
}
}
2002-10-12 11:37:38 +00:00
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
if(shi->refcol[0]==0.0) {
2002-10-12 11:37:38 +00:00
a= 255.0*( mat->r*ir +mat->ambr +isr);
if(a>255) a=255; else if(a<0) a= 0;
rect[0]= a;
a= 255.0*(mat->g*ig +mat->ambg +isg);
if(a>255) a=255; else if(a<0) a= 0;
rect[1]= a;
a= 255*(mat->b*ib +mat->ambb +isb);
if(a>255) a=255; else if(a<0) a= 0;
rect[2]= a;
}
else {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
a= 255.0*( mat->mirr*shi->refcol[1] + (1.0 - mat->mirr*shi->refcol[0])*(mat->r*ir +mat->ambr) +isr);
2002-10-12 11:37:38 +00:00
if(a>255) a=255; else if(a<0) a= 0;
rect[0]= a;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
a= 255.0*( mat->mirg*shi->refcol[2] + (1.0 - mat->mirg*shi->refcol[0])*(mat->g*ig +mat->ambg) +isg);
2002-10-12 11:37:38 +00:00
if(a>255) a=255; else if(a<0) a= 0;
rect[1]= a;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
a= 255.0*( mat->mirb*shi->refcol[3] + (1.0 - mat->mirb*shi->refcol[0])*(mat->b*ib +mat->ambb) +isb);
2002-10-12 11:37:38 +00:00
if(a>255) a=255; else if(a<0) a= 0;
rect[2]= a;
}
}
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
/* ztra shade */
if(mat->spectra!=0.0) {
inp = MAX3(isr, isg, isb);
inp *= mat->spectra;
if(inp>1.0) inp= 1.0;
alpha= (1.0-inp)*alpha+inp;
}
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
if(alpha!=1.0) {
if(mat->mode & MA_RAYTRANSP) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
refraction_prv(&x, &y, shi->vn, mat->ang);
2002-10-12 11:37:38 +00:00
}
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
2002-10-12 11:37:38 +00:00
tracol= previewback(mat->pr_back, x, y) & 255;
tracol= (1.0-alpha)*tracol;
rect[0]= tracol+ (rect[0]*alpha) ;
rect[1]= tracol+ (rect[1]*alpha) ;
rect[2]= tracol+ (rect[2]*alpha) ;
}
}
void BIF_previewrender(SpaceButs *sbuts)
{
ID *id, *idfrom;
Material *mat= NULL;
Tex *tex= NULL;
Lamp *la= NULL;
World *wrld= NULL;
LampRen *lar= NULL;
2002-10-12 11:37:38 +00:00
Image *ima;
HaloRen har;
Object *ob;
uiBlock *block;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
ShadeInput shi;
2002-10-12 11:37:38 +00:00
float lens = 0.0, vec[3];
int x, y, starty, startx, endy, endx, radsq, xsq, ysq, last = 0;
unsigned int *rect;
if(sbuts->cury>=PR_RECTY) return;
/* we safely assume curarea has panel "preview" */
/* quick hack for now, later on preview should become uiBlock itself */
block= uiFindOpenPanelBlockName(&curarea->uiblocks, "Preview");
if(block==NULL) return;
ob= ((G.scene->basact)? (G.scene->basact)->object: 0);
/* we cant trust this global lockpoin.. for example with headerless window */
buttons_active_id(&id, &idfrom);
G.buts->lockpoin= id;
if(sbuts->mainb==CONTEXT_SHADING) {
int tab= sbuts->tab[CONTEXT_SHADING];
if(tab==TAB_SHADING_MAT)
mat= sbuts->lockpoin;
else if(tab==TAB_SHADING_TEX)
tex= sbuts->lockpoin;
else if(tab==TAB_SHADING_LAMP) {
if(ob && ob->type==OB_LAMP) la= ob->data;
}
else if(tab==TAB_SHADING_WORLD)
wrld= sbuts->lockpoin;
}
else if(sbuts->mainb==CONTEXT_OBJECT) {
if(ob && ob->type==OB_LAMP) la= ob->data;
}
2002-10-12 11:37:38 +00:00
if(mat==NULL && tex==NULL && la==NULL && wrld==NULL) return;
2002-10-12 11:37:38 +00:00
har.flarec= 0; /* below is a test for postrender flare */
2002-10-12 11:37:38 +00:00
if(qtest()) {
addafterqueue(curarea->win, RENDERPREVIEW, 1);
return;
}
MTC_Mat4One(R.viewmat);
MTC_Mat4One(R.viewinv);
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi.osatex= 0;
Another commit for raytracing, now with glass refraction & fresnel! Changelog: - enable refraction with button "Ray Transp" in Material buttons. - set "Angular Index" value for amount of refraction. - use the "Alpha" value to define transparency. - remember to set a higher "Depth" too... glass can bounce quite some more than expected. - for correct refraction, 3D models MUST have normals pointing in the right direction (consistently pointing outside). - refraction 'sees' the thickness of glass based on what you model. So make for realistic glass both sides of a surface. - I needed to do some rewriting for correct mirroring/refraction, especially to prevent specularity being blended away. Solved this with localizing shading results in the rendercore.c. Now specularity correctly is added, and reduces the 'mirror' value. - Localizing more parts of the render code is being planned. The old render heavily relies on struct Render and struct Osa to store globals. For scanline render no problem, but recursive raytracing dislikes that. - done test with gamma-corrected summation of colors during tracing, is commented out still. But this will give more balanced reflections. Now dark reflections that are reflected in a bright surface seem incorrect. - Introduced 'Fresnel' effect for Mirror and Transparency. This influences the amount of mirror/transparency based at viewing angle. Next to a new Fresnel slider, also a 'falloff' button has been added to define the way it spreads. - Fresnel also works for Ztransp rendering - created new Panel for Raytrace options I have to evaluate still where it all should be logically located. - material preview shows fake reflection and fake refraction as well.
2003-12-16 14:12:01 +00:00
if(mat) {
2002-10-12 11:37:38 +00:00
/* rendervars */
init_render_world();
init_render_material(mat);
/* clear imats */
for(x=0; x<8; x++) {
if(mat->mtex[x]) {
if(mat->mtex[x]->tex) {
init_render_texture(mat->mtex[x]->tex);
if(mat->mtex[x]->tex->env && mat->mtex[x]->tex->env->object)
MTC_Mat4One(mat->mtex[x]->tex->env->object->imat);
}
if(mat->mtex[x]->object) MTC_Mat4One(mat->mtex[x]->object->imat);
if(mat->mtex[x]->object) MTC_Mat4One(mat->mtex[x]->object->imat);
}
}
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi.vlr= 0;
shi.mat= mat;
shi.matren= mat->ren;
2002-10-12 11:37:38 +00:00
if(mat->mode & MA_HALO) init_previewhalo(&har, mat);
}
else if(tex) {
2002-10-12 11:37:38 +00:00
ima= tex->ima;
if(ima) last= ima->lastframe;
init_render_texture(tex);
free_unused_animimages();
if(tex->ima) {
if(tex->ima!=ima) allqueue(REDRAWBUTSSHADING, 0);
else if(last!=ima->lastframe) allqueue(REDRAWBUTSSHADING, 0);
2002-10-12 11:37:38 +00:00
}
if(tex->env && tex->env->object)
MTC_Mat4Invert(tex->env->object->imat, tex->env->object->obmat);
}
else if(la) {
2002-10-12 11:37:38 +00:00
init_render_world();
init_render_textures(); /* do not do it twice!! (brightness) */
2002-10-12 11:37:38 +00:00
R.totlamp= 0;
RE_add_render_lamp(ob, 0); /* 0=no shadbuf */
lar= R.la[0];
/* exceptions: */
2002-10-12 11:37:38 +00:00
lar->spottexfac= 1.0;
lar->spotsi= cos( M_PI/3.0 );
lar->spotbl= (1.0-lar->spotsi)*la->spotblend;
MTC_Mat3One(lar->imat);
}
else if(wrld) {
2002-10-12 11:37:38 +00:00
lens= 35.0;
if(G.scene->camera) {
lens= ( (Camera *)G.scene->camera->data)->lens;
}
init_render_world();
init_render_textures(); /* dont do it twice!! (brightness) */
}
set_previewrect(sbuts->area->win, PR_XMIN, PR_YMIN, PR_XMAX, PR_YMAX);
if(sbuts->rect==0) {
sbuts->rect= MEM_callocN(sizeof(int)*PR_RECTX*PR_RECTY, "butsrect");
/* built in emboss */
rect= sbuts->rect;
for(y=0; y<PR_RECTY; y++, rect++) *rect= 0xFFFFFFFF;
rect= sbuts->rect + PR_RECTX-1;
for(y=0; y<PR_RECTY; y++, rect+=PR_RECTX) *rect= 0xFFFFFFFF;
}
starty= -PR_RECTY/2;
endy= starty+PR_RECTY;
starty+= sbuts->cury;
/* offset +1 for emboss */
startx= -PR_RECTX/2 +1;
endx= startx+PR_RECTX -2;
radsq= (PR_RECTX/2)*(PR_RECTY/2);
if(mat) {
if(mat->pr_type==MA_SPHERE) {
pr1_lamp[0]= 2.3; pr1_lamp[1]= -2.4; pr1_lamp[2]= -4.6;
pr2_lamp[0]= -8.8; pr2_lamp[1]= -5.6; pr2_lamp[2]= -1.5;
}
else {
pr1_lamp[0]= 1.9; pr1_lamp[1]= 3.1; pr1_lamp[2]= -8.5;
pr2_lamp[0]= 1.2; pr2_lamp[1]= -18; pr2_lamp[2]= 3.2;
}
}
/* here it starts! */
glDrawBuffer(GL_FRONT);
uiPanelPush(block);
2002-10-12 11:37:38 +00:00
for(y=starty; y<endy; y++) {
rect= sbuts->rect + 1 + PR_RECTX*sbuts->cury;
if(y== -PR_RECTY/2 || y==endy-1); /* emboss */
else if(mat) {
2002-10-12 11:37:38 +00:00
if(mat->mode & MA_HALO) {
for(x=startx; x<endx; x++, rect++) {
rect[0]= previewback(mat->pr_back, x, y);
}
if(har.flarec) {
if(y==endy-2) previewflare(sbuts, &har, sbuts->rect);
}
else {
halo_preview_pixel(&har, startx, endx, y, (char *) (rect-PR_RECTX));
}
}
else {
ysq= y*y;
for(x=startx; x<endx; x++, rect++) {
xsq= x*x;
if(mat->pr_type==MA_SPHERE) {
if(xsq+ysq <= radsq) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi.vn[0]= x;
shi.vn[1]= y;
shi.vn[2]= sqrt( (float)(radsq-xsq-ysq) );
Normalise(shi.vn);
2002-10-12 11:37:38 +00:00
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
vec[0]= shi.vn[0];
vec[1]= shi.vn[2];
vec[2]= -shi.vn[1];
2002-10-12 11:37:38 +00:00
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shade_preview_pixel(&shi, vec, x, y, (char *)rect, 1);
2002-10-12 11:37:38 +00:00
}
else {
rect[0]= previewback(mat->pr_back, x, y);
}
}
else if(mat->pr_type==MA_CUBE) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
if( ray_previewrender(x, y, vec, shi.vn) ) {
2002-10-12 11:37:38 +00:00
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shade_preview_pixel(&shi, vec, x, y, (char *)rect, 0);
2002-10-12 11:37:38 +00:00
}
else {
rect[0]= previewback(mat->pr_back, x, y);
}
}
else {
vec[0]= x*(2.0/PR_RECTX);
vec[1]= y*(2.0/PR_RECTX);
vec[2]= 0.0;
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shi.vn[0]= shi.vn[1]= 0.0;
shi.vn[2]= 1.0;
2002-10-12 11:37:38 +00:00
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
shade_preview_pixel(&shi, vec, x, y, (char *)rect, 0);
2002-10-12 11:37:38 +00:00
}
}
}
}
else if(tex) {
2002-10-12 11:37:38 +00:00
for(x=startx; x<endx; x++, rect++) {
texture_preview_pixel(tex, x, y, (char *)rect);
}
}
else if(la) {
2002-10-12 11:37:38 +00:00
for(x=startx; x<endx; x++, rect++) {
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
lamp_preview_pixel(&shi, lar, x, y, (char *)rect);
2002-10-12 11:37:38 +00:00
}
}
else {
for(x=startx; x<endx; x++, rect++) {
sky_preview_pixel(lens, x, y, (char *)rect);
}
}
if(y<endy-2) {
if(qtest()) {
addafterqueue(curarea->win, RENDERPREVIEW, 1);
break;
}
}
display_pr_scanline(sbuts->rect, sbuts->cury);
sbuts->cury++;
}
if(sbuts->cury>=PR_RECTY && tex)
if (sbuts->tab[CONTEXT_SHADING]==TAB_SHADING_TEX)
draw_tex_crop(sbuts->lockpoin);
2002-10-12 11:37:38 +00:00
glDrawBuffer(GL_BACK);
/* draw again for clean swapbufers */
BIF_previewdraw();
uiPanelPop(block);
2002-10-12 11:37:38 +00:00
if(mat) {
2002-10-12 11:37:38 +00:00
end_render_material(mat);
for(x=0; x<8; x++) {
if(mat->mtex[x] && mat->mtex[x]->tex) end_render_texture(mat->mtex[x]->tex);
}
}
else if(tex) {
2002-10-12 11:37:38 +00:00
end_render_texture(tex);
}
else if(la) {
2002-10-12 11:37:38 +00:00
if(R.totlamp) {
if(R.la[0]->org) MEM_freeN(R.la[0]->org);
MEM_freeN(R.la[0]);
}
R.totlamp= 0;
end_render_textures();
}
Area lights and more... - New lamp type added "Area". This uses the radiosity formula (Stoke) to calculate the amount of energy which is received from a plane. Result is very nice local light, which nicely spreads out. - Area lamps have a 'gamma' option to control the light spread - Area lamp builtin sizes: square, rect, cube & box. Only first 2 are implemented. Set a type, and define area size - Button area size won't affect the amount of energy. But scaling the lamp in 3d window will do. This is to cover the case when you scale an entire scene, the light then will remain identical If you just want to change area lamp size, use buttons when you dont want to make the scene too bright or too dark - Since area lights realistically are sensitive for distance (quadratic), the effect it has is quickly too much, or too less. For this the "Dist" value in Lamp can be used. Set it at Dist=10 to have reasonable light on distance 10 Blender units (assumed you didnt scale lamp object). - I tried square sized specularity, but this looked totally weird. Not committed - Plan is to extend area light with 3d dimensions, boxes and cubes. - Note that area light is one-sided, towards negative Z. I need to design a nice drawing method for it. Area Shadow - Since there are a lot of variables associated with soft shadow, they now only are available for Area lights. Allowing spot & normal lamp to have soft shadow is possible though, but will require a reorganisation of the Lamp buttons. Is a point of research & feedback still. - Apart from area size, you now can individually set amount of samples in X and Y direction (for area lamp type 'Rect'). For box type area lamp, this will become 3 dimensions - Area shadows have four options: "Clip circle" : only uses a circular shape of samples, gives smoother results "Dither" : use a 2x2 dither mask "Jitter" : applys a pseudo-random offset to samples "Umbra" : extra emphasis on area that's fully in shadow. Raytrace speedup - improved filling in faces in Octree. Large faces occupied too many nodes - added a coherence check; rays fired sequentially that begin and end in same octree nodes, and that don't intersect, are quickly rejected - rendering shadow scenes benefits from this 20-40%. My statue test monkey file now renders in 19 seconds (was 30). Plus: - adjusted specular max to 511, and made sure Blinn spec has again this incredible small spec size - for UI rounded theme: the color "button" displayed RGB color too dark - fixed countall() function, to also include Subsurf totals - removed setting the 'near' clipping for pressing dot-key numpad - when you press the buttons-window icon for 'Shading Context' the context automaticilly switches as with F5 hotkey Please be warned that this is not a release... settings in files might not work as it did, nor guaranteed to work when we do a release. :)
2003-12-29 16:52:51 +00:00
else if(wrld) {
end_render_textures();
}
2002-10-12 11:37:38 +00:00
}