2 * Glide64 - Glide video plugin for Nintendo 64 emulators.
3 * Copyright (c) 2002 Dave2001
4 * Copyright (c) 2003-2009 Sergey 'Gonetz' Lipski
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation; either version 2 of the License, or
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
16 * You should have received a copy of the GNU General Public License
17 * along with this program; if not, write to the Free Software
18 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
21 //****************************************************************
23 // Glide64 - Glide Plugin for Nintendo 64 emulators
24 // Project started on December 29th, 2001
27 // Dave2001, original author, founded the project in 2001, left it in 2002
28 // Gugaman, joined the project in 2002, left it in 2002
29 // Sergey 'Gonetz' Lipski, joined the project in 2002, main author since fall of 2002
30 // Hiroshi 'KoolSmoky' Morii, joined the project in 2007
32 //****************************************************************
35 // * Write your name and (optional)email, commented by your work, so I know who did it, and so that you can find which parts you modified when it comes time to send it to me.
36 // * Do NOT send me the whole project or file that you modified. Take out your modified code sections, and tell me where to put them. If people sent the whole thing, I would have many different versions, but no idea how to combine them all.
38 //****************************************************************
40 // December 2008 Created by Gonetz (Gonetz@ngs.ru)
42 //****************************************************************
46 typedef float M44[4][4];
53 } zSortRdp = {{0, 0}, {0, 0}, 0, 0};
56 static int Calc_invw (int w) {
66 Result.UW = 0x7FFFFFFF;
70 if (Result.UHW[1] == 0xFFFF && Result.HW[0] < 0) {
71 Result.W = ~Result.W + 1;
78 for (count = 31; count > 0; count--) {
79 if ((Result.W & (1 << count))) {
80 Result.W &= (0xFFC00000 >> (31 - count) );
84 Result.W = 0x7FFFFFFF / Result.W;
85 for (count = 31; count > 0; count--) {
86 if ((Result.W & (1 << count))) {
87 Result.W &= (0xFFFF8000 >> (31 - count) );
98 static void uc9_draw_object (wxUint8 * addr, wxUint32 type)
102 wxUint32 textured, vnum, vsize;
126 textured = vnum = vsize = 0;
130 for (wxUint32 i = 0; i < vnum; i++)
133 v.sx = zSortRdp.scale_x * ((short*)addr)[0^1];
134 v.sy = zSortRdp.scale_y * ((short*)addr)[1^1];
142 v.uv_calculated = 0xFFFFFFFF;
145 v.screen_translated = 2;
148 v.ou = ((short*)addr)[4^1];
149 v.ov = ((short*)addr)[5^1];
150 v.w = Calc_invw(((int*)addr)[3]) / 31.0f;
152 FRDP ("v%d - sx: %f, sy: %f ou: %f, ov: %f, w: %f, r=%d, g=%d, b=%d, a=%d\n", i, v.sx/rdp.scale_x, v.sy/rdp.scale_y, v.ou*rdp.tiles[rdp.cur_tile].s_scale, v.ov*rdp.tiles[rdp.cur_tile].t_scale, v.w, v.r, v.g, v.b, v.a);
157 FRDP ("v%d - sx: %f, sy: %f r=%d, g=%d, b=%d, a=%d\n", i, v.sx/rdp.scale_x, v.sy/rdp.scale_y, v.r, v.g, v.b, v.a);
170 FRDP("uc9:Tri #%d, #%d\n", rdp.tri_n, rdp.tri_n+1);
176 FRDP("uc9:Quad #%d, #%d\n", rdp.tri_n, rdp.tri_n+1);
183 static wxUint32 uc9_load_object (wxUint32 zHeader, wxUint32 * rdpcmds)
185 wxUint32 type = zHeader & 7;
186 wxUint8 * addr = gfx.RDRAM + (zHeader&0xFFFFFFF8);
191 rdp.cmd1 = ((wxUint32*)addr)[1];
192 if (rdp.cmd1 != rdpcmds[0])
194 rdpcmds[0] = rdp.cmd1;
198 uc9_draw_object(addr + 8, type);
205 rdp.cmd1 = ((wxUint32*)addr)[1];
206 if (rdp.cmd1 != rdpcmds[0])
208 rdpcmds[0] = rdp.cmd1;
211 rdp.cmd1 = ((wxUint32*)addr)[2];
212 if (rdp.cmd1 != rdpcmds[1])
215 rdpcmds[1] = rdp.cmd1;
217 rdp.cmd1 = ((wxUint32*)addr)[3];
218 if (rdp.cmd1 != rdpcmds[2])
221 rdpcmds[2] = rdp.cmd1;
226 uc9_draw_object(addr + 16, type);
231 return segoffset(((wxUint32*)addr)[0]);
234 static void uc9_object ()
236 LRDP("uc9:object\n");
237 wxUint32 rdpcmds[3] = {0, 0, 0};
238 wxUint32 cmd1 = rdp.cmd1;
239 wxUint32 zHeader = segoffset(rdp.cmd0);
241 zHeader = uc9_load_object(zHeader, rdpcmds);
242 zHeader = segoffset(cmd1);
244 zHeader = uc9_load_object(zHeader, rdpcmds);
247 static void uc9_mix ()
249 LRDP("uc9:mix IGNORED\n");
252 static void uc9_fmlight ()
254 int mid = rdp.cmd0&0xFF;
255 rdp.num_lights = 1 + ((rdp.cmd1>>12)&0xFF);
256 wxUint32 a = -1024 + (rdp.cmd1&0xFFF);
257 FRDP ("uc9:fmlight matrix: %d, num: %d, dmem: %04lx\n", mid, rdp.num_lights, a);
268 m = (M44*)rdp.combined;
272 rdp.light[rdp.num_lights].r = (float)(((wxUint8*)gfx.DMEM)[(a+0)^3]) / 255.0f;
273 rdp.light[rdp.num_lights].g = (float)(((wxUint8*)gfx.DMEM)[(a+1)^3]) / 255.0f;
274 rdp.light[rdp.num_lights].b = (float)(((wxUint8*)gfx.DMEM)[(a+2)^3]) / 255.0f;
275 rdp.light[rdp.num_lights].a = 1.0f;
276 FRDP ("ambient light: r: %.3f, g: %.3f, b: %.3f\n", rdp.light[rdp.num_lights].r, rdp.light[rdp.num_lights].g, rdp.light[rdp.num_lights].b);
279 for (i = 0; i < rdp.num_lights; i++)
281 rdp.light[i].r = (float)(((wxUint8*)gfx.DMEM)[(a+0)^3]) / 255.0f;
282 rdp.light[i].g = (float)(((wxUint8*)gfx.DMEM)[(a+1)^3]) / 255.0f;
283 rdp.light[i].b = (float)(((wxUint8*)gfx.DMEM)[(a+2)^3]) / 255.0f;
284 rdp.light[i].a = 1.0f;
285 rdp.light[i].dir_x = (float)(((char*)gfx.DMEM)[(a+8)^3]) / 127.0f;
286 rdp.light[i].dir_y = (float)(((char*)gfx.DMEM)[(a+9)^3]) / 127.0f;
287 rdp.light[i].dir_z = (float)(((char*)gfx.DMEM)[(a+10)^3]) / 127.0f;
288 FRDP ("light: n: %d, r: %.3f, g: %.3f, b: %.3f, x: %.3f, y: %.3f, z: %.3f\n",
289 i, rdp.light[i].r, rdp.light[i].g, rdp.light[i].b,
290 rdp.light[i].dir_x, rdp.light[i].dir_y, rdp.light[i].dir_z);
291 // TransformVector(&rdp.light[i].dir_x, rdp.light_vector[i], *m);
292 InverseTransformVector(&rdp.light[i].dir_x, rdp.light_vector[i], *m);
293 NormalizeVector (rdp.light_vector[i]);
294 FRDP ("light vector: n: %d, x: %.3f, y: %.3f, z: %.3f\n",
295 i, rdp.light_vector[i][0], rdp.light_vector[i][1], rdp.light_vector[i][2]);
298 for (i = 0; i < 2; i++)
300 float dir_x = (float)(((char*)gfx.DMEM)[(a+8)^3]) / 127.0f;
301 float dir_y = (float)(((char*)gfx.DMEM)[(a+9)^3]) / 127.0f;
302 float dir_z = (float)(((char*)gfx.DMEM)[(a+10)^3]) / 127.0f;
303 if (sqrt(dir_x*dir_x + dir_y*dir_y + dir_z*dir_z) < 0.98)
305 rdp.use_lookat = FALSE;
308 rdp.lookat[i][0] = dir_x;
309 rdp.lookat[i][1] = dir_y;
310 rdp.lookat[i][2] = dir_z;
313 rdp.use_lookat = TRUE;
316 static void uc9_light ()
318 wxUint32 csrs = -1024 + ((rdp.cmd0>>12)&0xFFF);
319 wxUint32 nsrs = -1024 + (rdp.cmd0&0xFFF);
320 wxUint32 num = 1 + ((rdp.cmd1>>24)&0xFF);
321 wxUint32 cdest = -1024 + ((rdp.cmd1>>12)&0xFFF);
322 wxUint32 tdest = -1024 + (rdp.cmd1&0xFFF);
323 int use_material = (csrs != 0x0ff0);
325 FRDP ("uc9:light n: %d, colsrs: %04lx, normales: %04lx, coldst: %04lx, texdst: %04lx\n", num, csrs, nsrs, cdest, tdest);
327 for (wxUint32 i = 0; i < num; i++)
329 v.vec[0] = ((char*)gfx.DMEM)[(nsrs++)^3];
330 v.vec[1] = ((char*)gfx.DMEM)[(nsrs++)^3];
331 v.vec[2] = ((char*)gfx.DMEM)[(nsrs++)^3];
334 NormalizeVector (v.vec);
339 v.r = (wxUint8)(((wxUint32)v.r * gfx.DMEM[(csrs++)^3])>>8);
340 v.g = (wxUint8)(((wxUint32)v.g * gfx.DMEM[(csrs++)^3])>>8);
341 v.b = (wxUint8)(((wxUint32)v.b * gfx.DMEM[(csrs++)^3])>>8);
342 v.a = gfx.DMEM[(csrs++)^3];
344 gfx.DMEM[(cdest++)^3] = v.r;
345 gfx.DMEM[(cdest++)^3] = v.g;
346 gfx.DMEM[(cdest++)^3] = v.b;
347 gfx.DMEM[(cdest++)^3] = v.a;
348 ((short*)gfx.DMEM)[(tdest++)^1] = (short)v.ou;
349 ((short*)gfx.DMEM)[(tdest++)^1] = (short)v.ov;
353 static void uc9_mtxtrnsp ()
355 LRDP("uc9:mtxtrnsp - ignored\n");
357 LRDP("uc9:mtxtrnsp ");
359 switch (rdp.cmd1&0xF) {
369 s = (M44*)rdp.combined;
385 static void uc9_mtxcat ()
390 wxUint32 S = rdp.cmd0&0xF;
391 wxUint32 T = (rdp.cmd1>>16)&0xF;
392 wxUint32 D = rdp.cmd1&0xF;
403 s = (M44*)rdp.combined;
418 t = (M44*)rdp.combined;
421 DECLAREALIGN16VAR(m[4][4]);
422 MulMatrices(*s, *t, m);
426 memcpy (rdp.model, m, 64);;
430 memcpy (rdp.proj, m, 64);;
434 memcpy (rdp.combined, m, 64);;
438 #ifdef EXTREME_LOGGING
439 FRDP ("\nmodel\n{%f,%f,%f,%f}\n", rdp.model[0][0], rdp.model[0][1], rdp.model[0][2], rdp.model[0][3]);
440 FRDP ("{%f,%f,%f,%f}\n", rdp.model[1][0], rdp.model[1][1], rdp.model[1][2], rdp.model[1][3]);
441 FRDP ("{%f,%f,%f,%f}\n", rdp.model[2][0], rdp.model[2][1], rdp.model[2][2], rdp.model[2][3]);
442 FRDP ("{%f,%f,%f,%f}\n", rdp.model[3][0], rdp.model[3][1], rdp.model[3][2], rdp.model[3][3]);
443 FRDP ("\nproj\n{%f,%f,%f,%f}\n", rdp.proj[0][0], rdp.proj[0][1], rdp.proj[0][2], rdp.proj[0][3]);
444 FRDP ("{%f,%f,%f,%f}\n", rdp.proj[1][0], rdp.proj[1][1], rdp.proj[1][2], rdp.proj[1][3]);
445 FRDP ("{%f,%f,%f,%f}\n", rdp.proj[2][0], rdp.proj[2][1], rdp.proj[2][2], rdp.proj[2][3]);
446 FRDP ("{%f,%f,%f,%f}\n", rdp.proj[3][0], rdp.proj[3][1], rdp.proj[3][2], rdp.proj[3][3]);
447 FRDP ("\ncombined\n{%f,%f,%f,%f}\n", rdp.combined[0][0], rdp.combined[0][1], rdp.combined[0][2], rdp.combined[0][3]);
448 FRDP ("{%f,%f,%f,%f}\n", rdp.combined[1][0], rdp.combined[1][1], rdp.combined[1][2], rdp.combined[1][3]);
449 FRDP ("{%f,%f,%f,%f}\n", rdp.combined[2][0], rdp.combined[2][1], rdp.combined[2][2], rdp.combined[2][3]);
450 FRDP ("{%f,%f,%f,%f}\n", rdp.combined[3][0], rdp.combined[3][1], rdp.combined[3][2], rdp.combined[3][3]);
465 static void uc9_mult_mpmtx ()
467 //int id = rdp.cmd0&0xFF;
468 int num = 1+ ((rdp.cmd1>>24)&0xFF);
469 int src = -1024 + ((rdp.cmd1>>12)&0xFFF);
470 int dst = -1024 + (rdp.cmd1&0xFFF);
471 FRDP ("uc9:mult_mpmtx from: %04lx to: %04lx n: %d\n", src, dst, num);
472 short * saddr = (short*)(gfx.DMEM+src);
473 zSortVDest * daddr = (zSortVDest*)(gfx.DMEM+dst);
476 memset(&v, 0, sizeof(zSortVDest));
477 //float scale_x = 4.0f/rdp.scale_x;
478 //float scale_y = 4.0f/rdp.scale_y;
479 for (int i = 0; i < num; i++)
481 short sx = saddr[(idx++)^1];
482 short sy = saddr[(idx++)^1];
483 short sz = saddr[(idx++)^1];
484 float x = sx*rdp.combined[0][0] + sy*rdp.combined[1][0] + sz*rdp.combined[2][0] + rdp.combined[3][0];
485 float y = sx*rdp.combined[0][1] + sy*rdp.combined[1][1] + sz*rdp.combined[2][1] + rdp.combined[3][1];
486 float z = sx*rdp.combined[0][2] + sy*rdp.combined[1][2] + sz*rdp.combined[2][2] + rdp.combined[3][2];
487 float w = sx*rdp.combined[0][3] + sy*rdp.combined[1][3] + sz*rdp.combined[2][3] + rdp.combined[3][3];
488 v.sx = (short)(zSortRdp.view_trans[0] + x / w * zSortRdp.view_scale[0]);
489 v.sy = (short)(zSortRdp.view_trans[1] + y / w * zSortRdp.view_scale[1]);
494 v.invw = Calc_invw((int)(w * 31.0));
500 int fog = (int)(z / w * rdp.fog_multiplier + rdp.fog_offset);
503 v.fog = (fog >= 0) ? (wxUint8)fog : 0;
507 if (x < -w) v.cc |= 0x10;
508 if (x > w) v.cc |= 0x01;
509 if (y < -w) v.cc |= 0x20;
510 if (y > w) v.cc |= 0x02;
511 if (w < 0.1f) v.cc |= 0x04;
514 //memcpy(gfx.DMEM+dst+sizeof(zSortVDest)*i, &v, sizeof(zSortVDest));
515 // FRDP("v%d x: %d, y: %d, z: %d -> sx: %d, sy: %d, w: %d, xi: %d, yi: %d, wi: %d, fog: %d\n", i, sx, sy, sz, v.sx, v.sy, v.invw, v.xi, v.yi, v.wi, v.fog);
516 FRDP("v%d x: %d, y: %d, z: %d -> sx: %04lx, sy: %04lx, invw: %08lx - %f, xi: %04lx, yi: %04lx, wi: %04lx, fog: %04lx\n", i, sx, sy, sz, v.sx, v.sy, v.invw, w, v.xi, v.yi, v.wi, v.fog);
520 static void uc9_link_subdl ()
522 LRDP("uc9:link_subdl IGNORED\n");
525 static void uc9_set_subdl ()
527 LRDP("uc9:set_subdl IGNORED\n");
530 static void uc9_wait_signal ()
532 LRDP("uc9:wait_signal IGNORED\n");
535 static void uc9_send_signal ()
537 LRDP("uc9:send_signal IGNORED\n");
542 LRDP("uc9:movemem\n");
543 int idx = rdp.cmd0 & 0x0E;
544 int ofs = ((rdp.cmd0>>6)&0x1ff)<<3;
545 int len = (1 + ((rdp.cmd0>>15)&0x1ff))<<3;
546 FRDP ("uc9:movemem ofs: %d, len: %d. ", ofs, len);
547 int flag = rdp.cmd0 & 0x01;
548 wxUint32 addr = segoffset(rdp.cmd1);
555 int dmem_addr = (idx<<3) + ofs;
556 FRDP ("Load to DMEM. %08lx -> %08lx\n", addr, dmem_addr);
557 memcpy(gfx.DMEM + dmem_addr, gfx.RDRAM + addr, len);
561 int dmem_addr = (idx<<3) + ofs;
562 FRDP ("Load from DMEM. %08lx -> %08lx\n", dmem_addr, addr);
563 memcpy(gfx.RDRAM + addr, gfx.DMEM + dmem_addr, len);
567 case 4: // model matrix
568 case 6: // projection matrix
569 case 8: // combined matrix
571 DECLAREALIGN16VAR(m[4][4]);
572 load_matrix(m, addr);
575 case 4: // model matrix
576 LRDP("Modelview load\n");
579 case 6: // projection matrix
580 LRDP("Projection load\n");
583 case 8: // projection matrix
584 LRDP("Combined load\n");
585 rdp.update &= ~UPDATE_MULT_MAT;
586 memcpy (rdp.combined, m, 64);;
589 #ifdef EXTREME_LOGGING
590 FRDP ("{%f,%f,%f,%f}\n", m[0][0], m[0][1], m[0][2], m[0][3]);
591 FRDP ("{%f,%f,%f,%f}\n", m[1][0], m[1][1], m[1][2], m[1][3]);
592 FRDP ("{%f,%f,%f,%f}\n", m[2][0], m[2][1], m[2][2], m[2][3]);
593 FRDP ("{%f,%f,%f,%f}\n", m[3][0], m[3][1], m[3][2], m[3][3]);
594 FRDP ("\nmodel\n{%f,%f,%f,%f}\n", rdp.model[0][0], rdp.model[0][1], rdp.model[0][2], rdp.model[0][3]);
595 FRDP ("{%f,%f,%f,%f}\n", rdp.model[1][0], rdp.model[1][1], rdp.model[1][2], rdp.model[1][3]);
596 FRDP ("{%f,%f,%f,%f}\n", rdp.model[2][0], rdp.model[2][1], rdp.model[2][2], rdp.model[2][3]);
597 FRDP ("{%f,%f,%f,%f}\n", rdp.model[3][0], rdp.model[3][1], rdp.model[3][2], rdp.model[3][3]);
598 FRDP ("\nproj\n{%f,%f,%f,%f}\n", rdp.proj[0][0], rdp.proj[0][1], rdp.proj[0][2], rdp.proj[0][3]);
599 FRDP ("{%f,%f,%f,%f}\n", rdp.proj[1][0], rdp.proj[1][1], rdp.proj[1][2], rdp.proj[1][3]);
600 FRDP ("{%f,%f,%f,%f}\n", rdp.proj[2][0], rdp.proj[2][1], rdp.proj[2][2], rdp.proj[2][3]);
601 FRDP ("{%f,%f,%f,%f}\n", rdp.proj[3][0], rdp.proj[3][1], rdp.proj[3][2], rdp.proj[3][3]);
607 LRDP("Othermode - IGNORED\n");
612 wxUint32 a = addr >> 1;
613 short scale_x = ((short*)gfx.RDRAM)[(a+0)^1] >> 2;
614 short scale_y = ((short*)gfx.RDRAM)[(a+1)^1] >> 2;
615 short scale_z = ((short*)gfx.RDRAM)[(a+2)^1];
616 rdp.fog_multiplier = ((short*)gfx.RDRAM)[(a+3)^1];
617 short trans_x = ((short*)gfx.RDRAM)[(a+4)^1] >> 2;
618 short trans_y = ((short*)gfx.RDRAM)[(a+5)^1] >> 2;
619 short trans_z = ((short*)gfx.RDRAM)[(a+6)^1];
620 rdp.fog_offset = ((short*)gfx.RDRAM)[(a+7)^1];
621 rdp.view_scale[0] = scale_x * rdp.scale_x;
622 rdp.view_scale[1] = scale_y * rdp.scale_y;
623 rdp.view_scale[2] = 32.0f * scale_z;
624 rdp.view_trans[0] = trans_x * rdp.scale_x;
625 rdp.view_trans[1] = trans_y * rdp.scale_y;
626 rdp.view_trans[2] = 32.0f * trans_z;
627 zSortRdp.view_scale[0] = (float)(scale_x*4);
628 zSortRdp.view_scale[1] = (float)(scale_y*4);
629 zSortRdp.view_trans[0] = (float)(trans_x*4);
630 zSortRdp.view_trans[1] = (float)(trans_y*4);
631 zSortRdp.scale_x = rdp.scale_x / 4.0f;
632 zSortRdp.scale_y = rdp.scale_y / 4.0f;
634 rdp.update |= UPDATE_VIEWPORT;
636 rdp.mipmap_level = 0;
638 TILE *tmp_tile = &rdp.tiles[0];
640 tmp_tile->org_s_scale = 0xFFFF;
641 tmp_tile->org_t_scale = 0xFFFF;
642 tmp_tile->s_scale = 0.031250f;
643 tmp_tile->t_scale = 0.031250f;
645 rdp.geom_mode |= 0x0200;
647 FRDP ("viewport scale(%d, %d, %d), trans(%d, %d, %d), from:%08lx\n", scale_x, scale_y, scale_z,
648 trans_x, trans_y, trans_z, a);
649 FRDP ("fog: multiplier: %f, offset: %f\n", rdp.fog_multiplier, rdp.fog_offset);
654 FRDP ("** UNKNOWN %d\n", idx);
659 static void uc9_setscissor()
663 if ((rdp.scissor_o.lr_x - rdp.scissor_o.ul_x) > (zSortRdp.view_scale[0] - zSortRdp.view_trans[0]))
665 float w = (rdp.scissor_o.lr_x - rdp.scissor_o.ul_x) / 2.0f;
666 float h = (rdp.scissor_o.lr_y - rdp.scissor_o.ul_y) / 2.0f;
667 rdp.view_scale[0] = w * rdp.scale_x;
668 rdp.view_scale[1] = h * rdp.scale_y;
669 rdp.view_trans[0] = w * rdp.scale_x;
670 rdp.view_trans[1] = h * rdp.scale_y;
671 zSortRdp.view_scale[0] = w * 4.0f;
672 zSortRdp.view_scale[1] = h * 4.0f;
673 zSortRdp.view_trans[0] = w * 4.0f;
674 zSortRdp.view_trans[1] = h * 4.0f;
675 zSortRdp.scale_x = rdp.scale_x / 4.0f;
676 zSortRdp.scale_y = rdp.scale_y / 4.0f;
677 rdp.update |= UPDATE_VIEWPORT;
679 rdp.mipmap_level = 0;
681 TILE *tmp_tile = &rdp.tiles[0];
683 tmp_tile->org_s_scale = 0xFFFF;
684 tmp_tile->org_t_scale = 0xFFFF;
685 tmp_tile->s_scale = 0.031250f;
686 tmp_tile->t_scale = 0.031250f;
688 rdp.geom_mode |= 0x0200;