[mupen64plus-pandora.git] / source / gles2glide64 / src / Glide64 / ucode08.h

/*
* Glide64 - Glide video plugin for Nintendo 64 emulators.
* Copyright (c) 2002  Dave2001
* Copyright (c) 2003-2009  Sergey 'Gonetz' Lipski
*
* 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
* any later version.
*
* 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
*/

//****************************************************************
//
// Glide64 - Glide Plugin for Nintendo 64 emulators
// Project started on December 29th, 2001
//
// Authors:
// Dave2001, original author, founded the project in 2001, left it in 2002
// Gugaman, joined the project in 2002, left it in 2002
// Sergey 'Gonetz' Lipski, joined the project in 2002, main author since fall of 2002
// Hiroshi 'KoolSmoky' Morii, joined the project in 2007
//
//****************************************************************
//
// To modify Glide64:
// * 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.
// * 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.
//
//****************************************************************
//
// January 2004 Created by Gonetz (Gonetz@ngs.ru)
//
//****************************************************************

wxUint32 uc8_normale_addr = 0;
float uc8_coord_mod[16];

static void uc8_vertex ()
{
Check_FrameSkip;

	if (rdp.update & UPDATE_MULT_MAT)
	{
		rdp.update ^= UPDATE_MULT_MAT;
    	MulMatrices(rdp.model, rdp.proj, rdp.combined);
	}

	wxUint32 addr = segoffset(rdp.cmd1);
	int v0, i, n;
	float x, y, z;

	rdp.vn = n = (rdp.cmd0 >> 12) & 0xFF;
	rdp.v0 = v0 = ((rdp.cmd0 >> 1) & 0x7F) - n;

	FRDP ("uc8:vertex n: %d, v0: %d, from: %08lx\n", n, v0, addr);

	if (v0 < 0)
	{
		RDP_E ("** ERROR: uc2:vertex v0 < 0\n");
		LRDP("** ERROR: uc2:vertex v0 < 0\n");
		return;
	}
	//*
	// This is special, not handled in update()
	if (rdp.update & UPDATE_LIGHTS)
	{
		rdp.update ^= UPDATE_LIGHTS;

		// Calculate light vectors
		for (wxUint32 l=0; l<rdp.num_lights; l++)
		{
			InverseTransformVector(&rdp.light[l].dir_x, rdp.light_vector[l], rdp.model);
			NormalizeVector (rdp.light_vector[l]);
#ifdef EXTREME_LOGGING
			FRDP("light_vector[%d] x: %f, y: %f, z: %f\n", l, rdp.light_vector[l][0], rdp.light_vector[l][1], rdp.light_vector[l][2]);
#endif
		}
	}
	//*/
    #ifdef __ARM_NEON__
    float32x4_t comb0, comb1, comb2, comb3;
    float32x4_t v_xyzw;
    comb0 = vld1q_f32(rdp.combined[0]);
    comb1 = vld1q_f32(rdp.combined[1]);
    comb2 = vld1q_f32(rdp.combined[2]);
    comb3 = vld1q_f32(rdp.combined[3]);
	
	float32x4_t uc8_8_11, uc8_12_15;
	uc8_8_11 = vld1q_f32(uc8_coord_mod+8);
	uc8_12_15 = vld1q_f32(uc8_coord_mod+12);
	float32x4_t unite={1.0f, 1.0f, 1.0f, 1.0f};
	
	float32x4_t rdplight[12], rdpcolor[12];
	for (i=0; i<rdp.num_lights; i++) {
		rdplight[i] = vld1q_f32(&rdp.light[i].x);
		rdpcolor[i] = vld1q_f32(&rdp.light[i].r);
	}
	
    #endif
	for (i=0; i < (n<<4); i+=16)
	{
		VERTEX *v = &rdp.vtx[v0 + (i>>4)];
		x   = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 0)^1];
		y   = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 1)^1];
		z   = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 2)^1];
		v->flags  = ((wxUint16*)gfx.RDRAM)[(((addr+i) >> 1) + 3)^1];
		v->ou   = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 4)^1];
		v->ov   = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 5)^1];
        v->uv_scaled = 0;
		v->a    = ((wxUint8*)gfx.RDRAM)[(addr+i + 15)^3];

#ifdef EXTREME_LOGGING
		FRDP ("before v%d - x: %f, y: %f, z: %f\n", i>>4, x, y, z);
#endif
		#ifdef __ARM_NEON__
		v_xyzw = x*comb0+y*comb1+z*comb2+comb3;
		//vst1q_f32((float*)v, v_xyzw);
		#else
		v->x = x*rdp.combined[0][0] + y*rdp.combined[1][0] + z*rdp.combined[2][0] + rdp.combined[3][0];
		v->y = x*rdp.combined[0][1] + y*rdp.combined[1][1] + z*rdp.combined[2][1] + rdp.combined[3][1];
		v->z = x*rdp.combined[0][2] + y*rdp.combined[1][2] + z*rdp.combined[2][2] + rdp.combined[3][2];
		v->w = x*rdp.combined[0][3] + y*rdp.combined[1][3] + z*rdp.combined[2][3] + rdp.combined[3][3];
		#endif

#ifdef EXTREME_LOGGING
		FRDP ("v%d - x: %f, y: %f, z: %f, w: %f, u: %f, v: %f, flags: %d\n", i>>4, v->x, v->y, v->z, v->w, v->ou, v->ov, v->flags);
#endif
		v->uv_calculated = 0xFFFFFFFF;
		v->screen_translated = 0;
		v->shade_mod = 0;
		#ifdef __ARM_NEON__
		v->x=v_xyzw[0];
		v->y=v_xyzw[1];
		v->z=v_xyzw[2];
		v->w=v_xyzw[3];
		#endif

		///*
		v->r = ((wxUint8*)gfx.RDRAM)[(addr+i + 12)^3];
		v->g = ((wxUint8*)gfx.RDRAM)[(addr+i + 13)^3];
		v->b = ((wxUint8*)gfx.RDRAM)[(addr+i + 14)^3];

        if (fabs(v->w) < 0.001) v->w = 0.001f;
		v->oow = 1.0f / v->w;
		#ifdef __ARM_NEON__
		v_xyzw *= v->oow;
		v->x_w=v_xyzw[0];
		v->y_w=v_xyzw[1];
		v->z_w=v_xyzw[2];
		#else
		v->x_w = v->x * v->oow;
		v->y_w = v->y * v->oow;
		v->z_w = v->z * v->oow;
		#endif

		v->scr_off = 0;
		if (v->x < -v->w) v->scr_off |= 1;
		if (v->x > v->w) v->scr_off |= 2;
		if (v->y < -v->w) v->scr_off |= 4;
		if (v->y > v->w) v->scr_off |= 8;
		if (v->w < 0.1f) v->scr_off |= 16;
#ifdef EXTREME_LOGGING
		FRDP ("r: %02lx, g: %02lx, b: %02lx, a: %02lx\n", v->r, v->g, v->b, v->a);
#endif

		if ((rdp.geom_mode & 0x00020000))
		{
			wxUint32 shift = v0 << 1;
			v->vec[0] = ((char*)gfx.RDRAM)[(uc8_normale_addr + (i>>3) + shift + 0)^3];
			v->vec[1] = ((char*)gfx.RDRAM)[(uc8_normale_addr + (i>>3) + shift + 1)^3];
			v->vec[2] = (char)(v->flags&0xff);

  			if (rdp.geom_mode & 0x80000)
  			{
  				calc_linear (v);
#ifdef EXTREME_LOGGING
  				FRDP ("calc linear: v%d - u: %f, v: %f\n", i>>4, v->ou, v->ov);
#endif
  			}
  			else if (rdp.geom_mode & 0x40000)
  			{
  				calc_sphere (v);
#ifdef EXTREME_LOGGING
  				FRDP ("calc sphere: v%d - u: %f, v: %f\n", i>>4, v->ou, v->ov);
#endif
  			}
			//     FRDP("calc light. r: 0x%02lx, g: 0x%02lx, b: 0x%02lx, nx: %.3f, ny: %.3f, nz: %.3f\n", v->r, v->g, v->b, v->vec[0], v->vec[1], v->vec[2]);
			FRDP("v[%d] calc light. r: 0x%02lx, g: 0x%02lx, b: 0x%02lx\n", i>>4, v->r, v->g, v->b);
			#ifdef __ARM_NEON__
			float32x4_t color = {rdp.light[rdp.num_lights].r, rdp.light[rdp.num_lights].g, rdp.light[rdp.num_lights].b, 1.0f};
			#else
			float color[3] = {rdp.light[rdp.num_lights].r, rdp.light[rdp.num_lights].g, rdp.light[rdp.num_lights].b};
			#endif
			FRDP("ambient light. r: %f, g: %f, b: %f\n", color[0], color[1], color[2]);
			float light_intensity = 0.0f;
			wxUint32 l;
			if (rdp.geom_mode & 0x00400000)
			{
				NormalizeVector (v->vec);
				for (l = 0; l < rdp.num_lights-1; l++)
        {
          if (!rdp.light[l].nonblack)
            continue;
          light_intensity = DotProduct (rdp.light_vector[l], v->vec);
          FRDP("light %d, intensity : %f\n", l, light_intensity);
          if (light_intensity < 0.0f)
            continue;
          //*
          if (rdp.light[l].ca > 0.0f)
          {
			#ifdef __ARM_NEON__
			float32x4_t vxyzw = (v_xyzw + uc8_8_11)*uc8_12_15 - rdplight[l];
			vxyzw = vxyzw*vxyzw;
			float invlen = 65536.0f/(vxyzw[0]+vxyzw[1]+vxyzw[2]+vxyzw[3]);
            float p_i = rdp.light[l].ca * invlen;
			#else
            float vx = (v->x + uc8_coord_mod[8])*uc8_coord_mod[12] - rdp.light[l].x;
            float vy = (v->y + uc8_coord_mod[9])*uc8_coord_mod[13] - rdp.light[l].y;
            float vz = (v->z + uc8_coord_mod[10])*uc8_coord_mod[14] - rdp.light[l].z;
            float vw = (v->w + uc8_coord_mod[11])*uc8_coord_mod[15] - rdp.light[l].w;
            float len = (vx*vx+vy*vy+vz*vz+vw*vw)/65536.0f;
            float p_i = rdp.light[l].ca / len;
			#endif
            if (p_i > 1.0f) p_i = 1.0f;
            light_intensity *= p_i;
			#if __ARM_NEON__
            FRDP("light %d, 1/len: %f, p_intensity : %f\n", l, invlen, p_i);
			#else
            FRDP("light %d, len: %f, p_intensity : %f\n", l, len, p_i);
			#endif
          }
          //*/
		  #ifdef __ARM_NEON__
		  color += rdpcolor[l] * light_intensity;
		  #else
          color[0] += rdp.light[l].r * light_intensity;
          color[1] += rdp.light[l].g * light_intensity;
          color[2] += rdp.light[l].b * light_intensity;
		  #endif
          FRDP("light %d r: %f, g: %f, b: %f\n", l, color[0], color[1], color[2]);
        }
        light_intensity = DotProduct (rdp.light_vector[l], v->vec);
        FRDP("light %d, intensity : %f\n", l, light_intensity);
        if (light_intensity > 0.0f)
        {
		  #ifdef __ARM_NEON__
		  color += rdpcolor[l] * light_intensity;
		  #else
          color[0] += rdp.light[l].r * light_intensity;
          color[1] += rdp.light[l].g * light_intensity;
          color[2] += rdp.light[l].b * light_intensity;
		  #endif
        }
        FRDP("light %d r: %f, g: %f, b: %f\n", l, color[0], color[1], color[2]);
      }
			else
			{
				for (l = 0; l < rdp.num_lights; l++)
				{
					if (rdp.light[l].nonblack && rdp.light[l].nonzero)
					{
						#ifdef __ARM_NEON__
						float32x4_t vxyzw = (v_xyzw + uc8_8_11)*uc8_12_15 - rdplight[l];
						vxyzw = vxyzw*vxyzw;
						float invlen = 65536.0f/(vxyzw[0]+vxyzw[1]+vxyzw[2]+vxyzw[3]);
						light_intensity = rdp.light[l].ca * invlen;
						#else
						float vx = (v->x + uc8_coord_mod[8])*uc8_coord_mod[12] - rdp.light[l].x;
						float vy = (v->y + uc8_coord_mod[9])*uc8_coord_mod[13] - rdp.light[l].y;
						float vz = (v->z + uc8_coord_mod[10])*uc8_coord_mod[14] - rdp.light[l].z;
						float vw = (v->w + uc8_coord_mod[11])*uc8_coord_mod[15] - rdp.light[l].w;
						float len = (vx*vx+vy*vy+vz*vz+vw*vw)/65536.0f;
						light_intensity = rdp.light[l].ca / len;
						#endif
						if (light_intensity > 1.0f) light_intensity = 1.0f;
						FRDP("light %d, p_intensity : %f\n", l, light_intensity);
					    #ifdef __ARM_NEON__
						color += rdpcolor[l] * light_intensity;
						#else
						color[0] += rdp.light[l].r * light_intensity;
						color[1] += rdp.light[l].g * light_intensity;
						color[2] += rdp.light[l].b * light_intensity;
						#endif
						//FRDP("light %d r: %f, g: %f, b: %f\n", l, color[0], color[1], color[2]);
					}
				}
			}
			#ifdef __ARM_NEON__
			color = vminq_f32(color, unite);
			#else
			if (color[0] > 1.0f) color[0] = 1.0f;
			if (color[1] > 1.0f) color[1] = 1.0f;
			if (color[2] > 1.0f) color[2] = 1.0f;
			#endif
			v->r = (wxUint8)(((float)v->r)*color[0]);
			v->g = (wxUint8)(((float)v->g)*color[1]);
			v->b = (wxUint8)(((float)v->b)*color[2]);
#ifdef EXTREME_LOGGING
		FRDP("color after light: r: 0x%02lx, g: 0x%02lx, b: 0x%02lx\n", v->r, v->g, v->b);
#endif
		}
  }
}

static void uc8_moveword ()
{
	wxUint8 index = (wxUint8)((rdp.cmd0 >> 16) & 0xFF);
	wxUint16 offset = (wxUint16)(rdp.cmd0 & 0xFFFF);
	wxUint32 data = rdp.cmd1;

	FRDP ("uc8:moveword ");

	switch (index)
	{
		// NOTE: right now it's assuming that it sets the integer part first.  This could
		//  be easily fixed, but only if i had something to test with.

	case 0x02:
		rdp.num_lights = (data / 48);
		rdp.update |= UPDATE_LIGHTS;
		FRDP ("numlights: %d\n", rdp.num_lights);
		break;

	case 0x04:
    if (offset == 0x04)
    {
      rdp.clip_ratio = sqrt((float)rdp.cmd1);
      rdp.update |= UPDATE_VIEWPORT;
    }
		FRDP ("mw_clip %08lx, %08lx\n", rdp.cmd0, rdp.cmd1);
		break;

	case 0x06:  // moveword SEGMENT
		{
			FRDP ("SEGMENT %08lx -> seg%d\n", data, offset >> 2);
			rdp.segment[(offset >> 2) & 0xF] = data;
		}
		break;

	case 0x08:
		{
			rdp.fog_multiplier = (short)(rdp.cmd1 >> 16);
			rdp.fog_offset = (short)(rdp.cmd1 & 0x0000FFFF);
			FRDP ("fog: multiplier: %f, offset: %f\n", rdp.fog_multiplier, rdp.fog_offset);
		}
		break;

	case 0x0c:
		RDP_E ("uc8:moveword forcemtx - IGNORED\n");
		LRDP("forcemtx - IGNORED\n");
		break;

	case 0x0e:
		LRDP("perspnorm - IGNORED\n");
		break;

	case 0x10:  // moveword coord mod
		{
			wxUint8 n = offset >> 2;

			FRDP ("coord mod:%d, %08lx\n", n, data);
			if (rdp.cmd0&8)
				return;
			wxUint32 idx = (rdp.cmd0>>1)&3;
			wxUint32 pos = rdp.cmd0&0x30;
			if (pos == 0)
			{
				uc8_coord_mod[0+idx] = (short)(rdp.cmd1>>16);
				uc8_coord_mod[1+idx] = (short)(rdp.cmd1&0xffff);
			}
			else if (pos == 0x10)
			{
				uc8_coord_mod[4+idx] = (rdp.cmd1>>16)/65536.0f;
				uc8_coord_mod[5+idx] = (rdp.cmd1&0xffff)/65536.0f;
				uc8_coord_mod[12+idx] = uc8_coord_mod[0+idx] + uc8_coord_mod[4+idx];
				uc8_coord_mod[13+idx] = uc8_coord_mod[1+idx] + uc8_coord_mod[5+idx];

			}
			else if (pos == 0x20)
			{
				uc8_coord_mod[8+idx] = (short)(rdp.cmd1>>16);
				uc8_coord_mod[9+idx] = (short)(rdp.cmd1&0xffff);
#ifdef EXTREME_LOGGING
				if (idx)
				{
					for (int k = 8; k < 16; k++)
					{
						FRDP("coord_mod[%d]=%f\n", k, uc8_coord_mod[k]);
					}
				}
#endif
			}

		}
		break;

	default:
		FRDP_E("uc8:moveword unknown (index: 0x%08lx, offset 0x%08lx)\n", index, offset);
		FRDP ("unknown (index: 0x%08lx, offset 0x%08lx)\n", index, offset);
  }
}

static void uc8_movemem ()
{
	int idx = rdp.cmd0 & 0xFF;
	wxUint32 addr = segoffset(rdp.cmd1);
	int ofs = (rdp.cmd0 >> 5) & 0x3FFF;

	FRDP ("uc8:movemem ofs:%d ", ofs);

	switch (idx)
	{
	case 8:   // VIEWPORT
		{
			wxUint32 a = addr >> 1;
			short scale_x = ((short*)gfx.RDRAM)[(a+0)^1] >> 2;
			short scale_y = ((short*)gfx.RDRAM)[(a+1)^1] >> 2;
			short scale_z = ((short*)gfx.RDRAM)[(a+2)^1];
			short trans_x = ((short*)gfx.RDRAM)[(a+4)^1] >> 2;
			short trans_y = ((short*)gfx.RDRAM)[(a+5)^1] >> 2;
			short trans_z = ((short*)gfx.RDRAM)[(a+6)^1];
			rdp.view_scale[0] = scale_x * rdp.scale_x;
			rdp.view_scale[1] = -scale_y * rdp.scale_y;
			rdp.view_scale[2] = 32.0f * scale_z;
			rdp.view_trans[0] = trans_x * rdp.scale_x;
			rdp.view_trans[1] = trans_y * rdp.scale_y;
			rdp.view_trans[2] = 32.0f * trans_z;

			rdp.update |= UPDATE_VIEWPORT;

			FRDP ("viewport scale(%d, %d), trans(%d, %d), from:%08lx\n", scale_x, scale_y,
				trans_x, trans_y, a);
		}
		break;

	case 10:  // LIGHT
		{
			int n = (ofs / 48);
            if (n < 2)
            {
              char dir_x = ((char*)gfx.RDRAM)[(addr+8)^3];
              rdp.lookat[n][0] = (float)(dir_x) / 127.0f;
              char dir_y = ((char*)gfx.RDRAM)[(addr+9)^3];
              rdp.lookat[n][1] = (float)(dir_y) / 127.0f;
              char dir_z = ((char*)gfx.RDRAM)[(addr+10)^3];
              rdp.lookat[n][2] = (float)(dir_z) / 127.0f;
              rdp.use_lookat = TRUE;
              if (n == 1)
              {
                 if (!dir_x && !dir_y)
                   rdp.use_lookat = FALSE;
              }
              FRDP("lookat_%d (%f, %f, %f)\n", n, rdp.lookat[n][0], rdp.lookat[n][1], rdp.lookat[n][2]);
              return;
            }
            n -= 2;
			wxUint8 col = gfx.RDRAM[(addr+0)^3];
			rdp.light[n].r = (float)col / 255.0f;
			rdp.light[n].nonblack = col;
			col = gfx.RDRAM[(addr+1)^3];
			rdp.light[n].g = (float)col / 255.0f;
			rdp.light[n].nonblack += col;
			col = gfx.RDRAM[(addr+2)^3];
			rdp.light[n].b = (float)col / 255.0f;
			rdp.light[n].nonblack += col;
			rdp.light[n].a = 1.0f;
			rdp.light[n].dir_x = (float)(((char*)gfx.RDRAM)[(addr+8)^3]) / 127.0f;
			rdp.light[n].dir_y = (float)(((char*)gfx.RDRAM)[(addr+9)^3]) / 127.0f;
			rdp.light[n].dir_z = (float)(((char*)gfx.RDRAM)[(addr+10)^3]) / 127.0f;
			// **
			wxUint32 a = addr >> 1;
			rdp.light[n].x = (float)(((short*)gfx.RDRAM)[(a+16)^1]);
			rdp.light[n].y = (float)(((short*)gfx.RDRAM)[(a+17)^1]);
			rdp.light[n].z = (float)(((short*)gfx.RDRAM)[(a+18)^1]);
			rdp.light[n].w = (float)(((short*)gfx.RDRAM)[(a+19)^1]);
			rdp.light[n].nonzero = gfx.RDRAM[(addr+12)^3];
			rdp.light[n].ca = (float)rdp.light[n].nonzero / 16.0f;
			//rdp.light[n].la = rdp.light[n].ca * 1.0f;
#ifdef EXTREME_LOGGING
			FRDP ("light: n: %d, pos: x: %f, y: %f, z: %f, w: %f, ca: %f\n",
				n, rdp.light[n].x, rdp.light[n].y, rdp.light[n].z, rdp.light[n].w, rdp.light[n].ca);
#endif
			FRDP ("light: n: %d, r: %f, g: %f, b: %f. dir: x: %.3f, y: %.3f, z: %.3f\n",
				n, rdp.light[n].r, rdp.light[n].g, rdp.light[n].b,
				rdp.light[n].dir_x, rdp.light[n].dir_y, rdp.light[n].dir_z);
#ifdef EXTREME_LOGGING
			for (int t=0; t < 24; t++)
			{
				FRDP ("light[%d] = 0x%04lx \n", t, ((wxUint16*)gfx.RDRAM)[(a+t)^1]);
			}
#endif
		}
		break;

	case 14: //Normales
		{
			uc8_normale_addr = segoffset(rdp.cmd1);
			FRDP ("Normale - addr: %08lx\n", uc8_normale_addr);
#ifdef EXTREME_LOGGING
      int i;
			for (i = 0; i < 32; i++)
			{
				char x = ((char*)gfx.RDRAM)[uc8_normale_addr + ((i<<1) + 0)^3];
				char y = ((char*)gfx.RDRAM)[uc8_normale_addr + ((i<<1) + 1)^3];
				FRDP("#%d x = %d, y = %d\n", i, x, y);
			}
			wxUint32 a = uc8_normale_addr >> 1;
			for (i = 0; i < 32; i++)
			{
				FRDP ("n[%d] = 0x%04lx \n", i, ((wxUint16*)gfx.RDRAM)[(a+i)^1]);
			}
#endif
		}
		break;

	default:
		FRDP ("uc8:movemem unknown (%d)\n", idx);
	}
}


static void uc8_tri4() //by Gugaman Apr 19 2002
{
Check_FrameSkip;

    if (rdp.skip_drawing)
    {
		LRDP("uc8:tri4. skipped\n");
		return;
    }

	FRDP("uc8:tri4 (#%d - #%d), %d-%d-%d, %d-%d-%d, %d-%d-%d, %d-%d-%d\n",
		rdp.tri_n,
		rdp.tri_n+3,
		((rdp.cmd0 >> 23) & 0x1F),
		((rdp.cmd0 >> 18) & 0x1F),
		((((rdp.cmd0 >> 15) & 0x7) << 2) | ((rdp.cmd1 >> 30) &0x3)),
		((rdp.cmd0 >> 10) & 0x1F),
		((rdp.cmd0 >> 5) & 0x1F),
		((rdp.cmd0 >> 0) & 0x1F),
		((rdp.cmd1 >> 25) & 0x1F),
		((rdp.cmd1 >> 20) & 0x1F),
		((rdp.cmd1 >> 15) & 0x1F),
		((rdp.cmd1 >> 10) & 0x1F),
		((rdp.cmd1 >> 5) & 0x1F),
		((rdp.cmd1 >> 0) & 0x1F));

	VERTEX *v[12] = {
		&rdp.vtx[(rdp.cmd0 >> 23) & 0x1F],
			&rdp.vtx[(rdp.cmd0 >> 18) & 0x1F],
			&rdp.vtx[((((rdp.cmd0 >> 15) & 0x7) << 2) | ((rdp.cmd1 >> 30) &0x3))],
			&rdp.vtx[(rdp.cmd0 >> 10) & 0x1F],
			&rdp.vtx[(rdp.cmd0 >> 5) & 0x1F],
			&rdp.vtx[(rdp.cmd0 >> 0) & 0x1F],
			&rdp.vtx[(rdp.cmd1 >> 25) & 0x1F],
			&rdp.vtx[(rdp.cmd1 >> 20) & 0x1F],
			&rdp.vtx[(rdp.cmd1 >> 15) & 0x1F],
			&rdp.vtx[(rdp.cmd1 >> 10) & 0x1F],
			&rdp.vtx[(rdp.cmd1 >> 5) & 0x1F],
			&rdp.vtx[(rdp.cmd1 >> 0) & 0x1F]
	};

	int updated = 0;

	if (cull_tri(v))
		rdp.tri_n ++;
	else
	{
		updated = 1;
		update ();

		draw_tri (v);
		rdp.tri_n ++;
	}

	if (cull_tri(v+3))
		rdp.tri_n ++;
	else
	{
		if (!updated)
		{
			updated = 1;
			update ();
		}

		draw_tri (v+3);
		rdp.tri_n ++;
	}

	if (cull_tri(v+6))
		rdp.tri_n ++;
	else
	{
		if (!updated)
		{
			updated = 1;
			update ();
		}

		draw_tri (v+6);
		rdp.tri_n ++;
	}

	if (cull_tri(v+9))
		rdp.tri_n ++;
	else
	{
		if (!updated)
		{
			updated = 1;
			update ();
		}

		draw_tri (v+9);
		rdp.tri_n ++;
	}
}
Commit	Line	Data
	1	/*
	2	* Glide64 - Glide video plugin for Nintendo 64 emulators.
	3	* Copyright (c) 2002 Dave2001
	4	* Copyright (c) 2003-2009 Sergey 'Gonetz' Lipski
	5	*
	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
	9	* any later version.
	10	*
	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.
	15	*
	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
	19	*/
	20
	21	//****************************************************************
	22	//
	23	// Glide64 - Glide Plugin for Nintendo 64 emulators
	24	// Project started on December 29th, 2001
	25	//
	26	// Authors:
	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
	31	//
	32	//****************************************************************
	33	//
	34	// To modify Glide64:
	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.
	37	//
	38	//****************************************************************
	39	//
	40	// January 2004 Created by Gonetz (Gonetz@ngs.ru)
	41	//
	42	//****************************************************************
	43
	44	wxUint32 uc8_normale_addr = 0;
	45	float uc8_coord_mod[16];
	46
	47	static void uc8_vertex ()
	48	{
	49	Check_FrameSkip;
	50
	51	if (rdp.update & UPDATE_MULT_MAT)
	52	{
	53	rdp.update ^= UPDATE_MULT_MAT;
	54	MulMatrices(rdp.model, rdp.proj, rdp.combined);
	55	}
	56
	57	wxUint32 addr = segoffset(rdp.cmd1);
	58	int v0, i, n;
	59	float x, y, z;
	60
	61	rdp.vn = n = (rdp.cmd0 >> 12) & 0xFF;
	62	rdp.v0 = v0 = ((rdp.cmd0 >> 1) & 0x7F) - n;
	63
	64	FRDP ("uc8:vertex n: %d, v0: %d, from: %08lx\n", n, v0, addr);
	65
	66	if (v0 < 0)
	67	{
	68	RDP_E ("** ERROR: uc2:vertex v0 < 0\n");
	69	LRDP("** ERROR: uc2:vertex v0 < 0\n");
	70	return;
	71	}
	72	//*
	73	// This is special, not handled in update()
	74	if (rdp.update & UPDATE_LIGHTS)
	75	{
	76	rdp.update ^= UPDATE_LIGHTS;
	77
	78	// Calculate light vectors
	79	for (wxUint32 l=0; l<rdp.num_lights; l++)
	80	{
	81	InverseTransformVector(&rdp.light[l].dir_x, rdp.light_vector[l], rdp.model);
	82	NormalizeVector (rdp.light_vector[l]);
	83	#ifdef EXTREME_LOGGING
	84	FRDP("light_vector[%d] x: %f, y: %f, z: %f\n", l, rdp.light_vector[l][0], rdp.light_vector[l][1], rdp.light_vector[l][2]);
	85	#endif
	86	}
	87	}
	88	//*/
	89	#ifdef __ARM_NEON__
	90	float32x4_t comb0, comb1, comb2, comb3;
	91	float32x4_t v_xyzw;
	92	comb0 = vld1q_f32(rdp.combined[0]);
	93	comb1 = vld1q_f32(rdp.combined[1]);
	94	comb2 = vld1q_f32(rdp.combined[2]);
	95	comb3 = vld1q_f32(rdp.combined[3]);
	96
	97	float32x4_t uc8_8_11, uc8_12_15;
	98	uc8_8_11 = vld1q_f32(uc8_coord_mod+8);
	99	uc8_12_15 = vld1q_f32(uc8_coord_mod+12);
	100	float32x4_t unite={1.0f, 1.0f, 1.0f, 1.0f};
	101
	102	float32x4_t rdplight[12], rdpcolor[12];
	103	for (i=0; i<rdp.num_lights; i++) {
	104	rdplight[i] = vld1q_f32(&rdp.light[i].x);
	105	rdpcolor[i] = vld1q_f32(&rdp.light[i].r);
	106	}
	107
	108	#endif
	109	for (i=0; i < (n<<4); i+=16)
	110	{
	111	VERTEX *v = &rdp.vtx[v0 + (i>>4)];
	112	x = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 0)^1];
	113	y = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 1)^1];
	114	z = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 2)^1];
	115	v->flags = ((wxUint16*)gfx.RDRAM)[(((addr+i) >> 1) + 3)^1];
	116	v->ou = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 4)^1];
	117	v->ov = (float)((short*)gfx.RDRAM)[(((addr+i) >> 1) + 5)^1];
	118	v->uv_scaled = 0;
	119	v->a = ((wxUint8*)gfx.RDRAM)[(addr+i + 15)^3];
	120
	121	#ifdef EXTREME_LOGGING
	122	FRDP ("before v%d - x: %f, y: %f, z: %f\n", i>>4, x, y, z);
	123	#endif
	124	#ifdef __ARM_NEON__
	125	v_xyzw = xcomb0+ycomb1+z*comb2+comb3;
	126	//vst1q_f32((float*)v, v_xyzw);
	127	#else
	128	v->x = xrdp.combined[0][0] + yrdp.combined[1][0] + z*rdp.combined[2][0] + rdp.combined[3][0];
	129	v->y = xrdp.combined[0][1] + yrdp.combined[1][1] + z*rdp.combined[2][1] + rdp.combined[3][1];
	130	v->z = xrdp.combined[0][2] + yrdp.combined[1][2] + z*rdp.combined[2][2] + rdp.combined[3][2];
	131	v->w = xrdp.combined[0][3] + yrdp.combined[1][3] + z*rdp.combined[2][3] + rdp.combined[3][3];
	132	#endif
	133
	134	#ifdef EXTREME_LOGGING
	135	FRDP ("v%d - x: %f, y: %f, z: %f, w: %f, u: %f, v: %f, flags: %d\n", i>>4, v->x, v->y, v->z, v->w, v->ou, v->ov, v->flags);
	136	#endif
	137	v->uv_calculated = 0xFFFFFFFF;
	138	v->screen_translated = 0;
	139	v->shade_mod = 0;
	140	#ifdef __ARM_NEON__
	141	v->x=v_xyzw[0];
	142	v->y=v_xyzw[1];
	143	v->z=v_xyzw[2];
	144	v->w=v_xyzw[3];
	145	#endif
	146
	147	///*
	148	v->r = ((wxUint8*)gfx.RDRAM)[(addr+i + 12)^3];
	149	v->g = ((wxUint8*)gfx.RDRAM)[(addr+i + 13)^3];
	150	v->b = ((wxUint8*)gfx.RDRAM)[(addr+i + 14)^3];
	151
	152	if (fabs(v->w) < 0.001) v->w = 0.001f;
	153	v->oow = 1.0f / v->w;
	154	#ifdef __ARM_NEON__
	155	v_xyzw *= v->oow;
	156	v->x_w=v_xyzw[0];
	157	v->y_w=v_xyzw[1];
	158	v->z_w=v_xyzw[2];
	159	#else
	160	v->x_w = v->x * v->oow;
	161	v->y_w = v->y * v->oow;
	162	v->z_w = v->z * v->oow;
	163	#endif
	164
	165	v->scr_off = 0;
	166	if (v->x < -v->w) v->scr_off \|= 1;
	167	if (v->x > v->w) v->scr_off \|= 2;
	168	if (v->y < -v->w) v->scr_off \|= 4;
	169	if (v->y > v->w) v->scr_off \|= 8;
	170	if (v->w < 0.1f) v->scr_off \|= 16;
	171	#ifdef EXTREME_LOGGING
	172	FRDP ("r: %02lx, g: %02lx, b: %02lx, a: %02lx\n", v->r, v->g, v->b, v->a);
	173	#endif
	174
	175	if ((rdp.geom_mode & 0x00020000))
	176	{
	177	wxUint32 shift = v0 << 1;
	178	v->vec[0] = ((char*)gfx.RDRAM)[(uc8_normale_addr + (i>>3) + shift + 0)^3];
	179	v->vec[1] = ((char*)gfx.RDRAM)[(uc8_normale_addr + (i>>3) + shift + 1)^3];
	180	v->vec[2] = (char)(v->flags&0xff);
	181
	182	if (rdp.geom_mode & 0x80000)
	183	{
	184	calc_linear (v);
	185	#ifdef EXTREME_LOGGING
	186	FRDP ("calc linear: v%d - u: %f, v: %f\n", i>>4, v->ou, v->ov);
	187	#endif
	188	}
	189	else if (rdp.geom_mode & 0x40000)
	190	{
	191	calc_sphere (v);
	192	#ifdef EXTREME_LOGGING
	193	FRDP ("calc sphere: v%d - u: %f, v: %f\n", i>>4, v->ou, v->ov);
	194	#endif
	195	}
	196	// FRDP("calc light. r: 0x%02lx, g: 0x%02lx, b: 0x%02lx, nx: %.3f, ny: %.3f, nz: %.3f\n", v->r, v->g, v->b, v->vec[0], v->vec[1], v->vec[2]);
	197	FRDP("v[%d] calc light. r: 0x%02lx, g: 0x%02lx, b: 0x%02lx\n", i>>4, v->r, v->g, v->b);
	198	#ifdef __ARM_NEON__
	199	float32x4_t color = {rdp.light[rdp.num_lights].r, rdp.light[rdp.num_lights].g, rdp.light[rdp.num_lights].b, 1.0f};
	200	#else
	201	float color[3] = {rdp.light[rdp.num_lights].r, rdp.light[rdp.num_lights].g, rdp.light[rdp.num_lights].b};
	202	#endif
	203	FRDP("ambient light. r: %f, g: %f, b: %f\n", color[0], color[1], color[2]);
	204	float light_intensity = 0.0f;
	205	wxUint32 l;
	206	if (rdp.geom_mode & 0x00400000)
	207	{
	208	NormalizeVector (v->vec);
	209	for (l = 0; l < rdp.num_lights-1; l++)
	210	{
	211	if (!rdp.light[l].nonblack)
	212	continue;
	213	light_intensity = DotProduct (rdp.light_vector[l], v->vec);
	214	FRDP("light %d, intensity : %f\n", l, light_intensity);
	215	if (light_intensity < 0.0f)
	216	continue;
	217	//*
	218	if (rdp.light[l].ca > 0.0f)
	219	{
	220	#ifdef __ARM_NEON__
	221	float32x4_t vxyzw = (v_xyzw + uc8_8_11)*uc8_12_15 - rdplight[l];
	222	vxyzw = vxyzw*vxyzw;
	223	float invlen = 65536.0f/(vxyzw[0]+vxyzw[1]+vxyzw[2]+vxyzw[3]);
	224	float p_i = rdp.light[l].ca * invlen;
	225	#else
	226	float vx = (v->x + uc8_coord_mod[8])*uc8_coord_mod[12] - rdp.light[l].x;
	227	float vy = (v->y + uc8_coord_mod[9])*uc8_coord_mod[13] - rdp.light[l].y;
	228	float vz = (v->z + uc8_coord_mod[10])*uc8_coord_mod[14] - rdp.light[l].z;
	229	float vw = (v->w + uc8_coord_mod[11])*uc8_coord_mod[15] - rdp.light[l].w;
	230	float len = (vxvx+vyvy+vzvz+vwvw)/65536.0f;
	231	float p_i = rdp.light[l].ca / len;
	232	#endif
	233	if (p_i > 1.0f) p_i = 1.0f;
	234	light_intensity *= p_i;
	235	#if __ARM_NEON__
	236	FRDP("light %d, 1/len: %f, p_intensity : %f\n", l, invlen, p_i);
	237	#else
	238	FRDP("light %d, len: %f, p_intensity : %f\n", l, len, p_i);
	239	#endif
	240	}
	241	//*/
	242	#ifdef __ARM_NEON__
	243	color += rdpcolor[l] * light_intensity;
	244	#else
	245	color[0] += rdp.light[l].r * light_intensity;
	246	color[1] += rdp.light[l].g * light_intensity;
	247	color[2] += rdp.light[l].b * light_intensity;
	248	#endif
	249	FRDP("light %d r: %f, g: %f, b: %f\n", l, color[0], color[1], color[2]);
	250	}
	251	light_intensity = DotProduct (rdp.light_vector[l], v->vec);
	252	FRDP("light %d, intensity : %f\n", l, light_intensity);
	253	if (light_intensity > 0.0f)
	254	{
	255	#ifdef __ARM_NEON__
	256	color += rdpcolor[l] * light_intensity;
	257	#else
	258	color[0] += rdp.light[l].r * light_intensity;
	259	color[1] += rdp.light[l].g * light_intensity;
	260	color[2] += rdp.light[l].b * light_intensity;
	261	#endif
	262	}
	263	FRDP("light %d r: %f, g: %f, b: %f\n", l, color[0], color[1], color[2]);
	264	}
	265	else
	266	{
	267	for (l = 0; l < rdp.num_lights; l++)
	268	{
	269	if (rdp.light[l].nonblack && rdp.light[l].nonzero)
	270	{
	271	#ifdef __ARM_NEON__
	272	float32x4_t vxyzw = (v_xyzw + uc8_8_11)*uc8_12_15 - rdplight[l];
	273	vxyzw = vxyzw*vxyzw;
	274	float invlen = 65536.0f/(vxyzw[0]+vxyzw[1]+vxyzw[2]+vxyzw[3]);
	275	light_intensity = rdp.light[l].ca * invlen;
	276	#else
	277	float vx = (v->x + uc8_coord_mod[8])*uc8_coord_mod[12] - rdp.light[l].x;
	278	float vy = (v->y + uc8_coord_mod[9])*uc8_coord_mod[13] - rdp.light[l].y;
	279	float vz = (v->z + uc8_coord_mod[10])*uc8_coord_mod[14] - rdp.light[l].z;
	280	float vw = (v->w + uc8_coord_mod[11])*uc8_coord_mod[15] - rdp.light[l].w;
	281	float len = (vxvx+vyvy+vzvz+vwvw)/65536.0f;
	282	light_intensity = rdp.light[l].ca / len;
	283	#endif
	284	if (light_intensity > 1.0f) light_intensity = 1.0f;
	285	FRDP("light %d, p_intensity : %f\n", l, light_intensity);
	286	#ifdef __ARM_NEON__
	287	color += rdpcolor[l] * light_intensity;
	288	#else
	289	color[0] += rdp.light[l].r * light_intensity;
	290	color[1] += rdp.light[l].g * light_intensity;
	291	color[2] += rdp.light[l].b * light_intensity;
	292	#endif
	293	//FRDP("light %d r: %f, g: %f, b: %f\n", l, color[0], color[1], color[2]);
	294	}
	295	}
	296	}
	297	#ifdef __ARM_NEON__
	298	color = vminq_f32(color, unite);
	299	#else
	300	if (color[0] > 1.0f) color[0] = 1.0f;
	301	if (color[1] > 1.0f) color[1] = 1.0f;
	302	if (color[2] > 1.0f) color[2] = 1.0f;
	303	#endif
	304	v->r = (wxUint8)(((float)v->r)*color[0]);
	305	v->g = (wxUint8)(((float)v->g)*color[1]);
	306	v->b = (wxUint8)(((float)v->b)*color[2]);
	307	#ifdef EXTREME_LOGGING
	308	FRDP("color after light: r: 0x%02lx, g: 0x%02lx, b: 0x%02lx\n", v->r, v->g, v->b);
	309	#endif
	310	}
	311	}
	312	}
	313
	314	static void uc8_moveword ()
	315	{
	316	wxUint8 index = (wxUint8)((rdp.cmd0 >> 16) & 0xFF);
	317	wxUint16 offset = (wxUint16)(rdp.cmd0 & 0xFFFF);
	318	wxUint32 data = rdp.cmd1;
	319
	320	FRDP ("uc8:moveword ");
	321
	322	switch (index)
	323	{
	324	// NOTE: right now it's assuming that it sets the integer part first. This could
	325	// be easily fixed, but only if i had something to test with.
	326
	327	case 0x02:
	328	rdp.num_lights = (data / 48);
	329	rdp.update \|= UPDATE_LIGHTS;
	330	FRDP ("numlights: %d\n", rdp.num_lights);
	331	break;
	332
	333	case 0x04:
	334	if (offset == 0x04)
	335	{
	336	rdp.clip_ratio = sqrt((float)rdp.cmd1);
	337	rdp.update \|= UPDATE_VIEWPORT;
	338	}
	339	FRDP ("mw_clip %08lx, %08lx\n", rdp.cmd0, rdp.cmd1);
	340	break;
	341
	342	case 0x06: // moveword SEGMENT
	343	{
	344	FRDP ("SEGMENT %08lx -> seg%d\n", data, offset >> 2);
	345	rdp.segment[(offset >> 2) & 0xF] = data;
	346	}
	347	break;
	348
	349	case 0x08:
	350	{
	351	rdp.fog_multiplier = (short)(rdp.cmd1 >> 16);
	352	rdp.fog_offset = (short)(rdp.cmd1 & 0x0000FFFF);
	353	FRDP ("fog: multiplier: %f, offset: %f\n", rdp.fog_multiplier, rdp.fog_offset);
	354	}
	355	break;
	356
	357	case 0x0c:
	358	RDP_E ("uc8:moveword forcemtx - IGNORED\n");
	359	LRDP("forcemtx - IGNORED\n");
	360	break;
	361
	362	case 0x0e:
	363	LRDP("perspnorm - IGNORED\n");
	364	break;
	365
	366	case 0x10: // moveword coord mod
	367	{
	368	wxUint8 n = offset >> 2;
	369
	370	FRDP ("coord mod:%d, %08lx\n", n, data);
	371	if (rdp.cmd0&8)
	372	return;
	373	wxUint32 idx = (rdp.cmd0>>1)&3;
	374	wxUint32 pos = rdp.cmd0&0x30;
	375	if (pos == 0)
	376	{
	377	uc8_coord_mod[0+idx] = (short)(rdp.cmd1>>16);
	378	uc8_coord_mod[1+idx] = (short)(rdp.cmd1&0xffff);
	379	}
	380	else if (pos == 0x10)
	381	{
	382	uc8_coord_mod[4+idx] = (rdp.cmd1>>16)/65536.0f;
	383	uc8_coord_mod[5+idx] = (rdp.cmd1&0xffff)/65536.0f;
	384	uc8_coord_mod[12+idx] = uc8_coord_mod[0+idx] + uc8_coord_mod[4+idx];
	385	uc8_coord_mod[13+idx] = uc8_coord_mod[1+idx] + uc8_coord_mod[5+idx];
	386
	387	}
	388	else if (pos == 0x20)
	389	{
	390	uc8_coord_mod[8+idx] = (short)(rdp.cmd1>>16);
	391	uc8_coord_mod[9+idx] = (short)(rdp.cmd1&0xffff);
	392	#ifdef EXTREME_LOGGING
	393	if (idx)
	394	{
	395	for (int k = 8; k < 16; k++)
	396	{
	397	FRDP("coord_mod[%d]=%f\n", k, uc8_coord_mod[k]);
	398	}
	399	}
	400	#endif
	401	}
	402
	403	}
	404	break;
	405
	406	default:
	407	FRDP_E("uc8:moveword unknown (index: 0x%08lx, offset 0x%08lx)\n", index, offset);
	408	FRDP ("unknown (index: 0x%08lx, offset 0x%08lx)\n", index, offset);
	409	}
	410	}
	411
	412	static void uc8_movemem ()
	413	{
	414	int idx = rdp.cmd0 & 0xFF;
	415	wxUint32 addr = segoffset(rdp.cmd1);
	416	int ofs = (rdp.cmd0 >> 5) & 0x3FFF;
	417
	418	FRDP ("uc8:movemem ofs:%d ", ofs);
	419
	420	switch (idx)
	421	{
	422	case 8: // VIEWPORT
	423	{
	424	wxUint32 a = addr >> 1;
	425	short scale_x = ((short*)gfx.RDRAM)[(a+0)^1] >> 2;
	426	short scale_y = ((short*)gfx.RDRAM)[(a+1)^1] >> 2;
	427	short scale_z = ((short*)gfx.RDRAM)[(a+2)^1];
	428	short trans_x = ((short*)gfx.RDRAM)[(a+4)^1] >> 2;
	429	short trans_y = ((short*)gfx.RDRAM)[(a+5)^1] >> 2;
	430	short trans_z = ((short*)gfx.RDRAM)[(a+6)^1];
	431	rdp.view_scale[0] = scale_x * rdp.scale_x;
	432	rdp.view_scale[1] = -scale_y * rdp.scale_y;
	433	rdp.view_scale[2] = 32.0f * scale_z;
	434	rdp.view_trans[0] = trans_x * rdp.scale_x;
	435	rdp.view_trans[1] = trans_y * rdp.scale_y;
	436	rdp.view_trans[2] = 32.0f * trans_z;
	437
	438	rdp.update \|= UPDATE_VIEWPORT;
	439
	440	FRDP ("viewport scale(%d, %d), trans(%d, %d), from:%08lx\n", scale_x, scale_y,
	441	trans_x, trans_y, a);
	442	}
	443	break;
	444
	445	case 10: // LIGHT
	446	{
	447	int n = (ofs / 48);
	448	if (n < 2)
	449	{
	450	char dir_x = ((char*)gfx.RDRAM)[(addr+8)^3];
	451	rdp.lookat[n][0] = (float)(dir_x) / 127.0f;
	452	char dir_y = ((char*)gfx.RDRAM)[(addr+9)^3];
	453	rdp.lookat[n][1] = (float)(dir_y) / 127.0f;
	454	char dir_z = ((char*)gfx.RDRAM)[(addr+10)^3];
	455	rdp.lookat[n][2] = (float)(dir_z) / 127.0f;
	456	rdp.use_lookat = TRUE;
	457	if (n == 1)
	458	{
	459	if (!dir_x && !dir_y)
	460	rdp.use_lookat = FALSE;
	461	}
	462	FRDP("lookat_%d (%f, %f, %f)\n", n, rdp.lookat[n][0], rdp.lookat[n][1], rdp.lookat[n][2]);
	463	return;
	464	}
	465	n -= 2;
	466	wxUint8 col = gfx.RDRAM[(addr+0)^3];
	467	rdp.light[n].r = (float)col / 255.0f;
	468	rdp.light[n].nonblack = col;
	469	col = gfx.RDRAM[(addr+1)^3];
	470	rdp.light[n].g = (float)col / 255.0f;
	471	rdp.light[n].nonblack += col;
	472	col = gfx.RDRAM[(addr+2)^3];
	473	rdp.light[n].b = (float)col / 255.0f;
	474	rdp.light[n].nonblack += col;
	475	rdp.light[n].a = 1.0f;
	476	rdp.light[n].dir_x = (float)(((char*)gfx.RDRAM)[(addr+8)^3]) / 127.0f;
	477	rdp.light[n].dir_y = (float)(((char*)gfx.RDRAM)[(addr+9)^3]) / 127.0f;
	478	rdp.light[n].dir_z = (float)(((char*)gfx.RDRAM)[(addr+10)^3]) / 127.0f;
	479	// **
	480	wxUint32 a = addr >> 1;
	481	rdp.light[n].x = (float)(((short*)gfx.RDRAM)[(a+16)^1]);
	482	rdp.light[n].y = (float)(((short*)gfx.RDRAM)[(a+17)^1]);
	483	rdp.light[n].z = (float)(((short*)gfx.RDRAM)[(a+18)^1]);
	484	rdp.light[n].w = (float)(((short*)gfx.RDRAM)[(a+19)^1]);
	485	rdp.light[n].nonzero = gfx.RDRAM[(addr+12)^3];
	486	rdp.light[n].ca = (float)rdp.light[n].nonzero / 16.0f;
	487	//rdp.light[n].la = rdp.light[n].ca * 1.0f;
	488	#ifdef EXTREME_LOGGING
	489	FRDP ("light: n: %d, pos: x: %f, y: %f, z: %f, w: %f, ca: %f\n",
	490	n, rdp.light[n].x, rdp.light[n].y, rdp.light[n].z, rdp.light[n].w, rdp.light[n].ca);
	491	#endif
	492	FRDP ("light: n: %d, r: %f, g: %f, b: %f. dir: x: %.3f, y: %.3f, z: %.3f\n",
	493	n, rdp.light[n].r, rdp.light[n].g, rdp.light[n].b,
	494	rdp.light[n].dir_x, rdp.light[n].dir_y, rdp.light[n].dir_z);
	495	#ifdef EXTREME_LOGGING
	496	for (int t=0; t < 24; t++)
	497	{
	498	FRDP ("light[%d] = 0x%04lx \n", t, ((wxUint16*)gfx.RDRAM)[(a+t)^1]);
	499	}
	500	#endif
	501	}
	502	break;
	503
	504	case 14: //Normales
	505	{
	506	uc8_normale_addr = segoffset(rdp.cmd1);
	507	FRDP ("Normale - addr: %08lx\n", uc8_normale_addr);
	508	#ifdef EXTREME_LOGGING
	509	int i;
	510	for (i = 0; i < 32; i++)
	511	{
	512	char x = ((char*)gfx.RDRAM)[uc8_normale_addr + ((i<<1) + 0)^3];
	513	char y = ((char*)gfx.RDRAM)[uc8_normale_addr + ((i<<1) + 1)^3];
	514	FRDP("#%d x = %d, y = %d\n", i, x, y);
	515	}
	516	wxUint32 a = uc8_normale_addr >> 1;
	517	for (i = 0; i < 32; i++)
	518	{
	519	FRDP ("n[%d] = 0x%04lx \n", i, ((wxUint16*)gfx.RDRAM)[(a+i)^1]);
	520	}
	521	#endif
	522	}
	523	break;
	524
	525	default:
	526	FRDP ("uc8:movemem unknown (%d)\n", idx);
	527	}
	528	}
	529
	530
	531	static void uc8_tri4() //by Gugaman Apr 19 2002
	532	{
	533	Check_FrameSkip;
	534
	535	if (rdp.skip_drawing)
	536	{
	537	LRDP("uc8:tri4. skipped\n");
	538	return;
	539	}
	540
	541	FRDP("uc8:tri4 (#%d - #%d), %d-%d-%d, %d-%d-%d, %d-%d-%d, %d-%d-%d\n",
	542	rdp.tri_n,
	543	rdp.tri_n+3,
	544	((rdp.cmd0 >> 23) & 0x1F),
	545	((rdp.cmd0 >> 18) & 0x1F),
	546	((((rdp.cmd0 >> 15) & 0x7) << 2) \| ((rdp.cmd1 >> 30) &0x3)),
	547	((rdp.cmd0 >> 10) & 0x1F),
	548	((rdp.cmd0 >> 5) & 0x1F),
	549	((rdp.cmd0 >> 0) & 0x1F),
	550	((rdp.cmd1 >> 25) & 0x1F),
	551	((rdp.cmd1 >> 20) & 0x1F),
	552	((rdp.cmd1 >> 15) & 0x1F),
	553	((rdp.cmd1 >> 10) & 0x1F),
	554	((rdp.cmd1 >> 5) & 0x1F),
	555	((rdp.cmd1 >> 0) & 0x1F));
	556
	557	VERTEX *v[12] = {
	558	&rdp.vtx[(rdp.cmd0 >> 23) & 0x1F],
	559	&rdp.vtx[(rdp.cmd0 >> 18) & 0x1F],
	560	&rdp.vtx[((((rdp.cmd0 >> 15) & 0x7) << 2) \| ((rdp.cmd1 >> 30) &0x3))],
	561	&rdp.vtx[(rdp.cmd0 >> 10) & 0x1F],
	562	&rdp.vtx[(rdp.cmd0 >> 5) & 0x1F],
	563	&rdp.vtx[(rdp.cmd0 >> 0) & 0x1F],
	564	&rdp.vtx[(rdp.cmd1 >> 25) & 0x1F],
	565	&rdp.vtx[(rdp.cmd1 >> 20) & 0x1F],
	566	&rdp.vtx[(rdp.cmd1 >> 15) & 0x1F],
	567	&rdp.vtx[(rdp.cmd1 >> 10) & 0x1F],
	568	&rdp.vtx[(rdp.cmd1 >> 5) & 0x1F],
	569	&rdp.vtx[(rdp.cmd1 >> 0) & 0x1F]
	570	};
	571
	572	int updated = 0;
	573
	574	if (cull_tri(v))
	575	rdp.tri_n ++;
	576	else
	577	{
	578	updated = 1;
	579	update ();
	580
	581	draw_tri (v);
	582	rdp.tri_n ++;
	583	}
	584
	585	if (cull_tri(v+3))
	586	rdp.tri_n ++;
	587	else
	588	{
	589	if (!updated)
	590	{
	591	updated = 1;
	592	update ();
	593	}
	594
	595	draw_tri (v+3);
	596	rdp.tri_n ++;
	597	}
	598
	599	if (cull_tri(v+6))
	600	rdp.tri_n ++;
	601	else
	602	{
	603	if (!updated)
	604	{
	605	updated = 1;
	606	update ();
	607	}
	608
	609	draw_tri (v+6);
	610	rdp.tri_n ++;
	611	}
	612
	613	if (cull_tri(v+9))
	614	rdp.tri_n ++;
	615	else
	616	{
	617	if (!updated)
	618	{
	619	updated = 1;
	620	update ();
	621	}
	622
	623	draw_tri (v+9);
	624	rdp.tri_n ++;
	625	}
	626	}