dolphin/Source/Core/VideoCommon/Src/PixelShader.cpp

847 lines
34 KiB
C++

// Copyright (C) 2003-2008 Dolphin Project.
// 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, version 2.0.
// 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 2.0 for more details.
// A copy of the GPL 2.0 should have been included with the program.
// If not, see http://www.gnu.org/licenses/
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include <stdio.h>
#include <cmath>
#include <assert.h>
#include "Profiler.h"
#include "PixelShader.h"
#include "XFMemory.h" // for texture projection mode
#include "BPMemory.h"
// old tev->pixelshader notes
//
// color for this stage (alpha, color) is given by bpmem.tevorders[0].colorchan0
// konstant for this stage (alpha, color) is given by bpmem.tevksel
// inputs are given by bpmem.combiners[0].colorC.a/b/c/d << could be current chan color
// according to GXTevColorArg table above
// output is given by .outreg
// tevtemp is set according to swapmodetables and
static void WriteStage(char *&p, int n, u32 texture_mask);
static void WrapNonPow2Tex(char* &p, const char* var, int texmap, u32 texture_mask);
static void WriteAlphaCompare(char *&p, int num, int comp);
static bool WriteAlphaTest(char *&p);
const float epsilon8bit = 1.0f / 255.0f;
static const char *tevKSelTableC[] = // KCSEL
{
"1.0f,1.0f,1.0f", //1 = 0x00
"0.875,0.875,0.875",//7_8 = 0x01
"0.75,0.75,0.75", //3_4 = 0x02
"0.625,0.625,0.625",//5_8 = 0x03
"0.5,0.5,0.5", //1_2 = 0x04
"0.375,0.375,0.375",//3_8 = 0x05
"0.25,0.25,0.25", //1_4 = 0x06
"0.125,0.125,0.125",//1_8 = 0x07
"ERROR", //0x08
"ERROR", //0x09
"ERROR", //0x0a
"ERROR", //0x0b
I_KCOLORS"[0].rgb",//K0 = 0x0C
I_KCOLORS"[1].rgb",//K1 = 0x0D
I_KCOLORS"[2].rgb",//K2 = 0x0E
I_KCOLORS"[3].rgb",//K3 = 0x0F
I_KCOLORS"[0].rrr",//K0_R = 0x10
I_KCOLORS"[1].rrr",//K1_R = 0x11
I_KCOLORS"[2].rrr",//K2_R = 0x12
I_KCOLORS"[3].rrr",//K3_R = 0x13
I_KCOLORS"[0].ggg",//K0_G = 0x14
I_KCOLORS"[1].ggg",//K1_G = 0x15
I_KCOLORS"[2].ggg",//K2_G = 0x16
I_KCOLORS"[3].ggg",//K3_G = 0x17
I_KCOLORS"[0].bbb",//K0_B = 0x18
I_KCOLORS"[1].bbb",//K1_B = 0x19
I_KCOLORS"[2].bbb",//K2_B = 0x1A
I_KCOLORS"[3].bbb",//K3_B = 0x1B
I_KCOLORS"[0].aaa",//K0_A = 0x1C
I_KCOLORS"[1].aaa",//K1_A = 0x1D
I_KCOLORS"[2].aaa",//K2_A = 0x1E
I_KCOLORS"[3].aaa",//K3_A = 0x1F
};
static const char *tevKSelTableA[] = // KASEL
{
"1.0f", //1 = 0x00
"0.875f",//7_8 = 0x01
"0.75f", //3_4 = 0x02
"0.625f",//5_8 = 0x03
"0.5f", //1_2 = 0x04
"0.375f",//3_8 = 0x05
"0.25f", //1_4 = 0x06
"0.125f",//1_8 = 0x07
"ERROR", //0x08
"ERROR", //0x09
"ERROR", //0x0a
"ERROR", //0x0b
"ERROR", //0x0c
"ERROR", //0x0d
"ERROR", //0x0e
"ERROR", //0x0f
I_KCOLORS"[0].r",//K0_R = 0x10
I_KCOLORS"[1].r",//K1_R = 0x11
I_KCOLORS"[2].r",//K2_R = 0x12
I_KCOLORS"[3].r",//K3_R = 0x13
I_KCOLORS"[0].g",//K0_G = 0x14
I_KCOLORS"[1].g",//K1_G = 0x15
I_KCOLORS"[2].g",//K2_G = 0x16
I_KCOLORS"[3].g",//K3_G = 0x17
I_KCOLORS"[0].b",//K0_B = 0x18
I_KCOLORS"[1].b",//K1_B = 0x19
I_KCOLORS"[2].b",//K2_B = 0x1A
I_KCOLORS"[3].b",//K3_B = 0x1B
I_KCOLORS"[0].a",//K0_A = 0x1C
I_KCOLORS"[1].a",//K1_A = 0x1D
I_KCOLORS"[2].a",//K2_A = 0x1E
I_KCOLORS"[3].a",//K3_A = 0x1F
};
static const char *tevScaleTable[] = // CS
{
"1.0f", //SCALE_1
"2.0f", //SCALE_2
"4.0f", //SCALE_4
"0.5f",//DIVIDE_2
};
static const char *tevBiasTable[] = // TB
{
"", //ZERO,
"+0.5f", //ADDHALF,
"-0.5f", //SUBHALF,
"",
};
static const char *tevOpTable[] = { // TEV
"+", //TEVOP_ADD = 0,
"-", //TEVOP_SUB = 1,
};
//static const char *tevCompOpTable[] = { ">", "==" };
#define TEVCMP_R8 0
#define TEVCMP_GR16 1
#define TEVCMP_BGR24 2
#define TEVCMP_RGB8 3
static const char *tevCInputTable[] = // CC
{
"prev.rgb", //CPREV,
"prev.aaa", //APREV,
"c0.rgb", //C0,
"c0.aaa", //A0,
"c1.rgb", //C1,
"c1.aaa", //A1,
"c2.rgb", //C2,
"c2.aaa", //A2,
"textemp.rgb", //TEXC,
"textemp.aaa", //TEXA,
"rastemp.rgb", //RASC,
"rastemp.aaa", //RASA,
"float3(1.0f,1.0f,1.0f)", //ONE,
"float3(.5f,.5f,.5f)", //HALF,
"konsttemp.rgb", //KONST,
"float3(0.0f,0.0f,0.0f)", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
static const char *tevCInputTable2[] = // CC
{
"prev", //CPREV,
"(prev.aaa)", //APREV,
"c0", //C0,
"(c0.aaa)", //A0,
"c1", //C1,
"(c1.aaa)", //A1,
"c2", //C2,
"(c2.aaa)", //A2,
"textemp", //TEXC,
"(textemp.aaa)", //TEXA,
"rastemp", //RASC,
"(rastemp.aaa)", //RASA,
"float3(1.0f,1.0f,1.0f)", //ONE,
"float3(.5f,.5f,.5f)", //HALF,
"konsttemp", //"konsttemp.rgb", //KONST,
"float3(0.0f,0.0f,0.0f)", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
static const char *tevAInputTable[] = // CA
{
"prev.a", //APREV,
"c0.a", //A0,
"c1.a", //A1,
"c2.a", //A2,
"textemp.a", //TEXA,
"rastemp.a", //RASA,
"konsttemp.a", //KONST
"0.0", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR",
};
static const char *tevAInputTable2[] = // CA
{
"prev", //APREV,
"c0", //A0,
"c1", //A1,
"c2", //A2,
"textemp", //TEXA,
"rastemp", //RASA,
"konsttemp", //KONST, (hw1 had quarter)
"float4(0,0,0,0)", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
static const char *tevRasTable[] =
{
"colors[0]",
"colors[1]",
"ERROR", //2
"ERROR", //3
"ERROR", //4
"alphabump", // use bump alpha
"(alphabump*(255.0f/248.0f))", //normalized
"float4(0,0,0,0)", // zero
};
static const char *alphaRef[2] =
{
I_ALPHA"[0].x",
I_ALPHA"[0].y"
};
//static const char *tevTexFunc[] = { "tex2D", "texRECT" };
static const char *tevCOutputTable[] = { "prev.rgb", "c0.rgb", "c1.rgb", "c2.rgb" };
static const char *tevAOutputTable[] = { "prev.a", "c0.a", "c1.a", "c2.a" };
static const char* tevIndAlphaSel[] = {"", "x", "y", "z"};
static const char* tevIndAlphaScale[] = {"", "*32","*16","*8"};
static const char* tevIndBiasField[] = {"", "x", "y", "xy", "z", "xz", "yz", "xyz"}; // indexed by bias
static const char* tevIndBiasAdd[] = {"-128.0f", "1.0f", "1.0f", "1.0f" }; // indexed by fmt
static const char* tevIndWrapStart[] = {"0", "256", "128", "64", "32", "16", "0.001" };
static const char* tevIndFmtScale[] = {"255.0f", "31.0f", "15.0f", "8.0f" };
#define WRITE p+=sprintf
static const char *swapColors = "rgba";
static char swapModeTable[4][5];
static char text[16384];
static void BuildSwapModeTable()
{
//bpmem.tevregs[0].
for (int i = 0; i < 4; i++)
{
swapModeTable[i][0] = swapColors[bpmem.tevksel[i*2].swap1];
swapModeTable[i][1] = swapColors[bpmem.tevksel[i*2].swap2];
swapModeTable[i][2] = swapColors[bpmem.tevksel[i*2+1].swap1];
swapModeTable[i][3] = swapColors[bpmem.tevksel[i*2+1].swap2];
swapModeTable[i][4] = 0;
}
}
char *GeneratePixelShader(u32 texture_mask, bool has_zbuffer_target, bool bRenderZToCol0)
{
text[sizeof(text) - 1] = 0x7C; // canary
DVSTARTPROFILE();
BuildSwapModeTable();
int numStages = bpmem.genMode.numtevstages + 1;
int numTexgen = bpmem.genMode.numtexgens;
char *p = text;
WRITE(p, "//Pixel Shader for TEV stages\n");
WRITE(p, "//%i TEV stages, %i texgens, %i IND stages\n",
numStages, numTexgen, bpmem.genMode.numindstages);
bool bRenderZ = has_zbuffer_target && bpmem.zmode.updateenable;
bool bOutputZ = bpmem.ztex2.op != ZTEXTURE_DISABLE;
bool bInputZ = bpmem.ztex2.op==ZTEXTURE_ADD || bRenderZ;
// bool bRenderZToCol0 = ; // output z and alpha to color0
assert( !bRenderZToCol0 || bRenderZ );
int ztexcoord = -1;
if (bInputZ)
ztexcoord = numTexgen == 0 ? 0 : numTexgen-1;
int nIndirectStagesUsed = 0;
if (bpmem.genMode.numindstages > 0) {
for (int i = 0; i < numStages; ++i) {
if (bpmem.tevind[i].IsActive() && bpmem.tevind[i].bt < bpmem.genMode.numindstages) {
nIndirectStagesUsed |= 1<<bpmem.tevind[i].bt;
}
}
}
// Declare samplers
if (texture_mask) {
WRITE(p, "uniform samplerRECT ");
bool bfirst = true;
for (int i = 0; i < 8; ++i) {
if (texture_mask & (1<<i)) {
WRITE(p, "%s samp%d : register(s%d)", bfirst?"":",", i, i);
bfirst = false;
}
}
WRITE(p, ";\n");
}
if (texture_mask != 0xff) {
WRITE(p, "uniform sampler2D ");
bool bfirst = true;
for (int i = 0; i < 8; ++i) {
if (!(texture_mask & (1<<i))) {
WRITE(p, "%s samp%d : register(s%d)", bfirst?"":",",i, i);
bfirst = false;
}
}
WRITE(p, ";\n");
}
WRITE(p, "\n");
WRITE(p, "uniform float4 "I_COLORS"[4] : register(c%d);\n", C_COLORS);
WRITE(p, "uniform float4 "I_KCOLORS"[4] : register(c%d);\n", C_KCOLORS);
WRITE(p, "uniform float4 "I_ALPHA"[1] : register(c%d);\n", C_ALPHA);
WRITE(p, "uniform float4 "I_TEXDIMS"[8] : register(c%d);\n", C_TEXDIMS);
WRITE(p, "uniform float4 "I_ZBIAS"[2] : register(c%d);\n", C_ZBIAS);
WRITE(p, "uniform float4 "I_INDTEXSCALE"[2] : register(c%d);\n", C_INDTEXSCALE);
WRITE(p, "uniform float4 "I_INDTEXMTX"[6] : register(c%d);\n", C_INDTEXMTX);
WRITE(p, "void main(\n");
WRITE(p, "out half4 ocol0 : COLOR0,\n");
if (bRenderZ && !bRenderZToCol0 )
WRITE(p, "out half4 ocol1 : COLOR1,\n");
if (bOutputZ )
WRITE(p, " out float depth : DEPTH,\n");
// if zcoord might come from vertex shader in texcoord
if (bInputZ) {
if (numTexgen) {
for (int i = 0; i < numTexgen; ++i)
WRITE(p, " in float%d uv%d : TEXCOORD%d, \n", i==ztexcoord?4:3, i,i);
}
else
WRITE(p, " in float4 uv0 : TEXCOORD0,"); //HACK
}
else {
if (numTexgen) {
for (int i = 0; i < numTexgen; ++i)
WRITE(p, " in float3 uv%d : TEXCOORD%d,\n",i,i);
}
else
WRITE(p, " in float3 uv0 : TEXCOORD0,\n"); //HACK
}
WRITE(p, " in float4 colors[2] : COLOR0){\n");
char* pmainstart = p;
WRITE(p, "float4 c0="I_COLORS"[1],c1="I_COLORS"[2],c2="I_COLORS"[3],prev=float4(0.0f,0.0f,0.0f,0.0f),textemp,rastemp,konsttemp=float4(0.0f,0.0f,0.0f,0.0f);\n"
"float3 comp16 = float3(1,255,0), comp24 = float3(1,255,255*255);\n"
"float4 alphabump=0;\n"
"float3 tevcoord;\n"
"float2 wrappedcoord, tempcoord;\n");
//if (bOutputZ ) WRITE(p, " float depth;\n");
// WRITE(p, "return 1;}\n");
// return PixelShaderMngr::CompilePixelShader(ps, text);
// indirect texture map lookup
for(u32 i = 0; i < bpmem.genMode.numindstages; ++i) {
if (nIndirectStagesUsed & (1<<i)) {
// perform indirect texture map lookup
// note that we have to scale by the regular texture map's coordinates since this is a texRECT call
// (and we have to match with the game's texscale calls)
int texcoord = bpmem.tevindref.getTexCoord(i);
if (texture_mask & (1<<bpmem.tevindref.getTexMap(i))) {
// TODO: I removed a superfluous argument, please check that the resulting expression is correct. (mthuurne 2008-08-27)
WRITE(p, "float2 induv%d=uv%d.xy * "I_INDTEXSCALE"[%d].%s;\n", i, texcoord, i/2, (i&1)?"zw":"xy"); //, bpmem.tevindref.getTexMap(i)
char str[16];
sprintf(str, "induv%d", i);
WrapNonPow2Tex(p, str, bpmem.tevindref.getTexMap(i), texture_mask);
WRITE(p, "float3 indtex%d=texRECT(samp%d,induv%d.xy).abg;\n", i, bpmem.tevindref.getTexMap(i), i);
}
else {
WRITE(p, "float3 indtex%d=tex2D(samp%d,uv%d.xy*"I_INDTEXSCALE"[%d].%s).abg;\n", i, bpmem.tevindref.getTexMap(i), texcoord, i/2, (i&1)?"zw":"xy");
}
}
}
for (int i = 0; i < numStages; i++)
WriteStage(p, i, texture_mask); //build the equation for this stage
if (bOutputZ) {
// use the texture input of the last texture stage (textemp), hopefully this has been read and is in correct format...
if (bpmem.ztex2.op == ZTEXTURE_ADD) {
WRITE(p, "depth = frac(dot("I_ZBIAS"[0].xyzw, textemp.xyzw) + "I_ZBIAS"[1].w + uv%d.w);\n", ztexcoord);
}
else {
_assert_(bpmem.ztex2.op == ZTEXTURE_REPLACE);
WRITE(p, "depth = frac(dot("I_ZBIAS"[0].xyz, textemp.xyz) + "I_ZBIAS"[0].w);\n");
}
}
//if (bpmem.genMode.numindstages ) WRITE(p, "prev.rg = indtex0.xy;\nprev.b = 0;\n");
if (!WriteAlphaTest(p)) {
// alpha test will always fail, so restart the shader and just make it an empty function
p = pmainstart;
WRITE(p, "discard;\n");
WRITE(p, "ocol0 = 0;\n");
}
else {
if (!bRenderZToCol0) {
/* donkopunchstania: NEEDS FIX - dstalpha does not change how fragments are blended with the EFB
once the blending is done, the dstalpha is written to the EFB in place of the
fragment alpha if dstalpha is enabled. this only matters if the EFB supports alpha.
Commenting this out fixed Metroids but causes glitches in Super Mario Sunshine.
if (bpmem.dstalpha.enable)
WRITE(p, " ocol0 = float4(prev.rgb,"I_ALPHA"[0].w);\n");
else
*/
WRITE(p, " ocol0 = prev;\n");
}
}
if (bRenderZ) {
// write depth as color
if (bRenderZToCol0) {
if (bOutputZ )
WRITE(p, "ocol0.xyz = frac(float3(256.0f*256.0f, 256.0f, 1.0f) * depth);\n");
else
WRITE(p, "ocol0.xyz = frac(float3(256.0f*256.0f, 256.0f, 1.0f) * uv%d.w);\n", ztexcoord);
WRITE(p, "ocol0.w = prev.w;\n");
}
else {
if (bOutputZ)
WRITE(p, "ocol1 = frac(float4(256.0f*256.0f, 256.0f, 1.0f, 0.0f) * depth);\n");
else
WRITE(p, "ocol1 = frac(float4(256.0f*256.0f, 256.0f, 1.0f, 0.0f) * uv%d.w);\n", ztexcoord);
}
}
WRITE(p, "}\n");
if (text[sizeof(text) - 1] != 0x7C)
PanicAlert("PixelShader generator - buffer too small, canary has been eaten!");
return text;
}
static void WriteStage(char *&p, int n, u32 texture_mask)
{
char *rasswap = swapModeTable[bpmem.combiners[n].alphaC.rswap];
char *texswap = swapModeTable[bpmem.combiners[n].alphaC.tswap];
int texcoord = bpmem.tevorders[n/2].getTexCoord(n&1);
int texfun = xfregs.texcoords[texcoord].texmtxinfo.projection;
bool bHasIndStage = bpmem.tevind[n].IsActive() && bpmem.tevind[n].bt < bpmem.genMode.numindstages;
if (bHasIndStage) {
// perform the indirect op on the incoming regular coordinates using indtex%d as the offset coords
bHasIndStage = true;
int texmap = bpmem.tevorders[n/2].getEnable(n&1) ? bpmem.tevorders[n/2].getTexMap(n&1) : bpmem.tevindref.getTexMap(bpmem.tevind[n].bt);
if (bpmem.tevind[n].bs != ITBA_OFF) {
// write the bump alpha
if (bpmem.tevind[n].fmt == ITF_8) {
WRITE(p, "alphabump = indtex%d.%s %s;\n", bpmem.tevind[n].bt,
tevIndAlphaSel[bpmem.tevind[n].bs], tevIndAlphaScale[bpmem.tevind[n].fmt]);
}
else {
// donkopunchstania: really bad way to do this
// cannot always use fract because fract(1.0) is 0.0 when it needs to be 1.0
// omitting fract seems to work as well
WRITE(p, "if (indtex%d.%s >= 1.0f )\n", bpmem.tevind[n].bt,
tevIndAlphaSel[bpmem.tevind[n].bs]);
WRITE(p, " alphabump = 1.0f;\n");
WRITE(p, "else\n");
WRITE(p, " alphabump = fract ( indtex%d.%s %s );\n", bpmem.tevind[n].bt,
tevIndAlphaSel[bpmem.tevind[n].bs], tevIndAlphaScale[bpmem.tevind[n].fmt]);
}
}
// bias
WRITE(p, "float3 indtevcrd%d = indtex%d;\n", n, bpmem.tevind[n].bt);
WRITE(p, "indtevcrd%d.xy *= %s;\n", n, tevIndFmtScale[bpmem.tevind[n].fmt]);
if (bpmem.tevind[n].bias != ITB_NONE )
WRITE(p, "indtevcrd%d.%s += %s;\n", n, tevIndBiasField[bpmem.tevind[n].bias], tevIndBiasAdd[bpmem.tevind[n].fmt]);
// multiply by offset matrix and scale
if (bpmem.tevind[n].mid != 0) {
if (bpmem.tevind[n].mid <= 3) {
int mtxidx = 2*(bpmem.tevind[n].mid-1);
WRITE(p, "float2 indtevtrans%d = float2(dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d), dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d));\n",
n, mtxidx, n, mtxidx+1, n);
}
else if (bpmem.tevind[n].mid <= 5) { // s matrix
int mtxidx = 2*(bpmem.tevind[n].mid-5);
WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.xx;\n", n, mtxidx, texcoord, n);
}
else if (bpmem.tevind[n].mid <= 9) { // t matrix
int mtxidx = 2*(bpmem.tevind[n].mid-9);
WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.yy;\n", n, mtxidx, texcoord, n);
}
else {
// TODO: I removed a superfluous argument, please check that the resulting expression is correct. (mthuurne 2008-08-27)
WRITE(p, "float2 indtevtrans%d = 0;\n", n); //, n
}
}
else {
// TODO: I removed a superfluous argument, please check that the resulting expression is correct. (mthuurne 2008-08-27)
WRITE(p, "float2 indtevtrans%d = 0;\n", n); //, n
}
// wrapping
if (!bpmem.tevorders[n/2].getEnable(n&1) || (texture_mask & (1<<texmap))) {
// non pow2
if (bpmem.tevind[n].sw != ITW_OFF || bpmem.tevind[n].tw != ITW_OFF) {
if (bpmem.tevind[n].sw == ITW_0) {
if (bpmem.tevind[n].tw == ITW_0) {
// zero out completely
WRITE(p, "wrappedcoord = float2(0.0f,0.0f);\n");
}
else {
WRITE(p, "wrappedcoord.x = fmod( (uv%d.x+%s)*"I_TEXDIMS"[%d].x*"I_TEXDIMS"[%d].z, %s);\n"
"wrappedcoord.y = 0;\n", texcoord, tevIndWrapStart[bpmem.tevind[n].sw], texmap, texmap, tevIndWrapStart[bpmem.tevind[n].sw]);
}
}
else if (bpmem.tevind[n].tw == ITW_0) {
WRITE(p, "wrappedcoord.y = fmod( (uv%d.y+%s)*"I_TEXDIMS"[%d].y*"I_TEXDIMS"[%d].w, %s);\n"
"wrappedcoord.x = 0;\n", texcoord, tevIndWrapStart[bpmem.tevind[n].tw], texmap, texmap, tevIndWrapStart[bpmem.tevind[n].tw]);
}
else {
WRITE(p, "wrappedcoord = fmod( (uv%d.xy+float2(%s,%s))*"I_TEXDIMS"[%d].xy*"I_TEXDIMS"[%d].zw, float2(%s,%s));\n", texcoord,
tevIndWrapStart[bpmem.tevind[n].sw], tevIndWrapStart[bpmem.tevind[n].tw],texmap,texmap,
tevIndWrapStart[bpmem.tevind[n].sw], tevIndWrapStart[bpmem.tevind[n].tw]);
}
}
else {
WRITE(p, "wrappedcoord = uv%d.xy*"I_TEXDIMS"[%d].xy;\n", texcoord, texmap);
}
}
else {
// pow of 2
WRITE(p, "indtevtrans%d.xy *= "I_TEXDIMS"[%d].xy * "I_TEXDIMS"[%d].zw;\n", n, texmap, texmap);
// mult by bitdepth / tex dimensions
if (bpmem.tevind[n].sw != ITW_OFF || bpmem.tevind[n].tw != ITW_OFF) {
if (bpmem.tevind[n].sw == ITW_0) {
if (bpmem.tevind[n].tw == ITW_0) {
// zero out completely
WRITE(p, "wrappedcoord = float2(0.0f,0.0f);\n");
}
else {
WRITE(p, "wrappedcoord.x = "I_TEXDIMS"[%d].x * fmod( uv%d.x+%s, "I_TEXDIMS"[%d].z*%s);\n"
"wrappedcoord.y = 0;\n", texmap, texcoord, tevIndWrapStart[bpmem.tevind[n].sw], texmap, tevIndWrapStart[bpmem.tevind[n].sw]);
}
}
else if (bpmem.tevind[n].tw == ITW_0) {
WRITE(p, "wrappedcoord.y = "I_TEXDIMS"[%d].y * fmod( uv%d.y+%s, "I_TEXDIMS"[%d].w*%s);\n"
"wrappedcoord.x = 0;\n", texmap, texcoord, tevIndWrapStart[bpmem.tevind[n].tw], texmap, tevIndWrapStart[bpmem.tevind[n].tw]);
}
else {
// have to add an offset or else might get negative values!
WRITE(p, "wrappedcoord = "I_TEXDIMS"[%d].xy * fmod( uv%d.xy+float2(%s,%s), "I_TEXDIMS"[%d].zw*float2(%s,%s));\n", texmap, texcoord,
tevIndWrapStart[bpmem.tevind[n].sw], tevIndWrapStart[bpmem.tevind[n].tw], texmap,
tevIndWrapStart[bpmem.tevind[n].sw], tevIndWrapStart[bpmem.tevind[n].tw]);
}
}
else {
WRITE(p, "wrappedcoord = uv%d.xy;\n", texcoord);
}
}
if (bpmem.tevind[n].fb_addprev) {
// add previous tevcoord
if (texfun == XF_TEXPROJ_STQ) {
WRITE(p, "tevcoord.xy += wrappedcoord/uv%d.z + indtevtrans%d;\n", texcoord, n);
//WRITE(p, "tevcoord.z += uv%d.z;\n", texcoord);
}
else {
WRITE(p, "tevcoord.xy += wrappedcoord + indtevtrans%d;\n", n);
}
}
else {
WRITE(p, "tevcoord.xy = wrappedcoord/uv%d.z + indtevtrans%d;\n", texcoord, n);
//if (texfun == XF_TEXPROJ_STQ )
// WRITE(p, "tevcoord.z = uv%d.z;\n", texcoord);
}
}
WRITE(p, "rastemp=%s.%s;\n",tevRasTable[bpmem.tevorders[n/2].getColorChan(n&1)],rasswap);
if (bpmem.tevorders[n/2].getEnable(n&1)) {
int texmap = bpmem.tevorders[n/2].getTexMap(n&1);
if(!bHasIndStage) {
// calc tevcord
//tevcoord.xy = texdim[1].xy * uv1.xy / uv1.z;
int OurTexCoord = 0;
if(bpmem.genMode.numtexgens)
OurTexCoord = texcoord;
else
OurTexCoord = 0;
if (texture_mask & (1<<texmap)) {
// nonpow2
if (texfun == XF_TEXPROJ_STQ )
WRITE(p, "tevcoord.xy = uv%d.xy / uv%d.z;\n", texcoord, OurTexCoord);
else
WRITE(p, "tevcoord.xy = uv%d.xy;\n", OurTexCoord);
WrapNonPow2Tex(p, "tevcoord", texmap, texture_mask);
}
else {
if (texfun == XF_TEXPROJ_STQ )
WRITE(p, "tevcoord.xy = "I_TEXDIMS"[%d].xy * uv%d.xy / uv%d.z;\n", texmap, OurTexCoord , OurTexCoord );
else
WRITE(p, "tevcoord.xy = "I_TEXDIMS"[%d].xy * uv%d.xy;\n", texmap, OurTexCoord);
}
}
else if (texture_mask & (1<<texmap)) {
// if non pow 2, have to manually repeat
//WrapNonPow2Tex(p, "tevcoord", texmap);
bool bwraps = !!(texture_mask & (0x100<<texmap));
bool bwrapt = !!(texture_mask & (0x10000<<texmap));
if (bwraps || bwrapt) {
const char* field = bwraps ? (bwrapt ? "xy" : "x") : "y";
WRITE(p, "tevcoord.%s = fmod(tevcoord.%s+32*"I_TEXDIMS"[%d].%s,"I_TEXDIMS"[%d].%s);\n", field, field, texmap, field, texmap, field);
}
}
if (texture_mask & (1<<texmap) )
WRITE(p, "textemp=texRECT(samp%d,tevcoord.xy).%s;\n", texmap, texswap);
else
WRITE(p, "textemp=tex2D(samp%d,tevcoord.xy).%s;\n", texmap, texswap);
}
else
WRITE(p, "textemp=float4(1,1,1,1);\n");
int kc = bpmem.tevksel[n/2].getKC(n&1);
int ka = bpmem.tevksel[n/2].getKA(n&1);
TevStageCombiner::ColorCombiner &cc = bpmem.combiners[n].colorC;
TevStageCombiner::AlphaCombiner &ac = bpmem.combiners[n].alphaC;
bool bCKonst = cc.a == TEVCOLORARG_KONST || cc.b == TEVCOLORARG_KONST || cc.c == TEVCOLORARG_KONST || cc.d == TEVCOLORARG_KONST;
bool bAKonst = ac.a == TEVALPHAARG_KONST || ac.b == TEVALPHAARG_KONST || ac.c == TEVALPHAARG_KONST || ac.d == TEVALPHAARG_KONST;
if (bCKonst || bAKonst )
WRITE(p, "konsttemp=float4(%s,%s);\n",tevKSelTableC[kc],tevKSelTableA[ka]);
WRITE(p, "%s= ", tevCOutputTable[cc.dest]);
// combine the color channel
if (cc.bias != 3) { // if not compare
//normal color combiner goes here
WRITE(p, " %s*(%s%s",tevScaleTable[cc.shift],tevCInputTable[cc.d],tevOpTable[cc.op]);
WRITE(p, "lerp(%s,%s,%s) %s);\n",
tevCInputTable[cc.a], tevCInputTable[cc.b],
tevCInputTable[cc.c], tevBiasTable[cc.bias]);
}
else {
int cmp = (cc.shift<<1)|cc.op|8; // comparemode stored here
switch(cmp) {
case TEVCMP_R8_GT:
case TEVCMP_RGB8_GT: // per component compares
WRITE(p, " %s + ((%s.%s > %s.%s) ? %s : float3(0.0f,0.0f,0.0f));\n",
tevCInputTable[cc.d], tevCInputTable2[cc.a], cmp==TEVCMP_R8_GT?"r":"rgb", tevCInputTable2[cc.b], cmp==TEVCMP_R8_GT?"r":"rgb", tevCInputTable[cc.c]);
break;
case TEVCMP_R8_EQ:
case TEVCMP_RGB8_EQ:
WRITE(p, " %s + (abs(%s.r - %s.r)<%f ? %s : float3(0.0f,0.0f,0.0f));\n",
tevCInputTable[cc.d], tevCInputTable2[cc.a], tevCInputTable2[cc.b], epsilon8bit, tevCInputTable[cc.c]);
break;
case TEVCMP_GR16_GT: // 16 bit compares: 255*g+r (probably used for ztextures, so make sure in ztextures, g is the most significant byte)
case TEVCMP_BGR24_GT: // 24 bit compares: 255*255*b+255*g+r
WRITE(p, " %s + (( dot(%s.rgb-%s.rgb, comp%s) > 0) ? %s : float3(0.0f,0.0f,0.0f));\n",
tevCInputTable[cc.d], tevCInputTable2[cc.a], tevCInputTable2[cc.b], cmp==TEVCMP_GR16_GT?"16":"24", tevCInputTable[cc.c]);
break;
case TEVCMP_GR16_EQ:
case TEVCMP_BGR24_EQ:
WRITE(p, " %s + (abs(dot(%s.rgb - %s.rgb, comp%s))<%f ? %s : float3(0.0f,0.0f,0.0f));\n",
tevCInputTable[cc.d], tevCInputTable2[cc.a], tevCInputTable2[cc.b], cmp==TEVCMP_GR16_GT?"16":"24", epsilon8bit, tevCInputTable[cc.c]);
break;
default:
WRITE(p, "float3(0.0f,0.0f,0.0f);\n");
break;
}
}
if (cc.clamp)
WRITE(p, "%s = clamp(%s,0.0f,1.0f);\n", tevCOutputTable[cc.dest],tevCOutputTable[cc.dest]);
// combine the alpha channel
WRITE(p, "%s= ", tevAOutputTable[ac.dest]);
if (ac.bias != 3) { // if not compare
//normal alpha combiner goes here
WRITE(p, " %s*(%s%s",tevScaleTable[ac.shift],tevAInputTable[ac.d],tevOpTable[ac.op]);
WRITE(p, "lerp(%s,%s,%s) %s)\n",
tevAInputTable[ac.a],tevAInputTable[ac.b],
tevAInputTable[ac.c],tevBiasTable[ac.bias]);
}
else {
//compare alpha combiner goes here
int cmp = (ac.shift<<1)|ac.op|8; // comparemode stored here
switch(cmp) {
case TEVCMP_R8_GT:
case TEVCMP_A8_GT:
WRITE(p, " %s + ((%s.%s > %s.%s) ? %s : 0)\n",
tevAInputTable[ac.d],tevAInputTable2[ac.a], cmp==TEVCMP_R8_GT?"r":"a", tevAInputTable2[ac.b], cmp==TEVCMP_R8_GT?"r":"a", tevAInputTable[ac.c]);
break;
case TEVCMP_R8_EQ:
case TEVCMP_A8_EQ:
WRITE(p, " %s + (abs(%s.r - %s.r)<%f ? %s : 0)\n",
tevAInputTable[ac.d],tevAInputTable2[ac.a], tevAInputTable2[ac.b],epsilon8bit,tevAInputTable[ac.c]);
break;
case TEVCMP_GR16_GT: // 16 bit compares: 255*g+r (probably used for ztextures, so make sure in ztextures, g is the most significant byte)
case TEVCMP_BGR24_GT: // 24 bit compares: 255*255*b+255*g+r
WRITE(p, " %s + (( dot(%s.rgb-%s.rgb, comp%s) > 0) ? %s : 0)\n",
tevAInputTable[ac.d],tevAInputTable2[ac.a], tevAInputTable2[ac.b], cmp==TEVCMP_GR16_GT?"16":"24", tevAInputTable[ac.c]);
break;
case TEVCMP_GR16_EQ:
case TEVCMP_BGR24_EQ:
WRITE(p, " %s + (abs(dot(%s.rgb - %s.rgb, comp%s))<%f ? %s : 0)\n",
tevAInputTable[ac.d],tevAInputTable2[ac.a], tevAInputTable2[ac.b],cmp==TEVCMP_GR16_GT?"16":"24",epsilon8bit,tevAInputTable[ac.c]);
break;
default:
WRITE(p, "0)\n");
break;
}
}
WRITE(p, ";\n");
if (ac.clamp)
WRITE(p, "%s = clamp(%s,0.0f,1.0f);\n", tevAOutputTable[ac.dest],tevAOutputTable[ac.dest]);
WRITE(p, "\n");
}
void WrapNonPow2Tex(char* &p, const char* var, int texmap, u32 texture_mask)
{
_assert_(texture_mask & (1<<texmap));
bool bwraps = !!(texture_mask & (0x100<<texmap));
bool bwrapt = !!(texture_mask & (0x10000<<texmap));
if (bwraps || bwrapt) {
const char* field = bwraps ? (bwrapt ? "xy" : "x") : "y";
const char* wrapfield = bwraps ? (bwrapt ? "zw" : "z") : "w";
WRITE(p, "%s.%s = "I_TEXDIMS"[%d].%s*frac(%s.%s*"I_TEXDIMS"[%d].%s+32);\n", var, field, texmap, field, var, field, texmap, wrapfield);
if (!bwraps )
WRITE(p, "%s.x *= "I_TEXDIMS"[%d].x * "I_TEXDIMS"[%d].z;\n", var, texmap, texmap);
if (!bwrapt )
WRITE(p, "%s.y *= "I_TEXDIMS"[%d].y * "I_TEXDIMS"[%d].w;\n", var, texmap, texmap);
}
else {
WRITE(p, "%s.xy *= "I_TEXDIMS"[%d].xy * "I_TEXDIMS"[%d].zw;\n", var, texmap, texmap);
}
}
static void WriteAlphaCompare(char *&p, int num, int comp)
{
switch(comp) {
case ALPHACMP_ALWAYS: WRITE(p, "(false)"); break;
case ALPHACMP_NEVER: WRITE(p, "(true)"); break;
case ALPHACMP_LEQUAL: WRITE(p, "(prev.a > %s)",alphaRef[num]); break;
case ALPHACMP_LESS: WRITE(p, "(prev.a >= %s - %f)",alphaRef[num],epsilon8bit*0.5f);break;
case ALPHACMP_GEQUAL: WRITE(p, "(prev.a < %s)",alphaRef[num]); break;
case ALPHACMP_GREATER: WRITE(p, "(prev.a <= %s + %f)",alphaRef[num],epsilon8bit*0.5f);break;
case ALPHACMP_EQUAL: WRITE(p, "(abs(prev.a-%s)>%f)",alphaRef[num],epsilon8bit*2); break;
case ALPHACMP_NEQUAL: WRITE(p, "(abs(prev.a-%s)<%f)",alphaRef[num],epsilon8bit*2); break;
}
}
static bool WriteAlphaTest(char *&p)
{
u32 op = bpmem.alphaFunc.logic;
u32 comp[2] = {bpmem.alphaFunc.comp0,bpmem.alphaFunc.comp1};
//first kill all the simple cases
switch(op) {
case 0: // and
if (comp[0] == ALPHACMP_ALWAYS && comp[1] == ALPHACMP_ALWAYS) return true;
if (comp[0] == ALPHACMP_NEVER || comp[1] == ALPHACMP_NEVER) {
WRITE(p, "discard;\n");
return false;
}
break;
case 1: // or
if (comp[0] == ALPHACMP_ALWAYS || comp[1] == ALPHACMP_ALWAYS) return true;
if (comp[0] == ALPHACMP_NEVER && comp[1] == ALPHACMP_NEVER) {
WRITE(p, "discard;\n");
return false;
}
break;
case 2: // xor
if ( (comp[0] == ALPHACMP_ALWAYS && comp[1] == ALPHACMP_NEVER) || (comp[0] == ALPHACMP_NEVER && comp[1] == ALPHACMP_ALWAYS) ) return true;
if ( (comp[0] == ALPHACMP_ALWAYS && comp[1] == ALPHACMP_ALWAYS) || (comp[0] == ALPHACMP_NEVER && comp[1] == ALPHACMP_NEVER)) {
WRITE(p, "discard;\n");
return false;
}
break;
case 3: // xnor
if ( (comp[0] == ALPHACMP_ALWAYS && comp[1] == ALPHACMP_NEVER) || (comp[0] == ALPHACMP_NEVER && comp[1] == ALPHACMP_ALWAYS)) {
WRITE(p, "discard;\n");
return false;
}
if ( (comp[0] == ALPHACMP_ALWAYS && comp[1] == ALPHACMP_ALWAYS) || (comp[0] == ALPHACMP_NEVER && comp[1] == ALPHACMP_NEVER) )
return true;
break;
}
WRITE(p, "discard( ");
WriteAlphaCompare(p, 0, bpmem.alphaFunc.comp0);
// negated because testing the inverse condition
switch(bpmem.alphaFunc.logic) {
case 0: WRITE(p, " || "); break; // and
case 1: WRITE(p, " && "); break; // or
case 2: WRITE(p, " == "); break; // xor
case 3: WRITE(p, " != "); break; // xnor
}
WriteAlphaCompare(p, 1, bpmem.alphaFunc.comp1);
WRITE(p, ");\n");
return true;
}