double ApplyLUT2LUMA(FIBITMAP *src, FIBITMAP *dst, char * LUT, int threadcount)
{
_mark();
unsigned spitch = FreeImage_GetPitch(src);
unsigned dpitch = FreeImage_GetPitch(dst);
unsigned w = FreeImage_GetWidth(src);
unsigned h = FreeImage_GetHeight(src);
BYTE * srcbits = FreeImage_GetBits(src);
BYTE * dstbits = FreeImage_GetBits(dst);
BYTE LUT8[256];
WORD LUT16[65536];
int bpp = FreeImage_GetBPP(src);
if (bpp == 24) {
for (int x=0; x<256; x++) {
LUT8[x] = ((BYTE *) LUT)[x];
}
#pragma omp parallel for num_threads(threadcount)
for(unsigned y = 0; y < h; y++) {
for(unsigned x = 0; x < w; x++) {
BYTE * bdstpix = (BYTE *) (dstbits + dpitch*y + 3*x);
BYTE * pixel = (BYTE *) (srcbits + spitch*y + 3*x);
int lumaorig = (int) LUMA(pixel[FI_RGBA_RED],pixel[FI_RGBA_GREEN],pixel[FI_RGBA_BLUE]);
int diff = lumaorig - LUT8[lumaorig];
bdstpix[FI_RGBA_RED] = std::min(std::max(bdstpix[FI_RGBA_RED] + diff,0),255);
bdstpix[FI_RGBA_GREEN] = std::min(std::max(bdstpix[FI_RGBA_GREEN] + diff,0),255);
bdstpix[FI_RGBA_BLUE] = std::min(std::max(bdstpix[FI_RGBA_BLUE] + diff,0),255);
}
}
}
if (bpp == 48) {
for (int x=0; x<65536; x++) {
LUT16[x] = ((WORD *)LUT)[x];
}
#pragma omp parallel for num_threads(threadcount)
for(unsigned y = 0; y < h-1; y++) {
for(unsigned x = 0; x < w-1; x++) {
FIRGB16 * wdstpix = (FIRGB16 *) (dstbits + dpitch*y + 6*x);
FIRGB16 * pixel = (FIRGB16 *) (srcbits + spitch*y + 6*x);
int lumaorig = (int) LUMA(pixel->red,pixel->green,pixel->blue);
int diff = lumaorig - LUT16[lumaorig];
wdstpix->red = std::min(std::max(pixel->red + diff,0),65535);
wdstpix->green = std::min(std::max(pixel->green + diff,0),65535);
wdstpix->blue = std::min(std::max(pixel->blue + diff,0),65535);
}
}
}
return _duration();
}
static void
set_rgb16_lum_from_rgb16(const uint16_t topr,
const uint16_t topg,
const uint16_t topb,
uint16_t *botr,
uint16_t *botg,
uint16_t *botb)
{
// Spec: SetLum()
// Colours potentially can go out of band to both sides, hence the
// temporary representation inflation.
const uint16_t botlum = LUMA(*botr, *botg, *botb) / (1<<15);
const uint16_t toplum = LUMA(topr, topg, topb) / (1<<15);
const int16_t diff = botlum - toplum;
int32_t r = topr + diff;
int32_t g = topg + diff;
int32_t b = topb + diff;
// Spec: ClipColor()
// Clip out of band values
int32_t lum = LUMA(r, g, b) / (1<<15);
int32_t cmin = MIN3(r, g, b);
int32_t cmax = MAX3(r, g, b);
if (cmin < 0) {
r = lum + (((r - lum) * lum) / (lum - cmin));
g = lum + (((g - lum) * lum) / (lum - cmin));
b = lum + (((b - lum) * lum) / (lum - cmin));
}
if (cmax > (1<<15)) {
r = lum + (((r - lum) * ((1<<15)-lum)) / (cmax - lum));
g = lum + (((g - lum) * ((1<<15)-lum)) / (cmax - lum));
b = lum + (((b - lum) * ((1<<15)-lum)) / (cmax - lum));
}
#ifdef HEAVY_DEBUG
assert((0 <= r) && (r <= (1<<15)));
assert((0 <= g) && (g <= (1<<15)));
assert((0 <= b) && (b <= (1<<15)));
#endif
*botr = r;
*botg = g;
*botb = b;
}
/*
===============
ComputeColors
===============
*/
static void ComputeColors( shaderStage_t *pStage )
{
int i;
//
// rgbGen
//
switch ( pStage->rgbGen )
{
case CGEN_IDENTITY:
Com_Memset( tess.svars.colors, 0xff, tess.numVertexes * 4 );
break;
default:
case CGEN_IDENTITY_LIGHTING:
Com_Memset( tess.svars.colors, tr.identityLightByte, tess.numVertexes * 4 );
break;
case CGEN_LIGHTING_DIFFUSE:
RB_CalcDiffuseColor( ( unsigned char * ) tess.svars.colors );
break;
case CGEN_EXACT_VERTEX:
Com_Memcpy( tess.svars.colors, tess.vertexColors, tess.numVertexes * sizeof( tess.vertexColors[0] ) );
break;
case CGEN_CONST:
for ( i = 0; i < tess.numVertexes; i++ ) {
*(int *)tess.svars.colors[i] = *(int *)pStage->constantColor;
}
break;
case CGEN_VERTEX:
if ( tr.identityLight == 1 )
{
Com_Memcpy( tess.svars.colors, tess.vertexColors, tess.numVertexes * sizeof( tess.vertexColors[0] ) );
}
else
{
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][0] = tess.vertexColors[i][0] * tr.identityLight;
tess.svars.colors[i][1] = tess.vertexColors[i][1] * tr.identityLight;
tess.svars.colors[i][2] = tess.vertexColors[i][2] * tr.identityLight;
tess.svars.colors[i][3] = tess.vertexColors[i][3];
}
}
break;
case CGEN_ONE_MINUS_VERTEX:
if ( tr.identityLight == 1 )
{
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][0] = 255 - tess.vertexColors[i][0];
tess.svars.colors[i][1] = 255 - tess.vertexColors[i][1];
tess.svars.colors[i][2] = 255 - tess.vertexColors[i][2];
}
}
else
{
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][0] = ( 255 - tess.vertexColors[i][0] ) * tr.identityLight;
tess.svars.colors[i][1] = ( 255 - tess.vertexColors[i][1] ) * tr.identityLight;
tess.svars.colors[i][2] = ( 255 - tess.vertexColors[i][2] ) * tr.identityLight;
}
}
break;
case CGEN_FOG:
{
fog_t *fog;
fog = tr.world->fogs + tess.fogNum;
for ( i = 0; i < tess.numVertexes; i++ ) {
* ( int * )&tess.svars.colors[i] = fog->colorInt;
}
}
break;
case CGEN_WAVEFORM:
RB_CalcWaveColor( &pStage->rgbWave, ( unsigned char * ) tess.svars.colors );
break;
case CGEN_ENTITY:
RB_CalcColorFromEntity( ( unsigned char * ) tess.svars.colors );
break;
case CGEN_ONE_MINUS_ENTITY:
RB_CalcColorFromOneMinusEntity( ( unsigned char * ) tess.svars.colors );
break;
}
//
// alphaGen
//
switch ( pStage->alphaGen )
{
case AGEN_SKIP:
break;
case AGEN_IDENTITY:
if ( pStage->rgbGen != CGEN_IDENTITY ) {
if ( ( pStage->rgbGen == CGEN_VERTEX && tr.identityLight != 1 ) ||
pStage->rgbGen != CGEN_VERTEX ) {
for ( i = 0; i < tess.numVertexes; i++ ) {
tess.svars.colors[i][3] = 0xff;
}
}
}
break;
case AGEN_CONST:
if ( pStage->rgbGen != CGEN_CONST ) {
for ( i = 0; i < tess.numVertexes; i++ ) {
tess.svars.colors[i][3] = pStage->constantColor[3];
}
}
break;
case AGEN_WAVEFORM:
RB_CalcWaveAlpha( &pStage->alphaWave, ( unsigned char * ) tess.svars.colors );
break;
case AGEN_LIGHTING_SPECULAR:
RB_CalcSpecularAlpha( ( unsigned char * ) tess.svars.colors );
break;
case AGEN_ENTITY:
RB_CalcAlphaFromEntity( ( unsigned char * ) tess.svars.colors );
break;
case AGEN_ONE_MINUS_ENTITY:
RB_CalcAlphaFromOneMinusEntity( ( unsigned char * ) tess.svars.colors );
break;
case AGEN_VERTEX:
if ( pStage->rgbGen != CGEN_VERTEX ) {
for ( i = 0; i < tess.numVertexes; i++ ) {
tess.svars.colors[i][3] = tess.vertexColors[i][3];
}
}
break;
case AGEN_ONE_MINUS_VERTEX:
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][3] = 255 - tess.vertexColors[i][3];
}
break;
case AGEN_PORTAL:
{
unsigned char alpha;
for ( i = 0; i < tess.numVertexes; i++ )
{
float len;
vec3_t v;
VectorSubtract( tess.xyz[i], backEnd.viewParms.or.origin, v );
len = VectorLength( v );
len /= tess.shader->portalRange;
if ( len < 0 )
{
alpha = 0;
}
else if ( len > 1 )
{
alpha = 0xff;
}
else
{
alpha = len * 0xff;
}
tess.svars.colors[i][3] = alpha;
}
}
break;
}
//
// fog adjustment for colors to fade out as fog increases
//
if ( tess.fogNum )
{
switch ( pStage->adjustColorsForFog )
{
case ACFF_MODULATE_RGB:
RB_CalcModulateColorsByFog( ( unsigned char * ) tess.svars.colors );
break;
case ACFF_MODULATE_ALPHA:
RB_CalcModulateAlphasByFog( ( unsigned char * ) tess.svars.colors );
break;
case ACFF_MODULATE_RGBA:
RB_CalcModulateRGBAsByFog( ( unsigned char * ) tess.svars.colors );
break;
case ACFF_NONE:
break;
}
}
// if in greyscale rendering mode turn all color values into greyscale.
if(r_greyscale->integer)
{
int scale;
for(i = 0; i < tess.numVertexes; i++)
{
scale = LUMA(tess.svars.colors[i][0], tess.svars.colors[i][1], tess.svars.colors[i][2]);
tess.svars.colors[i][0] = tess.svars.colors[i][1] = tess.svars.colors[i][2] = scale;
}
}
else if(r_greyscale->value)
{
float scale;
for(i = 0; i < tess.numVertexes; i++)
{
scale = LUMA(tess.svars.colors[i][0], tess.svars.colors[i][1], tess.svars.colors[i][2]);
tess.svars.colors[i][0] = LERP(tess.svars.colors[i][0], scale, r_greyscale->value);
tess.svars.colors[i][1] = LERP(tess.svars.colors[i][1], scale, r_greyscale->value);
tess.svars.colors[i][2] = LERP(tess.svars.colors[i][2], scale, r_greyscale->value);
}
}
}
static void ProjectDlightTexture_scalar( void ) {
int i, l;
vec3_t origin;
float *texCoords;
byte *colors;
byte clipBits[SHADER_MAX_VERTEXES];
float texCoordsArray[SHADER_MAX_VERTEXES][2];
byte colorArray[SHADER_MAX_VERTEXES][4];
glIndex_t hitIndexes[SHADER_MAX_INDEXES];
int numIndexes;
float scale;
float radius;
vec3_t floatColor;
float modulate = 0.0f;
if ( !backEnd.refdef.num_dlights ) {
return;
}
for ( l = 0 ; l < backEnd.refdef.num_dlights ; l++ ) {
dlight_t *dl;
if ( !( tess.dlightBits & ( 1 << l ) ) ) {
continue; // this surface definately doesn't have any of this light
}
texCoords = texCoordsArray[0];
colors = colorArray[0];
dl = &backEnd.refdef.dlights[l];
VectorCopy( dl->transformed, origin );
radius = dl->radius;
scale = 1.0f / radius;
if(r_greyscale->integer)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = floatColor[1] = floatColor[2] = luminance;
}
else if(r_greyscale->value)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = LERP(dl->color[0] * 255.0f, luminance, r_greyscale->value);
floatColor[1] = LERP(dl->color[1] * 255.0f, luminance, r_greyscale->value);
floatColor[2] = LERP(dl->color[2] * 255.0f, luminance, r_greyscale->value);
}
else
{
floatColor[0] = dl->color[0] * 255.0f;
floatColor[1] = dl->color[1] * 255.0f;
floatColor[2] = dl->color[2] * 255.0f;
}
for ( i = 0 ; i < tess.numVertexes ; i++, texCoords += 2, colors += 4 ) {
int clip = 0;
vec3_t dist;
VectorSubtract( origin, tess.xyz[i], dist );
backEnd.pc.c_dlightVertexes++;
texCoords[0] = 0.5f + dist[0] * scale;
texCoords[1] = 0.5f + dist[1] * scale;
if( !r_dlightBacks->integer &&
// dist . tess.normal[i]
( dist[0] * tess.normal[i][0] +
dist[1] * tess.normal[i][1] +
dist[2] * tess.normal[i][2] ) < 0.0f ) {
clip = 63;
} else {
if ( texCoords[0] < 0.0f ) {
clip |= 1;
} else if ( texCoords[0] > 1.0f ) {
clip |= 2;
}
if ( texCoords[1] < 0.0f ) {
clip |= 4;
} else if ( texCoords[1] > 1.0f ) {
clip |= 8;
}
texCoords[0] = texCoords[0];
texCoords[1] = texCoords[1];
// modulate the strength based on the height and color
if ( dist[2] > radius ) {
clip |= 16;
modulate = 0.0f;
} else if ( dist[2] < -radius ) {
clip |= 32;
modulate = 0.0f;
} else {
dist[2] = Q_fabs(dist[2]);
if ( dist[2] < radius * 0.5f ) {
modulate = 1.0f;
} else {
modulate = 2.0f * (radius - dist[2]) * scale;
}
}
}
clipBits[i] = clip;
colors[0] = ri.ftol(floatColor[0] * modulate);
colors[1] = ri.ftol(floatColor[1] * modulate);
colors[2] = ri.ftol(floatColor[2] * modulate);
colors[3] = 255;
}
// build a list of triangles that need light
numIndexes = 0;
for ( i = 0 ; i < tess.numIndexes ; i += 3 ) {
int a, b, c;
a = tess.indexes[i];
b = tess.indexes[i+1];
c = tess.indexes[i+2];
if ( clipBits[a] & clipBits[b] & clipBits[c] ) {
continue; // not lighted
}
hitIndexes[numIndexes] = a;
hitIndexes[numIndexes+1] = b;
hitIndexes[numIndexes+2] = c;
numIndexes += 3;
}
if ( !numIndexes ) {
continue;
}
qglEnableClientState( GL_TEXTURE_COORD_ARRAY );
qglTexCoordPointer( 2, GL_FLOAT, 0, texCoordsArray[0] );
qglEnableClientState( GL_COLOR_ARRAY );
qglColorPointer( 4, GL_UNSIGNED_BYTE, 0, colorArray );
GL_Bind( tr.dlightImage );
// include GLS_DEPTHFUNC_EQUAL so alpha tested surfaces don't add light
// where they aren't rendered
if ( dl->additive ) {
GL_State( GLS_SRCBLEND_ONE | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
else {
GL_State( GLS_SRCBLEND_DST_COLOR | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
R_DrawElements( numIndexes, hitIndexes );
backEnd.pc.c_totalIndexes += numIndexes;
backEnd.pc.c_dlightIndexes += numIndexes;
}
}
static void ProjectDlightTexture_altivec( void ) {
int i, l;
vec_t origin0, origin1, origin2;
float texCoords0, texCoords1;
vector float floatColorVec0, floatColorVec1;
vector float modulateVec, colorVec, zero;
vector short colorShort;
vector signed int colorInt;
vector unsigned char floatColorVecPerm, modulatePerm, colorChar;
vector unsigned char vSel = VECCONST_UINT8(0x00, 0x00, 0x00, 0xff,
0x00, 0x00, 0x00, 0xff,
0x00, 0x00, 0x00, 0xff,
0x00, 0x00, 0x00, 0xff);
float *texCoords;
byte *colors;
byte clipBits[SHADER_MAX_VERTEXES];
float texCoordsArray[SHADER_MAX_VERTEXES][2];
byte colorArray[SHADER_MAX_VERTEXES][4];
unsigned hitIndexes[SHADER_MAX_INDEXES];
int numIndexes;
float scale;
float radius;
vec3_t floatColor;
float modulate = 0.0f;
if ( !backEnd.refdef.num_dlights ) {
return;
}
// There has to be a better way to do this so that floatColor
// and/or modulate are already 16-byte aligned.
floatColorVecPerm = vec_lvsl(0,(float *)floatColor);
modulatePerm = vec_lvsl(0,(float *)&modulate);
modulatePerm = (vector unsigned char)vec_splat((vector unsigned int)modulatePerm,0);
zero = (vector float)vec_splat_s8(0);
for ( l = 0 ; l < backEnd.refdef.num_dlights ; l++ ) {
dlight_t *dl;
if ( !( tess.dlightBits & ( 1 << l ) ) ) {
continue; // this surface definately doesn't have any of this light
}
texCoords = texCoordsArray[0];
colors = colorArray[0];
dl = &backEnd.refdef.dlights[l];
origin0 = dl->transformed[0];
origin1 = dl->transformed[1];
origin2 = dl->transformed[2];
radius = dl->radius;
scale = 1.0f / radius;
if(r_greyscale->integer)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = floatColor[1] = floatColor[2] = luminance;
}
else if(r_greyscale->value)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = LERP(dl->color[0] * 255.0f, luminance, r_greyscale->value);
floatColor[1] = LERP(dl->color[1] * 255.0f, luminance, r_greyscale->value);
floatColor[2] = LERP(dl->color[2] * 255.0f, luminance, r_greyscale->value);
}
else
{
floatColor[0] = dl->color[0] * 255.0f;
floatColor[1] = dl->color[1] * 255.0f;
floatColor[2] = dl->color[2] * 255.0f;
}
floatColorVec0 = vec_ld(0, floatColor);
floatColorVec1 = vec_ld(11, floatColor);
floatColorVec0 = vec_perm(floatColorVec0,floatColorVec0,floatColorVecPerm);
for ( i = 0 ; i < tess.numVertexes ; i++, texCoords += 2, colors += 4 ) {
int clip = 0;
vec_t dist0, dist1, dist2;
dist0 = origin0 - tess.xyz[i][0];
dist1 = origin1 - tess.xyz[i][1];
dist2 = origin2 - tess.xyz[i][2];
backEnd.pc.c_dlightVertexes++;
texCoords0 = 0.5f + dist0 * scale;
texCoords1 = 0.5f + dist1 * scale;
if( !r_dlightBacks->integer &&
// dist . tess.normal[i]
( dist0 * tess.normal[i][0] +
dist1 * tess.normal[i][1] +
dist2 * tess.normal[i][2] ) < 0.0f ) {
clip = 63;
} else {
if ( texCoords0 < 0.0f ) {
clip |= 1;
} else if ( texCoords0 > 1.0f ) {
clip |= 2;
}
if ( texCoords1 < 0.0f ) {
clip |= 4;
} else if ( texCoords1 > 1.0f ) {
clip |= 8;
}
texCoords[0] = texCoords0;
texCoords[1] = texCoords1;
// modulate the strength based on the height and color
if ( dist2 > radius ) {
clip |= 16;
modulate = 0.0f;
} else if ( dist2 < -radius ) {
clip |= 32;
modulate = 0.0f;
} else {
dist2 = Q_fabs(dist2);
if ( dist2 < radius * 0.5f ) {
modulate = 1.0f;
} else {
modulate = 2.0f * (radius - dist2) * scale;
}
}
}
clipBits[i] = clip;
modulateVec = vec_ld(0,(float *)&modulate);
modulateVec = vec_perm(modulateVec,modulateVec,modulatePerm);
colorVec = vec_madd(floatColorVec0,modulateVec,zero);
colorInt = vec_cts(colorVec,0); // RGBx
colorShort = vec_pack(colorInt,colorInt); // RGBxRGBx
colorChar = vec_packsu(colorShort,colorShort); // RGBxRGBxRGBxRGBx
colorChar = vec_sel(colorChar,vSel,vSel); // RGBARGBARGBARGBA replace alpha with 255
vec_ste((vector unsigned int)colorChar,0,(unsigned int *)colors); // store color
}
// build a list of triangles that need light
numIndexes = 0;
for ( i = 0 ; i < tess.numIndexes ; i += 3 ) {
int a, b, c;
a = tess.indexes[i];
b = tess.indexes[i+1];
c = tess.indexes[i+2];
if ( clipBits[a] & clipBits[b] & clipBits[c] ) {
continue; // not lighted
}
hitIndexes[numIndexes] = a;
hitIndexes[numIndexes+1] = b;
hitIndexes[numIndexes+2] = c;
numIndexes += 3;
}
if ( !numIndexes ) {
continue;
}
qglEnableClientState( GL_TEXTURE_COORD_ARRAY );
qglTexCoordPointer( 2, GL_FLOAT, 0, texCoordsArray[0] );
qglEnableClientState( GL_COLOR_ARRAY );
qglColorPointer( 4, GL_UNSIGNED_BYTE, 0, colorArray );
GL_Bind( tr.dlightImage );
// include GLS_DEPTHFUNC_EQUAL so alpha tested surfaces don't add light
// where they aren't rendered
if ( dl->additive ) {
GL_State( GLS_SRCBLEND_ONE | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
else {
GL_State( GLS_SRCBLEND_DST_COLOR | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
R_DrawElements( numIndexes, hitIndexes );
backEnd.pc.c_totalIndexes += numIndexes;
backEnd.pc.c_dlightIndexes += numIndexes;
}
}
/*
===============
ComputeColors
===============
*/
static void ComputeColors( shaderStage_t *pStage ) {
int i;
//
// rgbGen
//
switch ( pStage->rgbGen )
{
case CGEN_IDENTITY:
memset( tess.svars.colors, 0xff, tess.numVertexes * 4 );
break;
default:
case CGEN_IDENTITY_LIGHTING:
memset( tess.svars.colors, tr.identityLightByte, tess.numVertexes * 4 );
break;
case CGEN_LIGHTING_DIFFUSE:
RB_CalcDiffuseColor( ( unsigned char * ) tess.svars.colors );
break;
case CGEN_EXACT_VERTEX:
memcpy( tess.svars.colors, tess.vertexColors, tess.numVertexes * sizeof( tess.vertexColors[0] ) );
break;
case CGEN_CONST:
for ( i = 0; i < tess.numVertexes; i++ ) {
*(int *)tess.svars.colors[i] = *(int *)pStage->constantColor;
}
break;
case CGEN_VERTEX:
if ( tr.identityLight == 1 ) {
memcpy( tess.svars.colors, tess.vertexColors, tess.numVertexes * sizeof( tess.vertexColors[0] ) );
} else
{
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][0] = tess.vertexColors[i][0] * tr.identityLight;
tess.svars.colors[i][1] = tess.vertexColors[i][1] * tr.identityLight;
tess.svars.colors[i][2] = tess.vertexColors[i][2] * tr.identityLight;
tess.svars.colors[i][3] = tess.vertexColors[i][3];
}
}
break;
case CGEN_ONE_MINUS_VERTEX:
if ( tr.identityLight == 1 ) {
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][0] = 255 - tess.vertexColors[i][0];
tess.svars.colors[i][1] = 255 - tess.vertexColors[i][1];
tess.svars.colors[i][2] = 255 - tess.vertexColors[i][2];
}
} else
{
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][0] = ( 255 - tess.vertexColors[i][0] ) * tr.identityLight;
tess.svars.colors[i][1] = ( 255 - tess.vertexColors[i][1] ) * tr.identityLight;
tess.svars.colors[i][2] = ( 255 - tess.vertexColors[i][2] ) * tr.identityLight;
}
}
break;
case CGEN_FOG:
{
fog_t *fog;
fog = tr.world->fogs + tess.fogNum;
for ( i = 0; i < tess.numVertexes; i++ ) {
*( int * )&tess.svars.colors[i] = fog->colorInt;
}
}
break;
case CGEN_WAVEFORM:
RB_CalcWaveColor( &pStage->rgbWave, ( unsigned char * ) tess.svars.colors );
break;
case CGEN_ENTITY:
RB_CalcColorFromEntity( ( unsigned char * ) tess.svars.colors );
break;
case CGEN_ONE_MINUS_ENTITY:
RB_CalcColorFromOneMinusEntity( ( unsigned char * ) tess.svars.colors );
break;
}
//
// alphaGen
//
switch ( pStage->alphaGen )
{
case AGEN_SKIP:
break;
case AGEN_IDENTITY:
if ( pStage->rgbGen != CGEN_IDENTITY ) {
if ( ( pStage->rgbGen == CGEN_VERTEX && tr.identityLight != 1 ) ||
pStage->rgbGen != CGEN_VERTEX ) {
for ( i = 0; i < tess.numVertexes; i++ ) {
tess.svars.colors[i][3] = 0xff;
}
}
}
break;
case AGEN_CONST:
if ( pStage->rgbGen != CGEN_CONST ) {
for ( i = 0; i < tess.numVertexes; i++ ) {
tess.svars.colors[i][3] = pStage->constantColor[3];
}
}
break;
case AGEN_WAVEFORM:
RB_CalcWaveAlpha( &pStage->alphaWave, ( unsigned char * ) tess.svars.colors );
break;
case AGEN_LIGHTING_SPECULAR:
RB_CalcSpecularAlpha( ( unsigned char * ) tess.svars.colors );
break;
case AGEN_ENTITY:
RB_CalcAlphaFromEntity( ( unsigned char * ) tess.svars.colors );
break;
case AGEN_ONE_MINUS_ENTITY:
RB_CalcAlphaFromOneMinusEntity( ( unsigned char * ) tess.svars.colors );
break;
// Ridah
case AGEN_NORMALZFADE:
{
float alpha, range, lowest, highest, dot;
vec3_t worldUp;
qboolean zombieEffect = qfalse;
if ( VectorCompare( backEnd.currentEntity->e.fireRiseDir, vec3_origin ) ) {
VectorSet( backEnd.currentEntity->e.fireRiseDir, 0, 0, 1 );
}
if ( backEnd.currentEntity->e.hModel ) { // world surfaces dont have an axis
VectorRotate( backEnd.currentEntity->e.fireRiseDir, backEnd.currentEntity->e.axis, worldUp );
} else {
VectorCopy( backEnd.currentEntity->e.fireRiseDir, worldUp );
}
lowest = pStage->zFadeBounds[0];
if ( lowest == -1000 ) { // use entity alpha
lowest = backEnd.currentEntity->e.shaderTime;
zombieEffect = qtrue;
}
highest = pStage->zFadeBounds[1];
if ( highest == -1000 ) { // use entity alpha
highest = backEnd.currentEntity->e.shaderTime;
zombieEffect = qtrue;
}
range = highest - lowest;
for ( i = 0; i < tess.numVertexes; i++ ) {
dot = DotProduct( tess.normal[i], worldUp );
// special handling for Zombie fade effect
if ( zombieEffect ) {
alpha = (float)backEnd.currentEntity->e.shaderRGBA[3] * ( dot + 1.0 ) / 2.0;
alpha += ( 2.0 * (float)backEnd.currentEntity->e.shaderRGBA[3] ) * ( 1.0 - ( dot + 1.0 ) / 2.0 );
if ( alpha > 255.0 ) {
alpha = 255.0;
} else if ( alpha < 0.0 ) {
alpha = 0.0;
}
tess.svars.colors[i][3] = (byte)( alpha );
continue;
}
if ( dot < highest ) {
if ( dot > lowest ) {
if ( dot < lowest + range / 2 ) {
alpha = ( (float)pStage->constantColor[3] * ( ( dot - lowest ) / ( range / 2 ) ) );
} else {
alpha = ( (float)pStage->constantColor[3] * ( 1.0 - ( ( dot - lowest - range / 2 ) / ( range / 2 ) ) ) );
}
if ( alpha > 255.0 ) {
alpha = 255.0;
} else if ( alpha < 0.0 ) {
alpha = 0.0;
}
// finally, scale according to the entity's alpha
if ( backEnd.currentEntity->e.hModel ) {
alpha *= (float)backEnd.currentEntity->e.shaderRGBA[3] / 255.0;
}
tess.svars.colors[i][3] = (byte)( alpha );
} else {
tess.svars.colors[i][3] = 0;
}
} else {
tess.svars.colors[i][3] = 0;
}
}
}
break;
// done.
case AGEN_VERTEX:
if ( pStage->rgbGen != CGEN_VERTEX ) {
for ( i = 0; i < tess.numVertexes; i++ ) {
tess.svars.colors[i][3] = tess.vertexColors[i][3];
}
}
break;
case AGEN_ONE_MINUS_VERTEX:
for ( i = 0; i < tess.numVertexes; i++ )
{
tess.svars.colors[i][3] = 255 - tess.vertexColors[i][3];
}
break;
case AGEN_PORTAL:
{
unsigned char alpha;
for ( i = 0; i < tess.numVertexes; i++ )
{
float len;
vec3_t v;
VectorSubtract( tess.xyz[i], backEnd.viewParms.or.origin, v );
len = VectorLength( v );
len /= tess.shader->portalRange;
if ( len < 0 ) {
alpha = 0;
} else if ( len > 1 ) {
alpha = 0xff;
} else
{
alpha = len * 0xff;
}
tess.svars.colors[i][3] = alpha;
}
}
break;
}
//
// fog adjustment for colors to fade out as fog increases
//
if ( tess.fogNum ) {
switch ( pStage->adjustColorsForFog )
{
case ACFF_MODULATE_RGB:
RB_CalcModulateColorsByFog( ( unsigned char * ) tess.svars.colors );
break;
case ACFF_MODULATE_ALPHA:
RB_CalcModulateAlphasByFog( ( unsigned char * ) tess.svars.colors );
break;
case ACFF_MODULATE_RGBA:
RB_CalcModulateRGBAsByFog( ( unsigned char * ) tess.svars.colors );
break;
case ACFF_NONE:
break;
}
}
// if in greyscale rendering mode turn all color values into greyscale.
if(r_greyscale->integer)
{
int scale;
for(i = 0; i < tess.numVertexes; i++)
{
scale = LUMA(tess.svars.colors[i][0], tess.svars.colors[i][1], tess.svars.colors[i][2]);
tess.svars.colors[i][0] = tess.svars.colors[i][1] = tess.svars.colors[i][2] = scale;
}
}
else if(r_greyscale->value)
{
float scale;
for(i = 0; i < tess.numVertexes; i++)
{
scale = LUMA(tess.svars.colors[i][0], tess.svars.colors[i][1], tess.svars.colors[i][2]);
tess.svars.colors[i][0] = LERP(tess.svars.colors[i][0], scale, r_greyscale->value);
tess.svars.colors[i][1] = LERP(tess.svars.colors[i][1], scale, r_greyscale->value);
tess.svars.colors[i][2] = LERP(tess.svars.colors[i][2], scale, r_greyscale->value);
}
}
}
void setupFilterPrimitives_c(EncoderPrimitives& p)
{
LUMA(4, 4);
LUMA(8, 8);
CHROMA_420(4, 4);
LUMA(4, 8);
CHROMA_420(2, 4);
LUMA(8, 4);
CHROMA_420(4, 2);
LUMA(16, 16);
CHROMA_420(8, 8);
LUMA(16, 8);
CHROMA_420(8, 4);
LUMA(8, 16);
CHROMA_420(4, 8);
LUMA(16, 12);
CHROMA_420(8, 6);
LUMA(12, 16);
CHROMA_420(6, 8);
LUMA(16, 4);
CHROMA_420(8, 2);
LUMA(4, 16);
CHROMA_420(2, 8);
LUMA(32, 32);
CHROMA_420(16, 16);
LUMA(32, 16);
CHROMA_420(16, 8);
LUMA(16, 32);
CHROMA_420(8, 16);
LUMA(32, 24);
CHROMA_420(16, 12);
LUMA(24, 32);
CHROMA_420(12, 16);
LUMA(32, 8);
CHROMA_420(16, 4);
LUMA(8, 32);
CHROMA_420(4, 16);
LUMA(64, 64);
CHROMA_420(32, 32);
LUMA(64, 32);
CHROMA_420(32, 16);
LUMA(32, 64);
CHROMA_420(16, 32);
LUMA(64, 48);
CHROMA_420(32, 24);
LUMA(48, 64);
CHROMA_420(24, 32);
LUMA(64, 16);
CHROMA_420(32, 8);
LUMA(16, 64);
CHROMA_420(8, 32);
CHROMA_422(4, 8);
CHROMA_422(4, 4);
CHROMA_422(2, 4);
CHROMA_422(2, 8);
CHROMA_422(8, 16);
CHROMA_422(8, 8);
CHROMA_422(4, 16);
CHROMA_422(8, 12);
CHROMA_422(6, 16);
CHROMA_422(8, 4);
CHROMA_422(2, 16);
CHROMA_422(16, 32);
CHROMA_422(16, 16);
CHROMA_422(8, 32);
CHROMA_422(16, 24);
CHROMA_422(12, 32);
CHROMA_422(16, 8);
CHROMA_422(4, 32);
CHROMA_422(32, 64);
CHROMA_422(32, 32);
CHROMA_422(16, 64);
CHROMA_422(32, 48);
CHROMA_422(24, 64);
CHROMA_422(32, 16);
CHROMA_422(8, 64);
CHROMA_444(4, 4);
CHROMA_444(8, 8);
CHROMA_444(4, 8);
CHROMA_444(8, 4);
CHROMA_444(16, 16);
CHROMA_444(16, 8);
CHROMA_444(8, 16);
CHROMA_444(16, 12);
CHROMA_444(12, 16);
CHROMA_444(16, 4);
CHROMA_444(4, 16);
CHROMA_444(32, 32);
CHROMA_444(32, 16);
CHROMA_444(16, 32);
CHROMA_444(32, 24);
CHROMA_444(24, 32);
CHROMA_444(32, 8);
CHROMA_444(8, 32);
CHROMA_444(64, 64);
CHROMA_444(64, 32);
CHROMA_444(32, 64);
CHROMA_444(64, 48);
CHROMA_444(48, 64);
CHROMA_444(64, 16);
CHROMA_444(16, 64);
p.extendRowBorder = extendCURowColBorder;
}
void Setup_C_IPFilterPrimitives(EncoderPrimitives& p)
{
LUMA(4, 4);
LUMA(8, 8);
CHROMA_420(4, 4);
LUMA(4, 8);
CHROMA_420(2, 4);
LUMA(8, 4);
CHROMA_420(4, 2);
LUMA(16, 16);
CHROMA_420(8, 8);
LUMA(16, 8);
CHROMA_420(8, 4);
LUMA(8, 16);
CHROMA_420(4, 8);
LUMA(16, 12);
CHROMA_420(8, 6);
LUMA(12, 16);
CHROMA_420(6, 8);
LUMA(16, 4);
CHROMA_420(8, 2);
LUMA(4, 16);
CHROMA_420(2, 8);
LUMA(32, 32);
CHROMA_420(16, 16);
LUMA(32, 16);
CHROMA_420(16, 8);
LUMA(16, 32);
CHROMA_420(8, 16);
LUMA(32, 24);
CHROMA_420(16, 12);
LUMA(24, 32);
CHROMA_420(12, 16);
LUMA(32, 8);
CHROMA_420(16, 4);
LUMA(8, 32);
CHROMA_420(4, 16);
LUMA(64, 64);
CHROMA_420(32, 32);
LUMA(64, 32);
CHROMA_420(32, 16);
LUMA(32, 64);
CHROMA_420(16, 32);
LUMA(64, 48);
CHROMA_420(32, 24);
LUMA(48, 64);
CHROMA_420(24, 32);
LUMA(64, 16);
CHROMA_420(32, 8);
LUMA(16, 64);
CHROMA_420(8, 32);
CHROMA_444(4, 4);
CHROMA_444(8, 8);
CHROMA_444(4, 8);
CHROMA_444(8, 4);
CHROMA_444(16, 16);
CHROMA_444(16, 8);
CHROMA_444(8, 16);
CHROMA_444(16, 12);
CHROMA_444(12, 16);
CHROMA_444(16, 4);
CHROMA_444(4, 16);
CHROMA_444(32, 32);
CHROMA_444(32, 16);
CHROMA_444(16, 32);
CHROMA_444(32, 24);
CHROMA_444(24, 32);
CHROMA_444(32, 8);
CHROMA_444(8, 32);
CHROMA_444(64, 64);
CHROMA_444(64, 32);
CHROMA_444(32, 64);
CHROMA_444(64, 48);
CHROMA_444(48, 64);
CHROMA_444(64, 16);
CHROMA_444(16, 64);
p.luma_p2s = filterConvertPelToShort_c<MAX_CU_SIZE>;
p.chroma_p2s[X265_CSP_I444] = filterConvertPelToShort_c<MAX_CU_SIZE>;
p.chroma_p2s[X265_CSP_I420] = filterConvertPelToShort_c<MAX_CU_SIZE / 2>;
p.extendRowBorder = extendCURowColBorder;
}
static void ProjectDlightTexture_scalar( void ) {
int i, l;
vec3_t origin;
float *texCoords;
byte *colors;
int *intColors;
byte clipBits[SHADER_MAX_VERTEXES];
float texCoordsArray[SHADER_MAX_VERTEXES][2];
byte colorArray[SHADER_MAX_VERTEXES][4];
glIndex_t hitIndexes[SHADER_MAX_INDEXES];
int numIndexes;
float scale;
float radius;
float radiusInverseCubed;
float intensity, remainder;
vec3_t floatColor;
float modulate = 0.0f;
qboolean vertexLight;
if ( !backEnd.refdef.num_dlights ) {
return;
}
for ( l = 0 ; l < backEnd.refdef.num_dlights ; l++ ) {
dlight_t *dl;
if ( !( tess.dlightBits & ( 1 << l ) ) ) {
continue; // this surface definately doesn't have any of this light
}
// clear colors
Com_Memset( colorArray, 0, sizeof( colorArray ) );
texCoords = texCoordsArray[0];
colors = colorArray[0];
dl = &backEnd.refdef.dlights[l];
VectorCopy( dl->transformed, origin );
radius = dl->radius;
scale = 1.0f / radius;
radiusInverseCubed = dl->radiusInverseCubed;
intensity = dl->intensity;
vertexLight = ( ( dl->flags & REF_DIRECTED_DLIGHT ) || ( dl->flags & REF_VERTEX_DLIGHT ) );
// directional lights have max intensity and washout remainder intensity
if ( dl->flags & REF_DIRECTED_DLIGHT ) {
remainder = intensity * 0.125;
} else {
remainder = 0.0f;
}
if(r_greyscale->integer)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = floatColor[1] = floatColor[2] = luminance;
}
else if(r_greyscale->value)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = LERP(dl->color[0] * 255.0f, luminance, r_greyscale->value);
floatColor[1] = LERP(dl->color[1] * 255.0f, luminance, r_greyscale->value);
floatColor[2] = LERP(dl->color[2] * 255.0f, luminance, r_greyscale->value);
}
else
{
floatColor[0] = dl->color[0] * 255.0f;
floatColor[1] = dl->color[1] * 255.0f;
floatColor[2] = dl->color[2] * 255.0f;
}
for ( i = 0 ; i < tess.numVertexes ; i++, texCoords += 2, colors += 4 ) {
int clip = 0;
vec3_t dist;
VectorSubtract( origin, tess.xyz[i], dist );
backEnd.pc.c_dlightVertexes++;
// directional dlight, origin is a directional normal
if ( dl->flags & REF_DIRECTED_DLIGHT ) {
// twosided surfaces use absolute value of the calculated lighting
modulate = intensity * DotProduct( dl->origin, tess.normal[ i ] );
if ( tess.shader->cullType == CT_TWO_SIDED ) {
modulate = fabs( modulate );
}
modulate += remainder;
}
// spherical vertex lit dlight
else if ( dl->flags & REF_VERTEX_DLIGHT )
{
vec3_t dir;
dir[ 0 ] = radius - fabs( dist[ 0 ] );
if ( dir[ 0 ] <= 0.0f ) {
continue;
}
dir[ 1 ] = radius - fabs( dist[ 1 ] );
if ( dir[ 1 ] <= 0.0f ) {
continue;
}
dir[ 2 ] = radius - fabs( dist[ 2 ] );
if ( dir[ 2 ] <= 0.0f ) {
continue;
}
modulate = intensity * dir[ 0 ] * dir[ 1 ] * dir[ 2 ] * radiusInverseCubed;
}
// vertical cylinder dlight
else
{
texCoords[0] = 0.5f + dist[0] * scale;
texCoords[1] = 0.5f + dist[1] * scale;
if( !r_dlightBacks->integer &&
// dist . tess.normal[i]
( dist[0] * tess.normal[i][0] +
dist[1] * tess.normal[i][1] +
dist[2] * tess.normal[i][2] ) < 0.0f ) {
clip = 63;
} else {
if ( texCoords[0] < 0.0f ) {
clip |= 1;
} else if ( texCoords[0] > 1.0f ) {
clip |= 2;
}
if ( texCoords[1] < 0.0f ) {
clip |= 4;
} else if ( texCoords[1] > 1.0f ) {
clip |= 8;
}
texCoords[0] = texCoords[0];
texCoords[1] = texCoords[1];
// modulate the strength based on the height and color
if ( dist[2] > radius ) {
clip |= 16;
modulate = 0.0f;
} else if ( dist[2] < -radius ) {
clip |= 32;
modulate = 0.0f;
} else {
dist[2] = Q_fabs(dist[2]);
if ( dist[2] < radius * 0.5f ) {
modulate = intensity;
} else {
modulate = intensity * 2.0f * (radius - dist[2]) * scale;
}
}
}
}
// optimizations
if ( vertexLight && modulate < ( 1.0f / 128.0f ) ) {
continue;
} else if ( modulate > 1.0f ) {
modulate = 1.0f;
}
clipBits[i] = clip;
colors[0] = Com_Clamp( 0, 255, ri.ftol(floatColor[0] * modulate) );
colors[1] = Com_Clamp( 0, 255, ri.ftol(floatColor[1] * modulate) );
colors[2] = Com_Clamp( 0, 255, ri.ftol(floatColor[2] * modulate) );
colors[3] = 255;
}
// build a list of triangles that need light
intColors = (int*) colorArray;
numIndexes = 0;
for ( i = 0 ; i < tess.numIndexes ; i += 3 ) {
int a, b, c;
a = tess.indexes[i];
b = tess.indexes[i+1];
c = tess.indexes[i+2];
if ( vertexLight ) {
if ( !( intColors[ a ] | intColors[ b ] | intColors[ c ] ) ) {
continue;
}
} else {
if ( clipBits[a] & clipBits[b] & clipBits[c] ) {
continue; // not lighted
}
}
hitIndexes[numIndexes] = a;
hitIndexes[numIndexes+1] = b;
hitIndexes[numIndexes+2] = c;
numIndexes += 3;
}
if ( !numIndexes ) {
continue;
}
if ( !vertexLight ) {
qglEnableClientState( GL_TEXTURE_COORD_ARRAY );
qglTexCoordPointer( 2, GL_FLOAT, 0, texCoordsArray[0] );
} else {
qglDisableClientState( GL_TEXTURE_COORD_ARRAY );
}
qglEnableClientState( GL_COLOR_ARRAY );
qglColorPointer( 4, GL_UNSIGNED_BYTE, 0, colorArray );
if ( dl->dlshader ) {
shader_t *dls = dl->dlshader;
for ( i = 0; i < dls->numUnfoggedPasses; i++ ) {
shaderStage_t *stage = dls->stages[i];
R_BindAnimatedImage( &dls->stages[i]->bundle[0] );
GL_State( stage->stateBits | GLS_DEPTHFUNC_EQUAL );
R_DrawElements( numIndexes, hitIndexes );
backEnd.pc.c_totalIndexes += numIndexes;
backEnd.pc.c_dlightIndexes += numIndexes;
}
} else {
R_FogOff();
if ( !vertexLight ) {
GL_Bind( tr.dlightImage );
} else {
GL_Bind( tr.whiteImage );
}
// include GLS_DEPTHFUNC_EQUAL so alpha tested surfaces don't add light
// where they aren't rendered
if ( dl->flags & REF_ADDITIVE_DLIGHT ) {
GL_State( GLS_SRCBLEND_ONE | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
else {
GL_State( GLS_SRCBLEND_DST_COLOR | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
R_DrawElements( numIndexes, hitIndexes );
backEnd.pc.c_totalIndexes += numIndexes;
backEnd.pc.c_dlightIndexes += numIndexes;
RB_FogOn();
}
}
}
static void ProjectDlightTexture_altivec( void ) {
int i, l;
vec_t origin0, origin1, origin2;
float texCoords0, texCoords1;
vector float floatColorVec0, floatColorVec1;
vector float modulateVec, colorVec, zero;
vector short colorShort;
vector signed int colorInt;
vector unsigned char floatColorVecPerm, modulatePerm, colorChar;
vector unsigned char vSel = VECCONST_UINT8(0x00, 0x00, 0x00, 0xff,
0x00, 0x00, 0x00, 0xff,
0x00, 0x00, 0x00, 0xff,
0x00, 0x00, 0x00, 0xff);
float *texCoords;
byte *colors;
int *intColors;
byte clipBits[SHADER_MAX_VERTEXES];
float texCoordsArray[SHADER_MAX_VERTEXES][2];
byte colorArray[SHADER_MAX_VERTEXES][4];
glIndex_t hitIndexes[SHADER_MAX_INDEXES];
int numIndexes;
float scale;
float radius;
float radiusInverseCubed;
float intensity, remainder;
vec3_t floatColor;
float modulate = 0.0f;
qboolean vertexLight;
if ( !backEnd.refdef.num_dlights ) {
return;
}
// There has to be a better way to do this so that floatColor
// and/or modulate are already 16-byte aligned.
floatColorVecPerm = vec_lvsl(0,(float *)floatColor);
modulatePerm = vec_lvsl(0,(float *)&modulate);
modulatePerm = (vector unsigned char)vec_splat((vector unsigned int)modulatePerm,0);
zero = (vector float)vec_splat_s8(0);
for ( l = 0 ; l < backEnd.refdef.num_dlights ; l++ ) {
dlight_t *dl;
if ( !( tess.dlightBits & ( 1 << l ) ) ) {
continue; // this surface definately doesn't have any of this light
}
// clear colors
Com_Memset( colorArray, 0, sizeof( colorArray ) );
texCoords = texCoordsArray[0];
colors = colorArray[0];
dl = &backEnd.refdef.dlights[l];
origin0 = dl->transformed[0];
origin1 = dl->transformed[1];
origin2 = dl->transformed[2];
radius = dl->radius;
scale = 1.0f / radius;
radiusInverseCubed = dl->radiusInverseCubed;
intensity = dl->intensity;
vertexLight = ( ( dl->flags & REF_DIRECTED_DLIGHT ) || ( dl->flags & REF_VERTEX_DLIGHT ) );
// directional lights have max intensity and washout remainder intensity
if ( dl->flags & REF_DIRECTED_DLIGHT ) {
remainder = intensity * 0.125;
} else {
remainder = 0.0f;
}
if(r_greyscale->integer)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = floatColor[1] = floatColor[2] = luminance;
}
else if(r_greyscale->value)
{
float luminance;
luminance = LUMA(dl->color[0], dl->color[1], dl->color[2]) * 255.0f;
floatColor[0] = LERP(dl->color[0] * 255.0f, luminance, r_greyscale->value);
floatColor[1] = LERP(dl->color[1] * 255.0f, luminance, r_greyscale->value);
floatColor[2] = LERP(dl->color[2] * 255.0f, luminance, r_greyscale->value);
}
else
{
floatColor[0] = dl->color[0] * 255.0f;
floatColor[1] = dl->color[1] * 255.0f;
floatColor[2] = dl->color[2] * 255.0f;
}
floatColorVec0 = vec_ld(0, floatColor);
floatColorVec1 = vec_ld(11, floatColor);
floatColorVec0 = vec_perm(floatColorVec0,floatColorVec0,floatColorVecPerm);
for ( i = 0 ; i < tess.numVertexes ; i++, texCoords += 2, colors += 4 ) {
int clip = 0;
vec_t dist0, dist1, dist2;
dist0 = origin0 - tess.xyz[i][0];
dist1 = origin1 - tess.xyz[i][1];
dist2 = origin2 - tess.xyz[i][2];
backEnd.pc.c_dlightVertexes++;
// directional dlight, origin is a directional normal
if ( dl->flags & REF_DIRECTED_DLIGHT ) {
// twosided surfaces use absolute value of the calculated lighting
modulate = intensity * DotProduct( dl->origin, tess.normal[ i ] );
if ( tess.shader->cullType == CT_TWO_SIDED ) {
modulate = fabs( modulate );
}
modulate += remainder;
}
// spherical vertex lit dlight
else if ( dl->flags & REF_VERTEX_DLIGHT )
{
vec3_t dir;
dir[ 0 ] = radius - fabs( dist0 );
if ( dir[ 0 ] <= 0.0f ) {
continue;
}
dir[ 1 ] = radius - fabs( dist1 );
if ( dir[ 1 ] <= 0.0f ) {
continue;
}
dir[ 2 ] = radius - fabs( dist2 );
if ( dir[ 2 ] <= 0.0f ) {
continue;
}
modulate = intensity * dir[ 0 ] * dir[ 1 ] * dir[ 2 ] * radiusInverseCubed;
}
// vertical cylinder dlight
else
{
texCoords0 = 0.5f + dist0 * scale;
texCoords1 = 0.5f + dist1 * scale;
if( !r_dlightBacks->integer &&
// dist . tess.normal[i]
( dist0 * tess.normal[i][0] +
dist1 * tess.normal[i][1] +
dist2 * tess.normal[i][2] ) < 0.0f ) {
clip = 63;
} else {
if ( texCoords0 < 0.0f ) {
clip |= 1;
} else if ( texCoords0 > 1.0f ) {
clip |= 2;
}
if ( texCoords1 < 0.0f ) {
clip |= 4;
} else if ( texCoords1 > 1.0f ) {
clip |= 8;
}
texCoords[0] = texCoords0;
texCoords[1] = texCoords1;
// modulate the strength based on the height and color
if ( dist2 > radius ) {
clip |= 16;
modulate = 0.0f;
} else if ( dist2 < -radius ) {
clip |= 32;
modulate = 0.0f;
} else {
dist2 = Q_fabs(dist2);
if ( dist2 < radius * 0.5f ) {
modulate = intensity;
} else {
modulate = intensity * 2.0f * (radius - dist2) * scale;
}
}
}
}
clipBits[i] = clip;
// optimizations
if ( vertexLight && modulate < ( 1.0f / 128.0f ) ) {
continue;
} else if ( modulate > 1.0f ) {
modulate = 1.0f;
}
// ZTM: FIXME: should probably clamp to 0-255 range before converting to char,
// but I don't know how to do altvec stuff or if it's even used anymore
modulateVec = vec_ld(0,(float *)&modulate);
modulateVec = vec_perm(modulateVec,modulateVec,modulatePerm);
colorVec = vec_madd(floatColorVec0,modulateVec,zero);
colorInt = vec_cts(colorVec,0); // RGBx
colorShort = vec_pack(colorInt,colorInt); // RGBxRGBx
colorChar = vec_packsu(colorShort,colorShort); // RGBxRGBxRGBxRGBx
colorChar = vec_sel(colorChar,vSel,vSel); // RGBARGBARGBARGBA replace alpha with 255
vec_ste((vector unsigned int)colorChar,0,(unsigned int *)colors); // store color
}
// build a list of triangles that need light
intColors = (int*) colorArray;
numIndexes = 0;
for ( i = 0 ; i < tess.numIndexes ; i += 3 ) {
int a, b, c;
a = tess.indexes[i];
b = tess.indexes[i+1];
c = tess.indexes[i+2];
if ( vertexLight ) {
if ( !( intColors[ a ] | intColors[ b ] | intColors[ c ] ) ) {
continue;
}
} else {
if ( clipBits[a] & clipBits[b] & clipBits[c] ) {
continue; // not lighted
}
}
hitIndexes[numIndexes] = a;
hitIndexes[numIndexes+1] = b;
hitIndexes[numIndexes+2] = c;
numIndexes += 3;
}
if ( !numIndexes ) {
continue;
}
if ( !vertexLight ) {
qglEnableClientState( GL_TEXTURE_COORD_ARRAY );
qglTexCoordPointer( 2, GL_FLOAT, 0, texCoordsArray[0] );
} else {
qglDisableClientState( GL_TEXTURE_COORD_ARRAY );
}
qglEnableClientState( GL_COLOR_ARRAY );
qglColorPointer( 4, GL_UNSIGNED_BYTE, 0, colorArray );
if ( dl->dlshader ) {
shader_t *dls = dl->dlshader;
for ( i = 0; i < dls->numUnfoggedPasses; i++ ) {
shaderStage_t *stage = dls->stages[i];
R_BindAnimatedImage( &dls->stages[i]->bundle[0] );
GL_State( stage->stateBits | GLS_DEPTHFUNC_EQUAL );
R_DrawElements( numIndexes, hitIndexes );
backEnd.pc.c_totalIndexes += numIndexes;
backEnd.pc.c_dlightIndexes += numIndexes;
}
} else {
R_FogOff();
if ( !vertexLight ) {
GL_Bind( tr.dlightImage );
} else {
GL_Bind( tr.whiteImage );
}
// include GLS_DEPTHFUNC_EQUAL so alpha tested surfaces don't add light
// where they aren't rendered
if ( dl->flags & REF_ADDITIVE_DLIGHT ) {
GL_State( GLS_SRCBLEND_ONE | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
else {
GL_State( GLS_SRCBLEND_DST_COLOR | GLS_DSTBLEND_ONE | GLS_DEPTHFUNC_EQUAL );
}
R_DrawElements( numIndexes, hitIndexes );
backEnd.pc.c_totalIndexes += numIndexes;
backEnd.pc.c_dlightIndexes += numIndexes;
RB_FogOn();
}
}
}
// opt can fit in four bits, if the width and height need more than eight bits each.
#define KEY(width, height, bits, opt) (unsigned)(width) << 24 | (height) << 16 | (bits) << 8 | (opt)
#if defined(MVTOOLS_X86)
#define LUMA_SSE2(width, height) \
{ KEY(width, height, 8, SSE2), luma_sse2<width, height> },
#else
#define LUMA_SSE2(width, height)
#endif
#define LUMA(width, height) \
{ KEY(width, height, 8, Scalar), luma_c<width, height, uint8_t> }, \
{ KEY(width, height, 16, Scalar), luma_c<width, height, uint16_t> },
static const std::unordered_map<uint32_t, LUMAFunction> luma_functions = {
LUMA(4, 4)
LUMA(8, 4)
LUMA(8, 8)
LUMA(16, 2)
LUMA(16, 8)
LUMA(16, 16)
LUMA(32, 16)
LUMA(32, 32)
LUMA(64, 32)
LUMA(64, 64)
LUMA(128, 64)
LUMA(128, 128)
LUMA_SSE2(4, 4)
LUMA_SSE2(8, 4)
LUMA_SSE2(8, 8)
LUMA_SSE2(16, 2)
