mirror of
https://github.com/cemu-project/cemu_graphic_packs.git
synced 2024-11-30 05:14:14 +01:00
6c05eebe65
update this first before adding splatoon caution: gpu taxing at high res, if get framerate drop, try replacing them with ones from lower res, bloom will appear smaller and brighter (concentrated) hope cemu will get more specific texture exclude so we don't have to upscale the blurs... any advice on optimization is appreciated
84 lines
2.8 KiB
Plaintext
84 lines
2.8 KiB
Plaintext
#version 420
|
|
#extension GL_ARB_texture_gather : enable
|
|
// shader 10584c6fc5857351 // bloom pre-blur
|
|
uniform ivec4 uf_remappedPS[2];
|
|
layout(binding = 0) uniform sampler2D textureUnitPS0;// Tex0 addr 0xf5c7b800 res 1280x720x1 dim 1 tm: 4 format 0816 compSel: 0 1 2 5 mipView: 0x0 (num 0x1) sliceView: 0x0 (num 0x1) Sampler0 ClampX/Y/Z: 2 2 2 border: 1
|
|
layout(location = 0) in vec4 passParameterSem0;
|
|
layout(location = 1) in vec4 passParameterSem1;
|
|
layout(location = 0) out vec4 passPixelColor0;
|
|
uniform vec2 uf_fragCoordScale;
|
|
int clampFI32(int v)
|
|
{
|
|
if( v == 0x7FFFFFFF )
|
|
return floatBitsToInt(1.0);
|
|
else if( v == 0xFFFFFFFF )
|
|
return floatBitsToInt(0.0);
|
|
return floatBitsToInt(clamp(intBitsToFloat(v), 0.0, 1.0));
|
|
}
|
|
float mul_nonIEEE(float a, float b){ if( a == 0.0 || b == 0.0 ) return 0.0; return a*b; }
|
|
void main()
|
|
{
|
|
vec2 R0f = vec2(0.0);
|
|
vec4 R1f = vec4(0.0);
|
|
vec4 R2f = vec4(0.0);
|
|
vec4 R3f = vec4(0.0);
|
|
vec4 R123f = vec4(0.0);
|
|
vec4 R126f = vec4(0.0);
|
|
vec4 R127f = vec4(0.0);
|
|
float backupReg0f, backupReg1f, backupReg2f, backupReg3f, backupReg4f;
|
|
vec4 PV0f = vec4(0.0), PV1f = vec4(0.0);
|
|
float PS0f = 0.0, PS1f = 0.0;
|
|
vec4 tempf = vec4(0.0);
|
|
float tempResultf;
|
|
int tempResulti;
|
|
ivec4 ARi = ivec4(0);
|
|
bool predResult = true;
|
|
vec3 cubeMapSTM;
|
|
int cubeMapFaceId;
|
|
R0f = vec2((passParameterSem0.x + passParameterSem1.x)/2, (passParameterSem0.y + passParameterSem0.w)/2); //center point
|
|
vec2 res = vec2( passParameterSem1.x - passParameterSem0.x, passParameterSem0.w - passParameterSem0.y ) * uf_fragCoordScale;
|
|
int r = int(floor(1.0 / uf_fragCoordScale.y + 0.5));
|
|
float count = 0.0;
|
|
for( int x=-r; x<=r; x++ ) {
|
|
for( int y=-r; y<=r; y++ ) {
|
|
if( pow(x,2) + pow(y,2) <= pow(r,2) ) {
|
|
R1f += texture( textureUnitPS0, R0f + vec2(x,y)*res );
|
|
count += 1.0;
|
|
}
|
|
}
|
|
}
|
|
PV0f = R1f/count;
|
|
R127f.xzw = PV0f.xzw;
|
|
R126f.y = PV0f.y;
|
|
// 3
|
|
tempf.x = dot(vec4(PV0f.x,PV0f.y,PV0f.z,-0.0),vec4(intBitsToFloat(uf_remappedPS[0].x),intBitsToFloat(uf_remappedPS[0].y),intBitsToFloat(uf_remappedPS[0].z),0.0));
|
|
PV1f.x = tempf.x;
|
|
PV1f.y = tempf.x;
|
|
PV1f.z = tempf.x;
|
|
PV1f.w = tempf.x;
|
|
PS1f = mul_nonIEEE(PV0f.w, intBitsToFloat(uf_remappedPS[0].w));
|
|
// 4
|
|
tempf.x = dot(vec4(R127f.x,R126f.y,R127f.z,PS1f),vec4(intBitsToFloat(uf_remappedPS[0].x),intBitsToFloat(uf_remappedPS[0].y),intBitsToFloat(uf_remappedPS[0].z),1.0));
|
|
tempf.x = clamp(tempf.x, 0.0, 1.0);
|
|
PV0f.x = tempf.x;
|
|
PV0f.y = tempf.x;
|
|
PV0f.z = tempf.x;
|
|
PV0f.w = tempf.x;
|
|
R127f.y = tempf.x;
|
|
PS0f = 1.0 / PV1f.x;
|
|
// 5
|
|
PV1f.z = PS0f * intBitsToFloat(uf_remappedPS[1].x);
|
|
PV1f.z = clamp(PV1f.z, 0.0, 1.0);
|
|
// 6
|
|
PV0f.y = mul_nonIEEE(R127f.y, PV1f.z);
|
|
// 7
|
|
PV1f.x = mul_nonIEEE(PV0f.y, intBitsToFloat(uf_remappedPS[1].z));
|
|
// 8
|
|
R1f.x = mul_nonIEEE(R127f.x, PV1f.x);
|
|
R1f.y = mul_nonIEEE(R126f.y, PV1f.x);
|
|
R1f.z = mul_nonIEEE(R127f.z, PV1f.x);
|
|
R1f.w = mul_nonIEEE(R127f.w, PV1f.x);
|
|
// export
|
|
passPixelColor0 = vec4(R1f.x, R1f.y, R1f.z, R1f.w);
|
|
}
|