[XCX] 21:9 fix flashlight, blur scaling less aggressive on low res

This commit is contained in:
getdls 2018-03-19 00:09:03 +01:00
parent da421eb94a
commit 9a9362e0c3
3 changed files with 282 additions and 4 deletions

View File

@ -0,0 +1,277 @@
#version 420
#extension GL_ARB_texture_gather : enable
#extension GL_ARB_separate_shader_objects : enable
#extension GL_ARB_shading_language_packing : enable
// shader 610a9c4cb60b0bdf
//flashlight
uniform ivec4 uf_remappedVS[13];
uniform vec2 uf_windowSpaceToClipSpaceTransform;
layout(location = 0) in uvec4 attrDataSem0;
layout(location = 1) in uvec4 attrDataSem1;
layout(location = 2) in uvec4 attrDataSem2;
layout(location = 3) in uvec4 attrDataSem3;
layout(location = 4) in uvec4 attrDataSem5;
layout(location = 5) in uvec4 attrDataSem6;
layout(location = 6) in uvec4 attrDataSem7;
out gl_PerVertex
{
vec4 gl_Position;
float gl_PointSize;
};
layout(location = 0) out vec4 passParameterSem0;
layout(location = 1) out vec4 passParameterSem2;
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()
{
ivec4 R0i = ivec4(0);
ivec4 R1i = ivec4(0);
ivec4 R2i = ivec4(0);
ivec4 R3i = ivec4(0);
ivec4 R4i = ivec4(0);
ivec4 R5i = ivec4(0);
ivec4 R6i = ivec4(0);
ivec4 R7i = ivec4(0);
ivec4 R123i = ivec4(0);
ivec4 R124i = ivec4(0);
ivec4 R125i = ivec4(0);
ivec4 R126i = ivec4(0);
ivec4 R127i = ivec4(0);
uvec4 attrDecoder;
int backupReg0i, backupReg1i, backupReg2i, backupReg3i, backupReg4i;
ivec4 PV0i = ivec4(0), PV1i = ivec4(0);
int PS0i = 0, PS1i = 0;
ivec4 tempi = ivec4(0);
float tempResultf;
int tempResulti;
ivec4 ARi = ivec4(0);
bool predResult = true;
vec3 cubeMapSTM;
int cubeMapFaceId;
R0i = ivec4(gl_VertexID, 0, 0, gl_InstanceID);
attrDecoder.xy = attrDataSem0.xy;
attrDecoder.xy = (attrDecoder.xy>>24)|((attrDecoder.xy>>8)&0xFF00)|((attrDecoder.xy<<8)&0xFF0000)|((attrDecoder.xy<<24));
attrDecoder.z = 0;
attrDecoder.w = 0;
R1i = ivec4(int(attrDecoder.x), int(attrDecoder.y), floatBitsToInt(0.0), floatBitsToInt(1.0));
attrDecoder = attrDataSem1;
attrDecoder = (attrDecoder>>24)|((attrDecoder>>8)&0xFF00)|((attrDecoder<<8)&0xFF0000)|((attrDecoder<<24));
R2i = ivec4(int(attrDecoder.x), int(attrDecoder.y), int(attrDecoder.z), int(attrDecoder.w));
attrDecoder = attrDataSem5;
attrDecoder = (attrDecoder>>24)|((attrDecoder>>8)&0xFF00)|((attrDecoder<<8)&0xFF0000)|((attrDecoder<<24));
R5i = ivec4(int(attrDecoder.x), int(attrDecoder.y), int(attrDecoder.z), int(attrDecoder.w));
attrDecoder.xyz = attrDataSem6.xyz;
attrDecoder.xyz = (attrDecoder.xyz>>24)|((attrDecoder.xyz>>8)&0xFF00)|((attrDecoder.xyz<<8)&0xFF0000)|((attrDecoder.xyz<<24));
attrDecoder.w = 0;
R6i = ivec4(int(attrDecoder.x), int(attrDecoder.y), int(attrDecoder.z), floatBitsToInt(1.0));
attrDecoder.xyz = attrDataSem3.xyz;
attrDecoder.xyz = (attrDecoder.xyz>>24)|((attrDecoder.xyz>>8)&0xFF00)|((attrDecoder.xyz<<8)&0xFF0000)|((attrDecoder.xyz<<24));
attrDecoder.w = 0;
R4i = ivec4(int(attrDecoder.x), int(attrDecoder.y), int(attrDecoder.z), floatBitsToInt(1.0));
attrDecoder = attrDataSem7;
attrDecoder = (attrDecoder>>24)|((attrDecoder>>8)&0xFF00)|((attrDecoder<<8)&0xFF0000)|((attrDecoder<<24));
R7i = ivec4(int(attrDecoder.x), int(attrDecoder.y), int(attrDecoder.z), int(attrDecoder.w));
attrDecoder.xyzw = floatBitsToUint(vec4(attrDataSem2.xyzw)/255.0);
R3i = ivec4(int(attrDecoder.x), int(attrDecoder.y), int(attrDecoder.z), int(attrDecoder.w));
// 0
R123i.x = floatBitsToInt((intBitsToFloat(R5i.z) * intBitsToFloat(0x3e22f983) + 0.5));
PV0i.x = R123i.x;
PV0i.y = floatBitsToInt(intBitsToFloat(R2i.w) + -(0.5));
PV0i.z = floatBitsToInt(mul_nonIEEE(intBitsToFloat(R4i.x), intBitsToFloat(R6i.x)));
PV0i.w = floatBitsToInt(mul_nonIEEE(intBitsToFloat(R4i.y), intBitsToFloat(R6i.y)));
PS0i = floatBitsToInt(mul_nonIEEE(intBitsToFloat(R4i.z), intBitsToFloat(R6i.z)));
// 1
R123i.x = ((intBitsToFloat(PV0i.y) >= 0.0)?(floatBitsToInt(1.0)):(0));
PV1i.x = R123i.x;
PV1i.y = floatBitsToInt(-(intBitsToFloat(R4i.x)) + intBitsToFloat(PV0i.z));
PV1i.z = floatBitsToInt(fract(intBitsToFloat(PV0i.x)));
PV1i.w = floatBitsToInt(-(intBitsToFloat(R4i.y)) + intBitsToFloat(PV0i.w));
PS1i = floatBitsToInt(-(intBitsToFloat(R4i.z)) + intBitsToFloat(PS0i));
// 2
R127i.x = floatBitsToInt((mul_nonIEEE(intBitsToFloat(PV1i.w),intBitsToFloat(PV1i.x)) + intBitsToFloat(R4i.y)));
R127i.y = floatBitsToInt((mul_nonIEEE(intBitsToFloat(PV1i.y),intBitsToFloat(PV1i.x)) + intBitsToFloat(R4i.x)));
R123i.z = floatBitsToInt((mul_nonIEEE(intBitsToFloat(PS1i),intBitsToFloat(PV1i.x)) + intBitsToFloat(R4i.z)));
PV0i.z = R123i.z;
R123i.w = floatBitsToInt((intBitsToFloat(PV1i.z) * intBitsToFloat(0x40c90fdb) + -(intBitsToFloat(0x40490fdb))));
PV0i.w = R123i.w;
PS0i = 0;
// 3
PV1i.x = floatBitsToInt(mul_nonIEEE(intBitsToFloat(R1i.x), intBitsToFloat(R6i.x)));
R126i.y = floatBitsToInt(intBitsToFloat(PV0i.w) * intBitsToFloat(0x3e22f983));
PV1i.y = R126i.y;
PV1i.z = floatBitsToInt(mul_nonIEEE(intBitsToFloat(R1i.y), intBitsToFloat(R6i.y)));
R125i.w = floatBitsToInt(intBitsToFloat(PS0i) + intBitsToFloat(PV0i.z));
R126i.z = 0;
PS1i = R126i.z;
// 4
R126i.x = floatBitsToInt(intBitsToFloat(PV1i.x) + intBitsToFloat(R127i.y));
R125i.y = floatBitsToInt(intBitsToFloat(PV1i.z) + intBitsToFloat(R127i.x));
R127i.z = floatBitsToInt(intBitsToFloat(uf_remappedVS[0].x) * 1.0);
R127i.w = floatBitsToInt(intBitsToFloat(uf_remappedVS[0].y) * 1.0);
PS0i = floatBitsToInt(sin((intBitsToFloat(PV1i.y))/0.1591549367));
// 5
backupReg0i = R126i.y;
R127i.x = PS0i;
R126i.y = floatBitsToInt(-(intBitsToFloat(PS0i)));
R125i.z = R125i.w;
R126i.w = floatBitsToInt(intBitsToFloat(uf_remappedVS[0].z) * 1.0);
PS1i = floatBitsToInt(cos((intBitsToFloat(backupReg0i))/0.1591549367));
// 6
backupReg0i = R127i.z;
backupReg1i = R127i.w;
PV0i.x = PS1i;
R127i.y = PS1i;
R127i.z = floatBitsToInt(intBitsToFloat(uf_remappedVS[0].w) * 1.0);
R127i.w = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.z),intBitsToFloat(uf_remappedVS[1].x)) + intBitsToFloat(backupReg0i)));
R124i.z = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.z),intBitsToFloat(uf_remappedVS[1].y)) + intBitsToFloat(backupReg1i)));
PS0i = R124i.z;
// 7
backupReg0i = R126i.w;
R125i.x = floatBitsToInt(dot(vec4(intBitsToFloat(PV0i.x),intBitsToFloat(R126i.y),intBitsToFloat(R126i.z),-0.0),vec4(intBitsToFloat(R126i.x),intBitsToFloat(R125i.y),intBitsToFloat(R125i.w),0.0)));
PV1i.x = R125i.x;
PV1i.y = R125i.x;
PV1i.z = R125i.x;
PV1i.w = R125i.x;
R126i.w = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.z),intBitsToFloat(uf_remappedVS[1].z)) + intBitsToFloat(backupReg0i)));
PS1i = R126i.w;
// 8
backupReg0i = R127i.y;
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R127i.x),intBitsToFloat(backupReg0i),intBitsToFloat(R126i.z),-0.0),vec4(intBitsToFloat(R126i.x),intBitsToFloat(R125i.y),intBitsToFloat(R125i.w),0.0)));
PV0i.x = tempi.x;
PV0i.y = tempi.x;
PV0i.z = tempi.x;
PV0i.w = tempi.x;
R127i.y = tempi.x;
R125i.y = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.z),intBitsToFloat(uf_remappedVS[1].w)) + intBitsToFloat(R127i.z)));
PS0i = R125i.y;
// 9
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R125i.x),intBitsToFloat(PV0i.x),intBitsToFloat(R125i.z),-0.0),vec4(intBitsToFloat(uf_remappedVS[2].x),intBitsToFloat(uf_remappedVS[2].y),intBitsToFloat(uf_remappedVS[2].z),0.0)));
PV1i.x = tempi.x;
PV1i.y = tempi.x;
PV1i.z = tempi.x;
PV1i.w = tempi.x;
R127i.z = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R1i.y),intBitsToFloat(R7i.w)) + intBitsToFloat(R7i.y)));
PS1i = R127i.z;
// 10
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R125i.x),intBitsToFloat(R127i.y),intBitsToFloat(R125i.z),-0.0),vec4(intBitsToFloat(uf_remappedVS[3].x),intBitsToFloat(uf_remappedVS[3].y),intBitsToFloat(uf_remappedVS[3].z),0.0)));
PV0i.x = tempi.x;
PV0i.y = tempi.x;
PV0i.z = tempi.x;
PV0i.w = tempi.x;
R127i.x = floatBitsToInt(intBitsToFloat(PV1i.x) * 1.0);
PS0i = R127i.x;
// 11
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R125i.x),intBitsToFloat(R127i.y),intBitsToFloat(R125i.z),-0.0),vec4(intBitsToFloat(uf_remappedVS[4].x),intBitsToFloat(uf_remappedVS[4].y),intBitsToFloat(uf_remappedVS[4].z),0.0)));
PV1i.x = tempi.x;
PV1i.y = tempi.x;
PV1i.z = tempi.x;
PV1i.w = tempi.x;
R125i.w = floatBitsToInt(intBitsToFloat(PV0i.x) * 1.0);
PS1i = R125i.w;
// 12
R123i.x = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.y),intBitsToFloat(uf_remappedVS[5].x)) + intBitsToFloat(R127i.w)));
PV0i.x = R123i.x;
R123i.y = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.y),intBitsToFloat(uf_remappedVS[5].w)) + intBitsToFloat(R125i.y)));
PV0i.y = R123i.y;
R123i.z = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.y),intBitsToFloat(uf_remappedVS[5].y)) + intBitsToFloat(R124i.z)));
PV0i.z = R123i.z;
R123i.w = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.y),intBitsToFloat(uf_remappedVS[5].z)) + intBitsToFloat(R126i.w)));
PV0i.w = R123i.w;
R124i.z = floatBitsToInt(intBitsToFloat(PV1i.x) * 1.0);
PS0i = R124i.z;
// 13
R123i.x = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.x),intBitsToFloat(uf_remappedVS[6].x)) + intBitsToFloat(PV0i.x)));
PV1i.x = R123i.x;
R123i.y = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.x),intBitsToFloat(uf_remappedVS[6].z)) + intBitsToFloat(PV0i.w)));
PV1i.y = R123i.y;
R123i.z = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.x),intBitsToFloat(uf_remappedVS[6].y)) + intBitsToFloat(PV0i.z)));
PV1i.z = R123i.z;
R126i.w = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R2i.x),intBitsToFloat(uf_remappedVS[6].w)) + intBitsToFloat(PV0i.y)));
R127i.w = floatBitsToInt((mul_nonIEEE(intBitsToFloat(R1i.x),intBitsToFloat(R7i.z)) + intBitsToFloat(R7i.x)));
PS1i = R127i.w;
// 14
backupReg0i = R127i.x;
backupReg1i = R124i.z;
R127i.x = floatBitsToInt(intBitsToFloat(PV1i.x) + intBitsToFloat(backupReg0i));
PV0i.x = R127i.x;
R127i.y = floatBitsToInt(intBitsToFloat(PV1i.z) + intBitsToFloat(R125i.w));
PV0i.y = R127i.y;
R124i.z = floatBitsToInt(intBitsToFloat(PV1i.y) + intBitsToFloat(backupReg1i));
PV0i.z = R124i.z;
R125i.w = ((0.0 >= intBitsToFloat(uf_remappedVS[7].w))?int(0xFFFFFFFF):int(0x0));
R125i.y = floatBitsToInt(intBitsToFloat(R127i.z) + 0.5);
PS0i = R125i.y;
// 15
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(PV0i.x),intBitsToFloat(PV0i.y),intBitsToFloat(PV0i.z),intBitsToFloat(R126i.w)),vec4(intBitsToFloat(uf_remappedVS[8].x),intBitsToFloat(uf_remappedVS[8].y),intBitsToFloat(uf_remappedVS[8].z),intBitsToFloat(uf_remappedVS[8].w))));
PV1i.x = tempi.x;
PV1i.y = tempi.x;
PV1i.z = tempi.x;
PV1i.w = tempi.x;
R7i.x = floatBitsToInt(intBitsToFloat(R127i.w) + 0.5);
PS1i = R7i.x;
// 16
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R127i.x),intBitsToFloat(R127i.y),intBitsToFloat(R124i.z),intBitsToFloat(R126i.w)),vec4(intBitsToFloat(uf_remappedVS[9].x),intBitsToFloat(uf_remappedVS[9].y),intBitsToFloat(uf_remappedVS[9].z),intBitsToFloat(uf_remappedVS[9].w))));
PV0i.x = tempi.x;
PV0i.y = tempi.x;
PV0i.z = tempi.x;
PV0i.w = tempi.x;
R2i.x = floatBitsToInt(intBitsToFloat(PV1i.x) + 0.0);
PS0i = R2i.x;
// 17
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R127i.x),intBitsToFloat(R127i.y),intBitsToFloat(R124i.z),intBitsToFloat(R126i.w)),vec4(intBitsToFloat(uf_remappedVS[10].x),intBitsToFloat(uf_remappedVS[10].y),intBitsToFloat(uf_remappedVS[10].z),intBitsToFloat(uf_remappedVS[10].w))));
PV1i.x = tempi.x;
PV1i.y = tempi.x;
PV1i.z = tempi.x;
PV1i.w = tempi.x;
R127i.z = tempi.x;
R2i.y = floatBitsToInt(intBitsToFloat(PV0i.x) + 0.0);
PS1i = R2i.y;
// 18
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R127i.x),intBitsToFloat(R127i.y),intBitsToFloat(R124i.z),intBitsToFloat(R126i.w)),vec4(intBitsToFloat(uf_remappedVS[11].x),intBitsToFloat(uf_remappedVS[11].y),intBitsToFloat(uf_remappedVS[11].z),intBitsToFloat(uf_remappedVS[11].w))));
PV0i.x = tempi.x;
PV0i.y = tempi.x;
PV0i.z = tempi.x;
PV0i.w = tempi.x;
R2i.w = tempi.x;
R1i.x = ((R125i.w == 0)?(R2i.x):(0x3f800000));
PS0i = R1i.x;
// 19
PV1i.x = floatBitsToInt(mul_nonIEEE(intBitsToFloat(R126i.w), intBitsToFloat(uf_remappedVS[12].w)));
R1i.y = ((R125i.w == 0)?(R2i.y):(0x3f800000));
R1i.z = ((R125i.w == 0)?(R127i.z):(0x3f800000));
R1i.w = ((R125i.w == 0)?(PV0i.x):(0x3f800000));
R7i.y = floatBitsToInt(-(intBitsToFloat(R125i.y)) + 1.0);
PS1i = R7i.y;
// 20
backupReg0i = R3i.w;
tempi.x = floatBitsToInt(dot(vec4(intBitsToFloat(R127i.x),intBitsToFloat(R127i.y),intBitsToFloat(R124i.z),intBitsToFloat(PV1i.x)),vec4(intBitsToFloat(uf_remappedVS[12].x),intBitsToFloat(uf_remappedVS[12].y),intBitsToFloat(uf_remappedVS[12].z),1.0)));
PV0i.x = tempi.x;
PV0i.y = tempi.x;
PV0i.z = tempi.x;
PV0i.w = tempi.x;
R3i.w = backupReg0i;
R3i.w = floatBitsToInt(intBitsToFloat(R3i.w) * 2.0);
PS0i = R3i.w;
// 21
R2i.z = floatBitsToInt(-(intBitsToFloat(PV0i.x)));
// export
float origRatio = 1.777f;
float newRatio = 2.370f;
// *(origRatio / newRatio)
gl_Position = vec4(intBitsToFloat(R1i.x)*(origRatio / newRatio), intBitsToFloat(R1i.y), intBitsToFloat(R1i.z), intBitsToFloat(R1i.w));
// export
passParameterSem0 = vec4(intBitsToFloat(R3i.x), intBitsToFloat(R3i.y), intBitsToFloat(R3i.z), intBitsToFloat(R3i.w));
// export
// skipped export to semanticId 255
// export
passParameterSem2 = vec4(intBitsToFloat(R7i.x), intBitsToFloat(R7i.y), intBitsToFloat(R7i.z), intBitsToFloat(R7i.z));
// 0
}

View File

@ -12,6 +12,7 @@ $scaleFactorY = always_decimal_format($fullHeight / 720.0);
// 1/4 -> 1/16 bloom pyramid . Pixelated unless scaled but still needs blur for light bleed..
// To-do. Check if screen res is * samples stable
const float resScale = <?=$scaleFactorX?>;
const int sampleScale = <?=round($scaleFactorX)?>;
const float lightBloom = 0.95;
highp float lineRand(vec2 co)
{
@ -31,7 +32,7 @@ uniform vec2 uf_fragCoordScale;
// FabriceNeyret2, single pass gaussian by intermediate MIPmap level. https://www.shadertoy.com/view/ltScRG
// I hereby pledge my loyalty to the FabriceNeyret2 fanclub, this is bloody beautiful!
const int samples = 16, //check if must scale to pascal levels
const int samples = 8 * sampleScale, //8 or 4 balances xy position
LOD = 2, // gaussian done on MIPmap at scale LOD
sLOD = 1 << LOD; // tile size = 2^LOD
const float sigma = float(samples) * .25;

View File

@ -12,7 +12,7 @@ $scaleFactorY = always_decimal_format($fullHeight / 720.0);
// shader d8e69e8df8c227f5
// Bloom/blur 1st step, 1280->640->1280
// To do, go back and remove all test blurs..
uniform ivec4 uf_remappedPS[3];
layout(binding = 0) uniform sampler2D textureUnitPS0;// Tex0 addr 0xf470a000 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: 0
layout(binding = 1) uniform sampler2D textureUnitPS1;// Tex1 addr 0xf4386000 res 1280x720x1 dim 1 tm: 4 format 001a compSel: 0 1 2 3 mipView: 0x0 (num 0x1) sliceView: 0x0 (num 0x1) Sampler1 ClampX/Y/Z: 2 2 2 border: 0
@ -21,7 +21,7 @@ layout(location = 0) out vec4 passPixelColor0;
uniform vec2 uf_fragCoordScale;
const float resScale = <?=$scaleFactorX?>;
const int sampleScale = <?=round($scaleFactorX)?>;
//const float resScale = 3.5; //3.5 looks nicer
highp float lineRand(vec2 co)
{
@ -37,7 +37,7 @@ highp float lineRand(vec2 co)
// FabriceNeyret2, single pass gaussian by intermediate MIPmap level. https://www.shadertoy.com/view/ltScRG
// I hereby pledge my loyalty to the FabriceNeyret2 fanclub, this is bloody beautiful!
const int samples = 12 * sampleScale, //check if must scale to pascal levels
const int samples = 8 * sampleScale, //8 or 4 balances xy position
LOD = 2, // gaussian done on MIPmap at scale LOD
sLOD = 1 << LOD; // tile size = 2^LOD
const float sigma = float(samples) * .25;