From 5bd7cbb8fddea23434efd2cf733473d15fc7a001 Mon Sep 17 00:00:00 2001 From: M&M Date: Sat, 4 Jul 2020 20:46:21 -0700 Subject: [PATCH] try to fix MK8 FXAA scaling with in-game mod --- .../be99d80628d31127_00000000000003c9_ps.txt | 210 ++++++++++++++++++ 1 file changed, 210 insertions(+) create mode 100644 Resolutions/MarioKart8_Resolution/be99d80628d31127_00000000000003c9_ps.txt diff --git a/Resolutions/MarioKart8_Resolution/be99d80628d31127_00000000000003c9_ps.txt b/Resolutions/MarioKart8_Resolution/be99d80628d31127_00000000000003c9_ps.txt new file mode 100644 index 00000000..e6ebb858 --- /dev/null +++ b/Resolutions/MarioKart8_Resolution/be99d80628d31127_00000000000003c9_ps.txt @@ -0,0 +1,210 @@ +#version 430 +#extension GL_ARB_texture_gather : enable +#extension GL_ARB_separate_shader_objects : enable +// shader be99d80628d31127 - FXAA in-game (with mod) +const float resScale = float($width)/float($gameWidth); +// start of shader inputs/outputs, predetermined by Cemu. Do not touch +#ifdef VULKAN +#define ATTR_LAYOUT(__vkSet, __location) layout(set = __vkSet, location = __location) +#define UNIFORM_BUFFER_LAYOUT(__glLocation, __vkSet, __vkLocation) layout(set = __vkSet, binding = __vkLocation, std140) +#define TEXTURE_LAYOUT(__glLocation, __vkSet, __vkLocation) layout(set = __vkSet, binding = __vkLocation) +#define GET_FRAGCOORD() vec4(gl_FragCoord.xy*uf_fragCoordScale.xy,gl_FragCoord.zw) +#else +#define ATTR_LAYOUT(__vkSet, __location) layout(location = __location) +#define UNIFORM_BUFFER_LAYOUT(__glLocation, __vkSet, __vkLocation) layout(binding = __glLocation, std140) +#define TEXTURE_LAYOUT(__glLocation, __vkSet, __vkLocation) layout(binding = __glLocation) +#define GET_FRAGCOORD() vec4(gl_FragCoord.xy*uf_fragCoordScale,gl_FragCoord.zw) +#endif +#ifdef VULKAN +layout(set = 1, binding = 2) uniform ufBlock +{ +uniform ivec4 uf_remappedPS[4]; +uniform vec4 uf_fragCoordScale; +}; +#else +uniform ivec4 uf_remappedPS[4]; +uniform vec2 uf_fragCoordScale; +#endif +TEXTURE_LAYOUT(0, 1, 0) uniform sampler2D textureUnitPS0; +TEXTURE_LAYOUT(1, 1, 1) uniform sampler2D textureUnitPS1; +layout(location = 0) in vec4 passParameterSem2; +layout(location = 0) out vec4 passPixelColor0; +// end of shader inputs/outputs +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() +{ +vec4 R0f = vec4(0.0); +vec4 R1f = vec4(0.0); +vec4 R2f = vec4(0.0); +vec4 R3f = vec4(0.0); +vec4 R4f = vec4(0.0); +vec4 R5f = 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; +bool activeMaskStack[2]; +bool activeMaskStackC[3]; +activeMaskStack[0] = false; +activeMaskStackC[0] = false; +activeMaskStackC[1] = false; +activeMaskStack[0] = true; +activeMaskStackC[0] = true; +activeMaskStackC[1] = true; +vec3 cubeMapSTM; +int cubeMapFaceId; +R0f = passParameterSem2; +if( activeMaskStackC[1] == true ) { +R4f.xyzw = (texture(textureUnitPS0, R0f.xy).xyzw); +R2f.xyzw = (textureGather(textureUnitPS1, R0f.xy).wzxy); +} +if( activeMaskStackC[1] == true ) { +activeMaskStack[1] = activeMaskStack[0]; +activeMaskStackC[2] = activeMaskStackC[1]; +// 0 +PV0f.x = min(R2f.z, R2f.x); +PV0f.y = max(R2f.z, R2f.x); +PV0f.z = mul_nonIEEE(R4f.x, intBitsToFloat(uf_remappedPS[0].x)); +PV0f.w = min(R2f.w, R2f.y); +PS0f = max(R2f.w, R2f.y); +// 1 +PV1f.x = min(PV0f.x, PV0f.w); +R123f.y = (mul_nonIEEE(R4f.y,intBitsToFloat(uf_remappedPS[0].y)) + PV0f.z); +PV1f.y = R123f.y; +R127f.z = R2f.z + -(R2f.y); +PV1f.z = R127f.z; +PV1f.w = max(PV0f.y, PS0f); +R126f.z = R2f.w + -(R2f.x); +PS1f = R126f.z; +// 2 +PV0f.x = mul_nonIEEE(PV1f.w, intBitsToFloat(uf_remappedPS[1].x)); +PV0f.y = max(PV1f.y, PV1f.w); +PV0f.z = min(PV1f.y, PV1f.x); +R3f.x = PV1f.z + PS1f; +PS0f = R3f.x; +// 3 +R1f.x = max(PV0f.x, intBitsToFloat(uf_remappedPS[1].y)); +R3f.y = -(PV0f.z) + PV0f.y; +R1f.y = R127f.z + -(R126f.z); +PS1f = R1f.y; +// 4 +predResult = (R3f.y > R1f.x); +activeMaskStack[1] = predResult; +activeMaskStackC[2] = predResult == true && activeMaskStackC[1] == true; +} +else { +activeMaskStack[1] = false; +activeMaskStackC[2] = false; +} +if( activeMaskStackC[2] == true ) { +// 0 +backupReg0f = R2f.y; +R1f.x = max(R3f.x, -(R3f.x)); +PV0f.x = R1f.x; +R2f.y = backupReg0f + R2f.x; +PV0f.y = R2f.y; +R0f.z = intBitsToFloat(uf_remappedPS[2].z) * 0.25; +R0f.w = max(R1f.y, -(R1f.y)); +PV0f.w = R0f.w; +R2f.x = -(intBitsToFloat(uf_remappedPS[3].x)); +PS0f = R2f.x; +// 1 +R3f.y = R2f.z + PV0f.y; +PV1f.y = R3f.y; +R2f.y = min(PV0f.x, PV0f.w); +PS1f = R2f.y; +// 2 +R3f.y = R2f.w + PV1f.y; +PV0f.y = R3f.y; +R1f.z = intBitsToFloat(uf_remappedPS[3].x) / resScale; +R0f.w = intBitsToFloat(uf_remappedPS[3].y) / resScale; +R5f.y = -(intBitsToFloat(uf_remappedPS[3].y)) / resScale; +PS0f = R5f.y; +// 3 +backupReg0f = R0f.z; +R0f.z = (mul_nonIEEE(backupReg0f,PV0f.y) + intBitsToFloat(uf_remappedPS[2].w)); +PV1f.z = R0f.z; +// 4 +backupReg0f = R2f.y; +R2f.y = max(PV1f.z, backupReg0f); +PV0f.y = R2f.y; +// 5 +R2f.y = 1.0 / PV0f.y; +PS1f = R2f.y; +// 6 +backupReg0f = R1f.y; +R1f.x = mul_nonIEEE(R3f.x, PS1f); +PV0f.x = R1f.x; +R1f.y = mul_nonIEEE(backupReg0f, PS1f); +PV0f.y = R1f.y; +// 7 +R1f.x = max(PV0f.x, -(intBitsToFloat(uf_remappedPS[2].y))); +PV1f.x = R1f.x; +R1f.y = max(PV0f.y, -(intBitsToFloat(uf_remappedPS[2].y))); +PV1f.y = R1f.y; +// 8 +R1f.x = min(PV1f.x, intBitsToFloat(uf_remappedPS[2].y)); +PV0f.x = R1f.x; +R1f.y = min(PV1f.y, intBitsToFloat(uf_remappedPS[2].y)); +PV0f.y = R1f.y; +// 9 +backupReg0f = R0f.x; +backupReg1f = R0f.y; +backupReg0f = R0f.x; +backupReg2f = R0f.w; +backupReg1f = R0f.y; +R0f.x = (mul_nonIEEE(PV0f.x,R2f.x) + backupReg0f); +R0f.y = (mul_nonIEEE(PV0f.y,R5f.y) + backupReg1f); +R0f.z = (mul_nonIEEE(PV0f.x,R1f.z) + backupReg0f); +R0f.w = (mul_nonIEEE(PV0f.y,backupReg2f) + backupReg1f); +} +if( activeMaskStackC[2] == true ) { +R1f.xyzw = (texture(textureUnitPS0, R0f.zw).xyzw); +R0f.xyzw = (texture(textureUnitPS0, R0f.xy).xyzw); +} +if( activeMaskStackC[2] == true ) { +// 0 +backupReg0f = R0f.y; +backupReg1f = R0f.x; +PV0f.x = R0f.w + R1f.w; +PV0f.x /= 2.0; +PV0f.y = R0f.z + R1f.z; +PV0f.y /= 2.0; +PV0f.z = backupReg0f + R1f.y; +PV0f.z /= 2.0; +PV0f.w = backupReg1f + R1f.x; +PV0f.w /= 2.0; +// 1 +PV1f.x = -(R4f.w) + PV0f.x; +PV1f.y = -(R4f.z) + PV0f.y; +PV1f.z = -(R4f.y) + PV0f.z; +PV1f.w = -(R4f.x) + PV0f.w; +// 2 +backupReg0f = R4f.x; +backupReg1f = R4f.y; +backupReg2f = R4f.z; +backupReg3f = R4f.w; +R4f.x = (PV1f.w * intBitsToFloat(0x3f4ccccd) + backupReg0f); +R4f.y = (PV1f.z * intBitsToFloat(0x3f4ccccd) + backupReg1f); +R4f.z = (PV1f.y * intBitsToFloat(0x3f4ccccd) + backupReg2f); +R4f.w = (PV1f.x * intBitsToFloat(0x3f4ccccd) + backupReg3f); +} +activeMaskStackC[1] = activeMaskStack[0] == true && activeMaskStackC[0] == true; +// export +passPixelColor0 = vec4(R4f.x, R4f.y, R4f.z, R4f.w); +}