2
1
Fork 0
mirror of https://github.com/yuzu-emu/yuzu.git synced 2024-07-04 23:31:19 +01:00

Pica/Rasterizer: Add initial implementation of texture combiners.

This commit is contained in:
Tony Wasserka 2014-08-15 16:33:17 +02:00
parent c4691b784b
commit 27cab6477e
2 changed files with 225 additions and 2 deletions

View file

@ -4,6 +4,7 @@
#pragma once #pragma once
#include <array>
#include <cstddef> #include <cstddef>
#include <initializer_list> #include <initializer_list>
#include <map> #include <map>
@ -133,7 +134,97 @@ struct Regs {
INSERT_PADDING_WORDS(0x8); INSERT_PADDING_WORDS(0x8);
BitField<0, 4, TextureFormat> texture0_format; BitField<0, 4, TextureFormat> texture0_format;
INSERT_PADDING_WORDS(0x81); INSERT_PADDING_WORDS(0x31);
// 0xc0-0xff: Texture Combiner (akin to glTexEnv)
struct TevStageConfig {
enum class Source : u32 {
PrimaryColor = 0x0,
Texture0 = 0x3,
Texture1 = 0x4,
Texture2 = 0x5,
Texture3 = 0x6,
// 0x7-0xc = primary color??
Constant = 0xe,
Previous = 0xf,
};
enum class ColorModifier : u32 {
SourceColor = 0,
OneMinusSourceColor = 1,
SourceAlpha = 2,
OneMinusSourceAlpha = 3,
// Other values seem to be non-standard extensions
};
enum class AlphaModifier : u32 {
SourceAlpha = 0,
OneMinusSourceAlpha = 1,
// Other values seem to be non-standard extensions
};
enum class Operation : u32 {
Replace = 0,
Modulate = 1,
Add = 2,
AddSigned = 3,
Lerp = 4,
Subtract = 5,
};
union {
BitField< 0, 4, Source> color_source1;
BitField< 4, 4, Source> color_source2;
BitField< 8, 4, Source> color_source3;
BitField<16, 4, Source> alpha_source1;
BitField<20, 4, Source> alpha_source2;
BitField<24, 4, Source> alpha_source3;
};
union {
BitField< 0, 4, ColorModifier> color_modifier1;
BitField< 4, 4, ColorModifier> color_modifier2;
BitField< 8, 4, ColorModifier> color_modifier3;
BitField<12, 3, AlphaModifier> alpha_modifier1;
BitField<16, 3, AlphaModifier> alpha_modifier2;
BitField<20, 3, AlphaModifier> alpha_modifier3;
};
union {
BitField< 0, 4, Operation> color_op;
BitField<16, 4, Operation> alpha_op;
};
union {
BitField< 0, 8, u32> const_r;
BitField< 8, 8, u32> const_g;
BitField<16, 8, u32> const_b;
BitField<24, 8, u32> const_a;
};
INSERT_PADDING_WORDS(0x1);
};
TevStageConfig tev_stage0;
INSERT_PADDING_WORDS(0x3);
TevStageConfig tev_stage1;
INSERT_PADDING_WORDS(0x3);
TevStageConfig tev_stage2;
INSERT_PADDING_WORDS(0x3);
TevStageConfig tev_stage3;
INSERT_PADDING_WORDS(0x13);
TevStageConfig tev_stage4;
INSERT_PADDING_WORDS(0x3);
TevStageConfig tev_stage5;
INSERT_PADDING_WORDS(0x13);
const std::array<Regs::TevStageConfig,6> GetTevStages() const {
return { tev_stage0, tev_stage1,
tev_stage2, tev_stage3,
tev_stage4, tev_stage5 };
};
struct { struct {
enum ColorFormat : u32 { enum ColorFormat : u32 {
@ -444,6 +535,12 @@ struct Regs {
ADD_FIELD(viewport_corner); ADD_FIELD(viewport_corner);
ADD_FIELD(texture0); ADD_FIELD(texture0);
ADD_FIELD(texture0_format); ADD_FIELD(texture0_format);
ADD_FIELD(tev_stage0);
ADD_FIELD(tev_stage1);
ADD_FIELD(tev_stage2);
ADD_FIELD(tev_stage3);
ADD_FIELD(tev_stage4);
ADD_FIELD(tev_stage5);
ADD_FIELD(framebuffer); ADD_FIELD(framebuffer);
ADD_FIELD(vertex_attributes); ADD_FIELD(vertex_attributes);
ADD_FIELD(index_array); ADD_FIELD(index_array);
@ -503,6 +600,12 @@ ASSERT_REG_POSITION(vs_output_attributes[1], 0x51);
ASSERT_REG_POSITION(viewport_corner, 0x68); ASSERT_REG_POSITION(viewport_corner, 0x68);
ASSERT_REG_POSITION(texture0, 0x81); ASSERT_REG_POSITION(texture0, 0x81);
ASSERT_REG_POSITION(texture0_format, 0x8e); ASSERT_REG_POSITION(texture0_format, 0x8e);
ASSERT_REG_POSITION(tev_stage0, 0xc0);
ASSERT_REG_POSITION(tev_stage1, 0xc8);
ASSERT_REG_POSITION(tev_stage2, 0xd0);
ASSERT_REG_POSITION(tev_stage3, 0xd8);
ASSERT_REG_POSITION(tev_stage4, 0xf0);
ASSERT_REG_POSITION(tev_stage5, 0xf8);
ASSERT_REG_POSITION(framebuffer, 0x110); ASSERT_REG_POSITION(framebuffer, 0x110);
ASSERT_REG_POSITION(vertex_attributes, 0x200); ASSERT_REG_POSITION(vertex_attributes, 0x200);
ASSERT_REG_POSITION(index_array, 0x227); ASSERT_REG_POSITION(index_array, 0x227);

View file

@ -165,12 +165,132 @@ void ProcessTriangle(const VertexShader::OutputVertex& v0,
(u8)(GetInterpolatedAttribute(v0.color.a(), v1.color.a(), v2.color.a()).ToFloat32() * 255) (u8)(GetInterpolatedAttribute(v0.color.a(), v1.color.a(), v2.color.a()).ToFloat32() * 255)
}; };
// Texture environment - consists of 6 stages of color and alpha combining.
//
// Color combiners take three input color values from some source (e.g. interpolated
// vertex color, texture color, previous stage, etc), perform some very simple
// operations on each of them (e.g. inversion) and then calculate the output color
// with some basic arithmetic. Alpha combiners can be configured separately but work
// analogously.
Math::Vec4<u8> combiner_output;
for (auto tev_stage : registers.GetTevStages()) {
using Source = Regs::TevStageConfig::Source;
using ColorModifier = Regs::TevStageConfig::ColorModifier;
using AlphaModifier = Regs::TevStageConfig::AlphaModifier;
using Operation = Regs::TevStageConfig::Operation;
auto GetColorSource = [&](Source source) -> Math::Vec3<u8> {
switch (source) {
case Source::PrimaryColor:
return primary_color.rgb();
case Source::Constant:
return {tev_stage.const_r, tev_stage.const_g, tev_stage.const_b};
case Source::Previous:
return combiner_output.rgb();
default:
ERROR_LOG(GPU, "Unknown color combiner source %d\n", (int)source);
return {};
}
};
auto GetAlphaSource = [&](Source source) -> u8 {
switch (source) {
case Source::PrimaryColor:
return primary_color.a();
case Source::Constant:
return tev_stage.const_a;
case Source::Previous:
return combiner_output.a();
default:
ERROR_LOG(GPU, "Unknown alpha combiner source %d\n", (int)source);
return 0;
}
};
auto GetColorModifier = [](ColorModifier factor, const Math::Vec3<u8>& values) -> Math::Vec3<u8> {
switch (factor)
{
case ColorModifier::SourceColor:
return values;
default:
ERROR_LOG(GPU, "Unknown color factor %d\n", (int)factor);
return {};
}
};
auto GetAlphaModifier = [](AlphaModifier factor, u8 value) -> u8 {
switch (factor) {
case AlphaModifier::SourceAlpha:
return value;
default:
ERROR_LOG(GPU, "Unknown color factor %d\n", (int)factor);
return 0;
}
};
auto ColorCombine = [](Operation op, const Math::Vec3<u8> input[3]) -> Math::Vec3<u8> {
switch (op) {
case Operation::Replace:
return input[0];
case Operation::Modulate:
return ((input[0] * input[1]) / 255).Cast<u8>();
default:
ERROR_LOG(GPU, "Unknown color combiner operation %d\n", (int)op);
return {};
}
};
auto AlphaCombine = [](Operation op, const std::array<u8,3>& input) -> u8 {
switch (op) {
case Operation::Replace:
return input[0];
case Operation::Modulate:
return input[0] * input[1] / 255;
default:
ERROR_LOG(GPU, "Unknown alpha combiner operation %d\n", (int)op);
return 0;
}
};
// color combiner
// NOTE: Not sure if the alpha combiner might use the color output of the previous
// stage as input. Hence, we currently don't directly write the result to
// combiner_output.rgb(), but instead store it in a temporary variable until
// alpha combining has been done.
Math::Vec3<u8> color_result[3] = {
GetColorModifier(tev_stage.color_modifier1, GetColorSource(tev_stage.color_source1)),
GetColorModifier(tev_stage.color_modifier2, GetColorSource(tev_stage.color_source2)),
GetColorModifier(tev_stage.color_modifier3, GetColorSource(tev_stage.color_source3))
};
auto color_output = ColorCombine(tev_stage.color_op, color_result);
// alpha combiner
std::array<u8,3> alpha_result = {
GetAlphaModifier(tev_stage.alpha_modifier1, GetAlphaSource(tev_stage.alpha_source1)),
GetAlphaModifier(tev_stage.alpha_modifier2, GetAlphaSource(tev_stage.alpha_source2)),
GetAlphaModifier(tev_stage.alpha_modifier3, GetAlphaSource(tev_stage.alpha_source3))
};
auto alpha_output = AlphaCombine(tev_stage.alpha_op, alpha_result);
combiner_output = Math::MakeVec(color_output, alpha_output);
}
u16 z = (u16)(((float)v0.screenpos[2].ToFloat32() * w0 + u16 z = (u16)(((float)v0.screenpos[2].ToFloat32() * w0 +
(float)v1.screenpos[2].ToFloat32() * w1 + (float)v1.screenpos[2].ToFloat32() * w1 +
(float)v2.screenpos[2].ToFloat32() * w2) * 65535.f / wsum); // TODO: Shouldn't need to multiply by 65536? (float)v2.screenpos[2].ToFloat32() * w2) * 65535.f / wsum); // TODO: Shouldn't need to multiply by 65536?
SetDepth(x >> 4, y >> 4, z); SetDepth(x >> 4, y >> 4, z);
DrawPixel(x >> 4, y >> 4, primary_color); DrawPixel(x >> 4, y >> 4, combiner_output);
} }
} }
} }