Merge pull request #2792 from wwylele/lutlutlut

gl_rasterizer: fix lighting LUT interpolation
This commit is contained in:
Yuri Kunde Schlesner 2017-06-21 14:55:23 -07:00 committed by GitHub
commit 72b69cea4b
7 changed files with 175 additions and 151 deletions

View file

@ -87,12 +87,18 @@ struct State {
// LUT value, encoded as 12-bit fixed point, with 12 fraction bits // LUT value, encoded as 12-bit fixed point, with 12 fraction bits
BitField<0, 12, u32> value; // 0.0.12 fixed point BitField<0, 12, u32> value; // 0.0.12 fixed point
// Used by HW for efficient interpolation, Citra does not use these // Used for efficient interpolation.
BitField<12, 12, s32> difference; // 1.0.11 fixed point BitField<12, 11, u32> difference; // 0.0.11 fixed point
BitField<23, 1, u32> neg_difference;
float ToFloat() { float ToFloat() const {
return static_cast<float>(value) / 4095.f; return static_cast<float>(value) / 4095.f;
} }
float DiffToFloat() const {
float diff = static_cast<float>(difference) / 2047.f;
return neg_difference ? -diff : diff;
}
}; };
std::array<std::array<LutEntry, 256>, 24> luts; std::array<std::array<LutEntry, 256>, 24> luts;

View file

@ -26,6 +26,8 @@ struct LightingRegs {
DistanceAttenuation = 16, DistanceAttenuation = 16,
}; };
static constexpr unsigned NumLightingSampler = 24;
static LightingSampler SpotlightAttenuationSampler(unsigned index) { static LightingSampler SpotlightAttenuationSampler(unsigned index) {
return static_cast<LightingSampler>( return static_cast<LightingSampler>(
static_cast<unsigned>(LightingSampler::SpotlightAttenuation) + index); static_cast<unsigned>(LightingSampler::SpotlightAttenuation) + index);

View file

@ -49,9 +49,7 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
uniform_block_data.dirty = true; uniform_block_data.dirty = true;
for (unsigned index = 0; index < lighting_luts.size(); index++) { uniform_block_data.lut_dirty.fill(true);
uniform_block_data.lut_dirty[index] = true;
}
uniform_block_data.fog_lut_dirty = true; uniform_block_data.fog_lut_dirty = true;
@ -96,18 +94,16 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
framebuffer.Create(); framebuffer.Create();
// Allocate and bind lighting lut textures // Allocate and bind lighting lut textures
for (size_t i = 0; i < lighting_luts.size(); ++i) { lighting_lut_buffer.Create();
lighting_luts[i].Create(); state.lighting_lut.texture_buffer = lighting_lut.handle;
state.lighting_luts[i].texture_1d = lighting_luts[i].handle;
}
state.Apply(); state.Apply();
lighting_lut.Create();
for (size_t i = 0; i < lighting_luts.size(); ++i) { glBindBuffer(GL_TEXTURE_BUFFER, lighting_lut_buffer.handle);
glActiveTexture(static_cast<GLenum>(GL_TEXTURE3 + i)); glBufferData(GL_TEXTURE_BUFFER,
glTexImage1D(GL_TEXTURE_1D, 0, GL_RGBA32F, 256, 0, GL_RGBA, GL_FLOAT, nullptr); sizeof(GLfloat) * 2 * 256 * Pica::LightingRegs::NumLightingSampler, nullptr,
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_LINEAR); GL_DYNAMIC_DRAW);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_LINEAR); glActiveTexture(TextureUnits::LightingLUT.Enum());
} glTexBuffer(GL_TEXTURE_BUFFER, GL_RG32F, lighting_lut_buffer.handle);
// Setup the LUT for the fog // Setup the LUT for the fog
{ {
@ -116,7 +112,7 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
} }
state.Apply(); state.Apply();
glActiveTexture(GL_TEXTURE9); glActiveTexture(TextureUnits::FogLUT.Enum());
glTexImage1D(GL_TEXTURE_1D, 0, GL_R32UI, 128, 0, GL_RED_INTEGER, GL_UNSIGNED_INT, nullptr); glTexImage1D(GL_TEXTURE_1D, 0, GL_R32UI, 128, 0, GL_RED_INTEGER, GL_UNSIGNED_INT, nullptr);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
@ -125,7 +121,7 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
proctex_noise_lut.Create(); proctex_noise_lut.Create();
state.proctex_noise_lut.texture_1d = proctex_noise_lut.handle; state.proctex_noise_lut.texture_1d = proctex_noise_lut.handle;
state.Apply(); state.Apply();
glActiveTexture(GL_TEXTURE10); glActiveTexture(TextureUnits::ProcTexNoiseLUT.Enum());
glTexImage1D(GL_TEXTURE_1D, 0, GL_RG32F, 128, 0, GL_RG, GL_FLOAT, nullptr); glTexImage1D(GL_TEXTURE_1D, 0, GL_RG32F, 128, 0, GL_RG, GL_FLOAT, nullptr);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
@ -134,7 +130,7 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
proctex_color_map.Create(); proctex_color_map.Create();
state.proctex_color_map.texture_1d = proctex_color_map.handle; state.proctex_color_map.texture_1d = proctex_color_map.handle;
state.Apply(); state.Apply();
glActiveTexture(GL_TEXTURE11); glActiveTexture(TextureUnits::ProcTexColorMap.Enum());
glTexImage1D(GL_TEXTURE_1D, 0, GL_RG32F, 128, 0, GL_RG, GL_FLOAT, nullptr); glTexImage1D(GL_TEXTURE_1D, 0, GL_RG32F, 128, 0, GL_RG, GL_FLOAT, nullptr);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
@ -143,7 +139,7 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
proctex_alpha_map.Create(); proctex_alpha_map.Create();
state.proctex_alpha_map.texture_1d = proctex_alpha_map.handle; state.proctex_alpha_map.texture_1d = proctex_alpha_map.handle;
state.Apply(); state.Apply();
glActiveTexture(GL_TEXTURE12); glActiveTexture(TextureUnits::ProcTexAlphaMap.Enum());
glTexImage1D(GL_TEXTURE_1D, 0, GL_RG32F, 128, 0, GL_RG, GL_FLOAT, nullptr); glTexImage1D(GL_TEXTURE_1D, 0, GL_RG32F, 128, 0, GL_RG, GL_FLOAT, nullptr);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
@ -152,7 +148,7 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
proctex_lut.Create(); proctex_lut.Create();
state.proctex_lut.texture_1d = proctex_lut.handle; state.proctex_lut.texture_1d = proctex_lut.handle;
state.Apply(); state.Apply();
glActiveTexture(GL_TEXTURE13); glActiveTexture(TextureUnits::ProcTexLUT.Enum());
glTexImage1D(GL_TEXTURE_1D, 0, GL_RGBA32F, 256, 0, GL_RGBA, GL_FLOAT, nullptr); glTexImage1D(GL_TEXTURE_1D, 0, GL_RGBA32F, 256, 0, GL_RGBA, GL_FLOAT, nullptr);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
@ -161,7 +157,7 @@ RasterizerOpenGL::RasterizerOpenGL() : shader_dirty(true) {
proctex_diff_lut.Create(); proctex_diff_lut.Create();
state.proctex_diff_lut.texture_1d = proctex_diff_lut.handle; state.proctex_diff_lut.texture_1d = proctex_diff_lut.handle;
state.Apply(); state.Apply();
glActiveTexture(GL_TEXTURE14); glActiveTexture(TextureUnits::ProcTexDiffLUT.Enum());
glTexImage1D(GL_TEXTURE_1D, 0, GL_RGBA32F, 256, 0, GL_RGBA, GL_FLOAT, nullptr); glTexImage1D(GL_TEXTURE_1D, 0, GL_RGBA32F, 256, 0, GL_RGBA, GL_FLOAT, nullptr);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
@ -313,7 +309,7 @@ void RasterizerOpenGL::DrawTriangles() {
} }
// Sync the lighting luts // Sync the lighting luts
for (unsigned index = 0; index < lighting_luts.size(); index++) { for (unsigned index = 0; index < uniform_block_data.lut_dirty.size(); index++) {
if (uniform_block_data.lut_dirty[index]) { if (uniform_block_data.lut_dirty[index]) {
SyncLightingLUT(index); SyncLightingLUT(index);
uniform_block_data.lut_dirty[index] = false; uniform_block_data.lut_dirty[index] = false;
@ -851,7 +847,7 @@ void RasterizerOpenGL::NotifyPicaRegisterChanged(u32 id) {
case PICA_REG_INDEX_WORKAROUND(lighting.lut_data[6], 0x1ce): case PICA_REG_INDEX_WORKAROUND(lighting.lut_data[6], 0x1ce):
case PICA_REG_INDEX_WORKAROUND(lighting.lut_data[7], 0x1cf): { case PICA_REG_INDEX_WORKAROUND(lighting.lut_data[7], 0x1cf): {
auto& lut_config = regs.lighting.lut_config; auto& lut_config = regs.lighting.lut_config;
uniform_block_data.lut_dirty[lut_config.type / 4] = true; uniform_block_data.lut_dirty[lut_config.type] = true;
break; break;
} }
} }
@ -1187,77 +1183,57 @@ void RasterizerOpenGL::SetShader() {
state.Apply(); state.Apply();
// Set the texture samplers to correspond to different texture units // Set the texture samplers to correspond to different texture units
GLuint uniform_tex = glGetUniformLocation(shader->shader.handle, "tex[0]"); GLint uniform_tex = glGetUniformLocation(shader->shader.handle, "tex[0]");
if (uniform_tex != -1) { if (uniform_tex != -1) {
glUniform1i(uniform_tex, 0); glUniform1i(uniform_tex, TextureUnits::PicaTexture(0).id);
} }
uniform_tex = glGetUniformLocation(shader->shader.handle, "tex[1]"); uniform_tex = glGetUniformLocation(shader->shader.handle, "tex[1]");
if (uniform_tex != -1) { if (uniform_tex != -1) {
glUniform1i(uniform_tex, 1); glUniform1i(uniform_tex, TextureUnits::PicaTexture(1).id);
} }
uniform_tex = glGetUniformLocation(shader->shader.handle, "tex[2]"); uniform_tex = glGetUniformLocation(shader->shader.handle, "tex[2]");
if (uniform_tex != -1) { if (uniform_tex != -1) {
glUniform1i(uniform_tex, 2); glUniform1i(uniform_tex, TextureUnits::PicaTexture(2).id);
} }
// Set the texture samplers to correspond to different lookup table texture units // Set the texture samplers to correspond to different lookup table texture units
GLuint uniform_lut = glGetUniformLocation(shader->shader.handle, "lut[0]"); GLint uniform_lut = glGetUniformLocation(shader->shader.handle, "lighting_lut");
if (uniform_lut != -1) { if (uniform_lut != -1) {
glUniform1i(uniform_lut, 3); glUniform1i(uniform_lut, TextureUnits::LightingLUT.id);
}
uniform_lut = glGetUniformLocation(shader->shader.handle, "lut[1]");
if (uniform_lut != -1) {
glUniform1i(uniform_lut, 4);
}
uniform_lut = glGetUniformLocation(shader->shader.handle, "lut[2]");
if (uniform_lut != -1) {
glUniform1i(uniform_lut, 5);
}
uniform_lut = glGetUniformLocation(shader->shader.handle, "lut[3]");
if (uniform_lut != -1) {
glUniform1i(uniform_lut, 6);
}
uniform_lut = glGetUniformLocation(shader->shader.handle, "lut[4]");
if (uniform_lut != -1) {
glUniform1i(uniform_lut, 7);
}
uniform_lut = glGetUniformLocation(shader->shader.handle, "lut[5]");
if (uniform_lut != -1) {
glUniform1i(uniform_lut, 8);
} }
GLuint uniform_fog_lut = glGetUniformLocation(shader->shader.handle, "fog_lut"); GLint uniform_fog_lut = glGetUniformLocation(shader->shader.handle, "fog_lut");
if (uniform_fog_lut != -1) { if (uniform_fog_lut != -1) {
glUniform1i(uniform_fog_lut, 9); glUniform1i(uniform_fog_lut, TextureUnits::FogLUT.id);
} }
GLuint uniform_proctex_noise_lut = GLint uniform_proctex_noise_lut =
glGetUniformLocation(shader->shader.handle, "proctex_noise_lut"); glGetUniformLocation(shader->shader.handle, "proctex_noise_lut");
if (uniform_proctex_noise_lut != -1) { if (uniform_proctex_noise_lut != -1) {
glUniform1i(uniform_proctex_noise_lut, 10); glUniform1i(uniform_proctex_noise_lut, TextureUnits::ProcTexNoiseLUT.id);
} }
GLuint uniform_proctex_color_map = GLint uniform_proctex_color_map =
glGetUniformLocation(shader->shader.handle, "proctex_color_map"); glGetUniformLocation(shader->shader.handle, "proctex_color_map");
if (uniform_proctex_color_map != -1) { if (uniform_proctex_color_map != -1) {
glUniform1i(uniform_proctex_color_map, 11); glUniform1i(uniform_proctex_color_map, TextureUnits::ProcTexColorMap.id);
} }
GLuint uniform_proctex_alpha_map = GLint uniform_proctex_alpha_map =
glGetUniformLocation(shader->shader.handle, "proctex_alpha_map"); glGetUniformLocation(shader->shader.handle, "proctex_alpha_map");
if (uniform_proctex_alpha_map != -1) { if (uniform_proctex_alpha_map != -1) {
glUniform1i(uniform_proctex_alpha_map, 12); glUniform1i(uniform_proctex_alpha_map, TextureUnits::ProcTexAlphaMap.id);
} }
GLuint uniform_proctex_lut = glGetUniformLocation(shader->shader.handle, "proctex_lut"); GLint uniform_proctex_lut = glGetUniformLocation(shader->shader.handle, "proctex_lut");
if (uniform_proctex_lut != -1) { if (uniform_proctex_lut != -1) {
glUniform1i(uniform_proctex_lut, 13); glUniform1i(uniform_proctex_lut, TextureUnits::ProcTexLUT.id);
} }
GLuint uniform_proctex_diff_lut = GLint uniform_proctex_diff_lut =
glGetUniformLocation(shader->shader.handle, "proctex_diff_lut"); glGetUniformLocation(shader->shader.handle, "proctex_diff_lut");
if (uniform_proctex_diff_lut != -1) { if (uniform_proctex_diff_lut != -1) {
glUniform1i(uniform_proctex_diff_lut, 14); glUniform1i(uniform_proctex_diff_lut, TextureUnits::ProcTexDiffLUT.id);
} }
current_shader = shader_cache.emplace(config, std::move(shader)).first->second.get(); current_shader = shader_cache.emplace(config, std::move(shader)).first->second.get();
@ -1387,7 +1363,7 @@ void RasterizerOpenGL::SyncFogLUT() {
if (new_data != fog_lut_data) { if (new_data != fog_lut_data) {
fog_lut_data = new_data; fog_lut_data = new_data;
glActiveTexture(GL_TEXTURE9); glActiveTexture(TextureUnits::FogLUT.Enum());
glTexSubImage1D(GL_TEXTURE_1D, 0, 0, 128, GL_RED_INTEGER, GL_UNSIGNED_INT, glTexSubImage1D(GL_TEXTURE_1D, 0, 0, 128, GL_RED_INTEGER, GL_UNSIGNED_INT,
fog_lut_data.data()); fog_lut_data.data());
} }
@ -1426,17 +1402,18 @@ static void SyncProcTexValueLUT(const std::array<Pica::State::ProcTex::ValueEntr
} }
void RasterizerOpenGL::SyncProcTexNoiseLUT() { void RasterizerOpenGL::SyncProcTexNoiseLUT() {
SyncProcTexValueLUT(Pica::g_state.proctex.noise_table, proctex_noise_lut_data, GL_TEXTURE10); SyncProcTexValueLUT(Pica::g_state.proctex.noise_table, proctex_noise_lut_data,
TextureUnits::ProcTexNoiseLUT.Enum());
} }
void RasterizerOpenGL::SyncProcTexColorMap() { void RasterizerOpenGL::SyncProcTexColorMap() {
SyncProcTexValueLUT(Pica::g_state.proctex.color_map_table, proctex_color_map_data, SyncProcTexValueLUT(Pica::g_state.proctex.color_map_table, proctex_color_map_data,
GL_TEXTURE11); TextureUnits::ProcTexColorMap.Enum());
} }
void RasterizerOpenGL::SyncProcTexAlphaMap() { void RasterizerOpenGL::SyncProcTexAlphaMap() {
SyncProcTexValueLUT(Pica::g_state.proctex.alpha_map_table, proctex_alpha_map_data, SyncProcTexValueLUT(Pica::g_state.proctex.alpha_map_table, proctex_alpha_map_data,
GL_TEXTURE12); TextureUnits::ProcTexAlphaMap.Enum());
} }
void RasterizerOpenGL::SyncProcTexLUT() { void RasterizerOpenGL::SyncProcTexLUT() {
@ -1451,7 +1428,7 @@ void RasterizerOpenGL::SyncProcTexLUT() {
if (new_data != proctex_lut_data) { if (new_data != proctex_lut_data) {
proctex_lut_data = new_data; proctex_lut_data = new_data;
glActiveTexture(GL_TEXTURE13); glActiveTexture(TextureUnits::ProcTexLUT.Enum());
glTexSubImage1D(GL_TEXTURE_1D, 0, 0, 256, GL_RGBA, GL_FLOAT, proctex_lut_data.data()); glTexSubImage1D(GL_TEXTURE_1D, 0, 0, 256, GL_RGBA, GL_FLOAT, proctex_lut_data.data());
} }
} }
@ -1468,7 +1445,7 @@ void RasterizerOpenGL::SyncProcTexDiffLUT() {
if (new_data != proctex_diff_lut_data) { if (new_data != proctex_diff_lut_data) {
proctex_diff_lut_data = new_data; proctex_diff_lut_data = new_data;
glActiveTexture(GL_TEXTURE14); glActiveTexture(TextureUnits::ProcTexDiffLUT.Enum());
glTexSubImage1D(GL_TEXTURE_1D, 0, 0, 256, GL_RGBA, GL_FLOAT, proctex_diff_lut_data.data()); glTexSubImage1D(GL_TEXTURE_1D, 0, 0, 256, GL_RGBA, GL_FLOAT, proctex_diff_lut_data.data());
} }
} }
@ -1571,20 +1548,17 @@ void RasterizerOpenGL::SyncGlobalAmbient() {
} }
void RasterizerOpenGL::SyncLightingLUT(unsigned lut_index) { void RasterizerOpenGL::SyncLightingLUT(unsigned lut_index) {
std::array<GLvec4, 256> new_data; std::array<GLvec2, 256> new_data;
const auto& source_lut = Pica::g_state.lighting.luts[lut_index];
for (unsigned offset = 0; offset < new_data.size(); ++offset) { std::transform(source_lut.begin(), source_lut.end(), new_data.begin(), [](const auto& entry) {
new_data[offset][0] = Pica::g_state.lighting.luts[(lut_index * 4) + 0][offset].ToFloat(); return GLvec2{entry.ToFloat(), entry.DiffToFloat()};
new_data[offset][1] = Pica::g_state.lighting.luts[(lut_index * 4) + 1][offset].ToFloat(); });
new_data[offset][2] = Pica::g_state.lighting.luts[(lut_index * 4) + 2][offset].ToFloat();
new_data[offset][3] = Pica::g_state.lighting.luts[(lut_index * 4) + 3][offset].ToFloat();
}
if (new_data != lighting_lut_data[lut_index]) { if (new_data != lighting_lut_data[lut_index]) {
lighting_lut_data[lut_index] = new_data; lighting_lut_data[lut_index] = new_data;
glActiveTexture(GL_TEXTURE3 + lut_index); glBindBuffer(GL_TEXTURE_BUFFER, lighting_lut_buffer.handle);
glTexSubImage1D(GL_TEXTURE_1D, 0, 0, 256, GL_RGBA, GL_FLOAT, glBufferSubData(GL_TEXTURE_BUFFER, lut_index * new_data.size() * sizeof(GLvec2),
lighting_lut_data[lut_index].data()); new_data.size() * sizeof(GLvec2), new_data.data());
} }
} }

View file

@ -263,7 +263,7 @@ private:
struct { struct {
UniformData data; UniformData data;
bool lut_dirty[6]; std::array<bool, Pica::LightingRegs::NumLightingSampler> lut_dirty;
bool fog_lut_dirty; bool fog_lut_dirty;
bool proctex_noise_lut_dirty; bool proctex_noise_lut_dirty;
bool proctex_color_map_dirty; bool proctex_color_map_dirty;
@ -279,8 +279,9 @@ private:
OGLBuffer uniform_buffer; OGLBuffer uniform_buffer;
OGLFramebuffer framebuffer; OGLFramebuffer framebuffer;
std::array<OGLTexture, 6> lighting_luts; OGLBuffer lighting_lut_buffer;
std::array<std::array<GLvec4, 256>, 6> lighting_lut_data{}; OGLTexture lighting_lut;
std::array<std::array<GLvec2, 256>, Pica::LightingRegs::NumLightingSampler> lighting_lut_data{};
OGLTexture fog_lut; OGLTexture fog_lut;
std::array<GLuint, 128> fog_lut_data{}; std::array<GLuint, 128> fog_lut_data{};

View file

@ -562,9 +562,9 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
out += "vec3 normal = quaternion_rotate(normalized_normquat, surface_normal);\n"; out += "vec3 normal = quaternion_rotate(normalized_normquat, surface_normal);\n";
out += "vec3 tangent = quaternion_rotate(normalized_normquat, surface_tangent);\n"; out += "vec3 tangent = quaternion_rotate(normalized_normquat, surface_tangent);\n";
// Gets the index into the specified lookup table for specular lighting // Samples the specified lookup table for specular lighting
auto GetLutIndex = [&lighting](unsigned light_num, LightingRegs::LightingLutInput input, auto GetLutValue = [&lighting](LightingRegs::LightingSampler sampler, unsigned light_num,
bool abs) { LightingRegs::LightingLutInput input, bool abs) {
std::string index; std::string index;
switch (input) { switch (input) {
case LightingRegs::LightingLutInput::NH: case LightingRegs::LightingLutInput::NH:
@ -610,22 +610,18 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
break; break;
} }
std::string sampler_string = std::to_string(static_cast<unsigned>(sampler));
if (abs) { if (abs) {
// LUT index is in the range of (0.0, 1.0) // LUT index is in the range of (0.0, 1.0)
index = lighting.light[light_num].two_sided_diffuse ? "abs(" + index + ")" index = lighting.light[light_num].two_sided_diffuse ? "abs(" + index + ")"
: "max(" + index + ", 0.0)"; : "max(" + index + ", 0.0)";
return "LookupLightingLUTUnsigned(" + sampler_string + ", " + index + ")";
} else { } else {
// LUT index is in the range of (-1.0, 1.0) // LUT index is in the range of (-1.0, 1.0)
index = "((" + index + " < 0) ? " + index + " + 2.0 : " + index + ") / 2.0"; return "LookupLightingLUTSigned(" + sampler_string + ", " + index + ")";
} }
return "(OFFSET_256 + SCALE_256 * clamp(" + index + ", 0.0, 1.0))";
};
// Gets the lighting lookup table value given the specified sampler and index
auto GetLutValue = [](LightingRegs::LightingSampler sampler, std::string lut_index) {
return std::string("texture(lut[" + std::to_string((unsigned)sampler / 4) + "], " +
lut_index + ")[" + std::to_string((unsigned)sampler & 3) + "]");
}; };
// Write the code to emulate each enabled light // Write the code to emulate each enabled light
@ -653,21 +649,21 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
if (light_config.spot_atten_enable && if (light_config.spot_atten_enable &&
LightingRegs::IsLightingSamplerSupported( LightingRegs::IsLightingSamplerSupported(
lighting.config, LightingRegs::LightingSampler::SpotlightAttenuation)) { lighting.config, LightingRegs::LightingSampler::SpotlightAttenuation)) {
std::string index = std::string value =
GetLutIndex(light_config.num, lighting.lut_sp.type, lighting.lut_sp.abs_input); GetLutValue(LightingRegs::SpotlightAttenuationSampler(light_config.num),
auto sampler = LightingRegs::SpotlightAttenuationSampler(light_config.num); light_config.num, lighting.lut_sp.type, lighting.lut_sp.abs_input);
spot_atten = "(" + std::to_string(lighting.lut_sp.scale) + " * " + spot_atten = "(" + std::to_string(lighting.lut_sp.scale) + " * " + value + ")";
GetLutValue(sampler, index) + ")";
} }
// If enabled, compute distance attenuation value // If enabled, compute distance attenuation value
std::string dist_atten = "1.0"; std::string dist_atten = "1.0";
if (light_config.dist_atten_enable) { if (light_config.dist_atten_enable) {
std::string index = "(" + light_src + ".dist_atten_scale * length(-view - " + std::string index = "clamp(" + light_src + ".dist_atten_scale * length(-view - " +
light_src + ".position) + " + light_src + ".dist_atten_bias)"; light_src + ".position) + " + light_src +
index = "(OFFSET_256 + SCALE_256 * clamp(" + index + ", 0.0, 1.0))"; ".dist_atten_bias, 0.0, 1.0)";
auto sampler = LightingRegs::DistanceAttenuationSampler(light_config.num); auto sampler = LightingRegs::DistanceAttenuationSampler(light_config.num);
dist_atten = GetLutValue(sampler, index); dist_atten = "LookupLightingLUTUnsigned(" +
std::to_string(static_cast<unsigned>(sampler)) + "," + index + ")";
} }
// If enabled, clamp specular component if lighting result is negative // If enabled, clamp specular component if lighting result is negative
@ -686,10 +682,10 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
LightingRegs::IsLightingSamplerSupported( LightingRegs::IsLightingSamplerSupported(
lighting.config, LightingRegs::LightingSampler::Distribution0)) { lighting.config, LightingRegs::LightingSampler::Distribution0)) {
// Lookup specular "distribution 0" LUT value // Lookup specular "distribution 0" LUT value
std::string index = std::string value =
GetLutIndex(light_config.num, lighting.lut_d0.type, lighting.lut_d0.abs_input); GetLutValue(LightingRegs::LightingSampler::Distribution0, light_config.num,
d0_lut_value = "(" + std::to_string(lighting.lut_d0.scale) + " * " + lighting.lut_d0.type, lighting.lut_d0.abs_input);
GetLutValue(LightingRegs::LightingSampler::Distribution0, index) + ")"; d0_lut_value = "(" + std::to_string(lighting.lut_d0.scale) + " * " + value + ")";
} }
std::string specular_0 = "(" + d0_lut_value + " * " + light_src + ".specular_0)"; std::string specular_0 = "(" + d0_lut_value + " * " + light_src + ".specular_0)";
if (light_config.geometric_factor_0) { if (light_config.geometric_factor_0) {
@ -700,10 +696,10 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
if (lighting.lut_rr.enable && if (lighting.lut_rr.enable &&
LightingRegs::IsLightingSamplerSupported(lighting.config, LightingRegs::IsLightingSamplerSupported(lighting.config,
LightingRegs::LightingSampler::ReflectRed)) { LightingRegs::LightingSampler::ReflectRed)) {
std::string index = std::string value =
GetLutIndex(light_config.num, lighting.lut_rr.type, lighting.lut_rr.abs_input); GetLutValue(LightingRegs::LightingSampler::ReflectRed, light_config.num,
std::string value = "(" + std::to_string(lighting.lut_rr.scale) + " * " + lighting.lut_rr.type, lighting.lut_rr.abs_input);
GetLutValue(LightingRegs::LightingSampler::ReflectRed, index) + ")"; value = "(" + std::to_string(lighting.lut_rr.scale) + " * " + value + ")";
out += "refl_value.r = " + value + ";\n"; out += "refl_value.r = " + value + ";\n";
} else { } else {
out += "refl_value.r = 1.0;\n"; out += "refl_value.r = 1.0;\n";
@ -713,11 +709,10 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
if (lighting.lut_rg.enable && if (lighting.lut_rg.enable &&
LightingRegs::IsLightingSamplerSupported(lighting.config, LightingRegs::IsLightingSamplerSupported(lighting.config,
LightingRegs::LightingSampler::ReflectGreen)) { LightingRegs::LightingSampler::ReflectGreen)) {
std::string index = std::string value =
GetLutIndex(light_config.num, lighting.lut_rg.type, lighting.lut_rg.abs_input); GetLutValue(LightingRegs::LightingSampler::ReflectGreen, light_config.num,
std::string value = "(" + std::to_string(lighting.lut_rg.scale) + " * " + lighting.lut_rg.type, lighting.lut_rg.abs_input);
GetLutValue(LightingRegs::LightingSampler::ReflectGreen, index) + value = "(" + std::to_string(lighting.lut_rg.scale) + " * " + value + ")";
")";
out += "refl_value.g = " + value + ";\n"; out += "refl_value.g = " + value + ";\n";
} else { } else {
out += "refl_value.g = refl_value.r;\n"; out += "refl_value.g = refl_value.r;\n";
@ -727,11 +722,10 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
if (lighting.lut_rb.enable && if (lighting.lut_rb.enable &&
LightingRegs::IsLightingSamplerSupported(lighting.config, LightingRegs::IsLightingSamplerSupported(lighting.config,
LightingRegs::LightingSampler::ReflectBlue)) { LightingRegs::LightingSampler::ReflectBlue)) {
std::string index = std::string value =
GetLutIndex(light_config.num, lighting.lut_rb.type, lighting.lut_rb.abs_input); GetLutValue(LightingRegs::LightingSampler::ReflectBlue, light_config.num,
std::string value = "(" + std::to_string(lighting.lut_rb.scale) + " * " + lighting.lut_rb.type, lighting.lut_rb.abs_input);
GetLutValue(LightingRegs::LightingSampler::ReflectBlue, index) + value = "(" + std::to_string(lighting.lut_rb.scale) + " * " + value + ")";
")";
out += "refl_value.b = " + value + ";\n"; out += "refl_value.b = " + value + ";\n";
} else { } else {
out += "refl_value.b = refl_value.r;\n"; out += "refl_value.b = refl_value.r;\n";
@ -743,10 +737,10 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
LightingRegs::IsLightingSamplerSupported( LightingRegs::IsLightingSamplerSupported(
lighting.config, LightingRegs::LightingSampler::Distribution1)) { lighting.config, LightingRegs::LightingSampler::Distribution1)) {
// Lookup specular "distribution 1" LUT value // Lookup specular "distribution 1" LUT value
std::string index = std::string value =
GetLutIndex(light_config.num, lighting.lut_d1.type, lighting.lut_d1.abs_input); GetLutValue(LightingRegs::LightingSampler::Distribution1, light_config.num,
d1_lut_value = "(" + std::to_string(lighting.lut_d1.scale) + " * " + lighting.lut_d1.type, lighting.lut_d1.abs_input);
GetLutValue(LightingRegs::LightingSampler::Distribution1, index) + ")"; d1_lut_value = "(" + std::to_string(lighting.lut_d1.scale) + " * " + value + ")";
} }
std::string specular_1 = std::string specular_1 =
"(" + d1_lut_value + " * refl_value * " + light_src + ".specular_1)"; "(" + d1_lut_value + " * refl_value * " + light_src + ".specular_1)";
@ -759,10 +753,10 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) {
LightingRegs::IsLightingSamplerSupported(lighting.config, LightingRegs::IsLightingSamplerSupported(lighting.config,
LightingRegs::LightingSampler::Fresnel)) { LightingRegs::LightingSampler::Fresnel)) {
// Lookup fresnel LUT value // Lookup fresnel LUT value
std::string index = std::string value =
GetLutIndex(light_config.num, lighting.lut_fr.type, lighting.lut_fr.abs_input); GetLutValue(LightingRegs::LightingSampler::Fresnel, light_config.num,
std::string value = "(" + std::to_string(lighting.lut_fr.scale) + " * " + lighting.lut_fr.type, lighting.lut_fr.abs_input);
GetLutValue(LightingRegs::LightingSampler::Fresnel, index) + ")"; value = "(" + std::to_string(lighting.lut_fr.scale) + " * " + value + ")";
// Enabled for difffuse lighting alpha component // Enabled for difffuse lighting alpha component
if (lighting.fresnel_selector == LightingRegs::LightingFresnelSelector::PrimaryAlpha || if (lighting.fresnel_selector == LightingRegs::LightingFresnelSelector::PrimaryAlpha ||
@ -1016,10 +1010,6 @@ std::string GenerateFragmentShader(const PicaShaderConfig& config) {
#define NUM_TEV_STAGES 6 #define NUM_TEV_STAGES 6
#define NUM_LIGHTS 8 #define NUM_LIGHTS 8
// Texture coordinate offsets and scales
#define OFFSET_256 (0.5 / 256.0)
#define SCALE_256 (255.0 / 256.0)
in vec4 primary_color; in vec4 primary_color;
in vec2 texcoord[3]; in vec2 texcoord[3];
in float texcoord0_w; in float texcoord0_w;
@ -1061,7 +1051,7 @@ layout (std140) uniform shader_data {
}; };
uniform sampler2D tex[3]; uniform sampler2D tex[3];
uniform sampler1D lut[6]; uniform samplerBuffer lighting_lut;
uniform usampler1D fog_lut; uniform usampler1D fog_lut;
uniform sampler1D proctex_noise_lut; uniform sampler1D proctex_noise_lut;
uniform sampler1D proctex_color_map; uniform sampler1D proctex_color_map;
@ -1074,6 +1064,24 @@ vec3 quaternion_rotate(vec4 q, vec3 v) {
return v + 2.0 * cross(q.xyz, cross(q.xyz, v) + q.w * v); return v + 2.0 * cross(q.xyz, cross(q.xyz, v) + q.w * v);
} }
float LookupLightingLUT(int lut_index, int index, float delta) {
vec2 entry = texelFetch(lighting_lut, lut_index * 256 + index).rg;
return entry.r + entry.g * delta;
}
float LookupLightingLUTUnsigned(int lut_index, float pos) {
int index = clamp(int(pos * 256.0), 0, 255);
float delta = pos * 256.0 - index;
return LookupLightingLUT(lut_index, index, delta);
}
float LookupLightingLUTSigned(int lut_index, float pos) {
int index = clamp(int(pos * 128.0), -128, 127);
float delta = pos * 128.0 - index;
if (index < 0) index += 256;
return LookupLightingLUT(lut_index, index, delta);
}
)"; )";
if (config.state.proctex.enable) if (config.state.proctex.enable)

View file

@ -52,9 +52,7 @@ OpenGLState::OpenGLState() {
texture_unit.sampler = 0; texture_unit.sampler = 0;
} }
for (auto& lut : lighting_luts) { lighting_lut.texture_buffer = 0;
lut.texture_1d = 0;
}
fog_lut.texture_1d = 0; fog_lut.texture_1d = 0;
@ -185,7 +183,7 @@ void OpenGLState::Apply() const {
// Textures // Textures
for (unsigned i = 0; i < ARRAY_SIZE(texture_units); ++i) { for (unsigned i = 0; i < ARRAY_SIZE(texture_units); ++i) {
if (texture_units[i].texture_2d != cur_state.texture_units[i].texture_2d) { if (texture_units[i].texture_2d != cur_state.texture_units[i].texture_2d) {
glActiveTexture(GL_TEXTURE0 + i); glActiveTexture(TextureUnits::PicaTexture(i).Enum());
glBindTexture(GL_TEXTURE_2D, texture_units[i].texture_2d); glBindTexture(GL_TEXTURE_2D, texture_units[i].texture_2d);
} }
if (texture_units[i].sampler != cur_state.texture_units[i].sampler) { if (texture_units[i].sampler != cur_state.texture_units[i].sampler) {
@ -194,46 +192,44 @@ void OpenGLState::Apply() const {
} }
// Lighting LUTs // Lighting LUTs
for (unsigned i = 0; i < ARRAY_SIZE(lighting_luts); ++i) { if (lighting_lut.texture_buffer != cur_state.lighting_lut.texture_buffer) {
if (lighting_luts[i].texture_1d != cur_state.lighting_luts[i].texture_1d) { glActiveTexture(TextureUnits::LightingLUT.Enum());
glActiveTexture(GL_TEXTURE3 + i); glBindTexture(GL_TEXTURE_BUFFER, cur_state.lighting_lut.texture_buffer);
glBindTexture(GL_TEXTURE_1D, lighting_luts[i].texture_1d);
}
} }
// Fog LUT // Fog LUT
if (fog_lut.texture_1d != cur_state.fog_lut.texture_1d) { if (fog_lut.texture_1d != cur_state.fog_lut.texture_1d) {
glActiveTexture(GL_TEXTURE9); glActiveTexture(TextureUnits::FogLUT.Enum());
glBindTexture(GL_TEXTURE_1D, fog_lut.texture_1d); glBindTexture(GL_TEXTURE_1D, fog_lut.texture_1d);
} }
// ProcTex Noise LUT // ProcTex Noise LUT
if (proctex_noise_lut.texture_1d != cur_state.proctex_noise_lut.texture_1d) { if (proctex_noise_lut.texture_1d != cur_state.proctex_noise_lut.texture_1d) {
glActiveTexture(GL_TEXTURE10); glActiveTexture(TextureUnits::ProcTexNoiseLUT.Enum());
glBindTexture(GL_TEXTURE_1D, proctex_noise_lut.texture_1d); glBindTexture(GL_TEXTURE_1D, proctex_noise_lut.texture_1d);
} }
// ProcTex Color Map // ProcTex Color Map
if (proctex_color_map.texture_1d != cur_state.proctex_color_map.texture_1d) { if (proctex_color_map.texture_1d != cur_state.proctex_color_map.texture_1d) {
glActiveTexture(GL_TEXTURE11); glActiveTexture(TextureUnits::ProcTexColorMap.Enum());
glBindTexture(GL_TEXTURE_1D, proctex_color_map.texture_1d); glBindTexture(GL_TEXTURE_1D, proctex_color_map.texture_1d);
} }
// ProcTex Alpha Map // ProcTex Alpha Map
if (proctex_alpha_map.texture_1d != cur_state.proctex_alpha_map.texture_1d) { if (proctex_alpha_map.texture_1d != cur_state.proctex_alpha_map.texture_1d) {
glActiveTexture(GL_TEXTURE12); glActiveTexture(TextureUnits::ProcTexAlphaMap.Enum());
glBindTexture(GL_TEXTURE_1D, proctex_alpha_map.texture_1d); glBindTexture(GL_TEXTURE_1D, proctex_alpha_map.texture_1d);
} }
// ProcTex LUT // ProcTex LUT
if (proctex_lut.texture_1d != cur_state.proctex_lut.texture_1d) { if (proctex_lut.texture_1d != cur_state.proctex_lut.texture_1d) {
glActiveTexture(GL_TEXTURE13); glActiveTexture(TextureUnits::ProcTexLUT.Enum());
glBindTexture(GL_TEXTURE_1D, proctex_lut.texture_1d); glBindTexture(GL_TEXTURE_1D, proctex_lut.texture_1d);
} }
// ProcTex Diff LUT // ProcTex Diff LUT
if (proctex_diff_lut.texture_1d != cur_state.proctex_diff_lut.texture_1d) { if (proctex_diff_lut.texture_1d != cur_state.proctex_diff_lut.texture_1d) {
glActiveTexture(GL_TEXTURE14); glActiveTexture(TextureUnits::ProcTexDiffLUT.Enum());
glBindTexture(GL_TEXTURE_1D, proctex_diff_lut.texture_1d); glBindTexture(GL_TEXTURE_1D, proctex_diff_lut.texture_1d);
} }
@ -274,6 +270,20 @@ void OpenGLState::ResetTexture(GLuint handle) {
unit.texture_2d = 0; unit.texture_2d = 0;
} }
} }
if (cur_state.lighting_lut.texture_buffer == handle)
cur_state.lighting_lut.texture_buffer = 0;
if (cur_state.fog_lut.texture_1d == handle)
cur_state.fog_lut.texture_1d = 0;
if (cur_state.proctex_noise_lut.texture_1d == handle)
cur_state.proctex_noise_lut.texture_1d = 0;
if (cur_state.proctex_color_map.texture_1d == handle)
cur_state.proctex_color_map.texture_1d = 0;
if (cur_state.proctex_alpha_map.texture_1d == handle)
cur_state.proctex_alpha_map.texture_1d = 0;
if (cur_state.proctex_lut.texture_1d == handle)
cur_state.proctex_lut.texture_1d = 0;
if (cur_state.proctex_diff_lut.texture_1d == handle)
cur_state.proctex_diff_lut.texture_1d = 0;
} }
void OpenGLState::ResetSampler(GLuint handle) { void OpenGLState::ResetSampler(GLuint handle) {

View file

@ -6,6 +6,29 @@
#include <glad/glad.h> #include <glad/glad.h>
namespace TextureUnits {
struct TextureUnit {
GLint id;
constexpr GLenum Enum() const {
return static_cast<GLenum>(GL_TEXTURE0 + id);
}
};
constexpr TextureUnit PicaTexture(int unit) {
return TextureUnit{unit};
}
constexpr TextureUnit LightingLUT{3};
constexpr TextureUnit FogLUT{4};
constexpr TextureUnit ProcTexNoiseLUT{5};
constexpr TextureUnit ProcTexColorMap{6};
constexpr TextureUnit ProcTexAlphaMap{7};
constexpr TextureUnit ProcTexLUT{8};
constexpr TextureUnit ProcTexDiffLUT{9};
} // namespace TextureUnits
class OpenGLState { class OpenGLState {
public: public:
struct { struct {
@ -64,8 +87,8 @@ public:
} texture_units[3]; } texture_units[3];
struct { struct {
GLuint texture_1d; // GL_TEXTURE_BINDING_1D GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
} lighting_luts[6]; } lighting_lut;
struct { struct {
GLuint texture_1d; // GL_TEXTURE_BINDING_1D GLuint texture_1d; // GL_TEXTURE_BINDING_1D