Mirrors/citra-nightly
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

common/vector_math: Move Vec[x] types into the Common namespace

Browse source 
These types are within the common library, so they should be using the
Common namespace.
This commit is contained in:
Lioncash 2019-02-26 22:38:34 -05:00 • committed by fearlessTobi
parent db58652680
commit 643472e24a
40 changed files with 309 additions and 301 deletions
Download patch file Download diff file Expand all files Collapse all files

35
src/video_core/swrasterizer/clipper.cpp
View file

@ -23,13 +23,15 @@ namespace Pica::Clipper {
struct ClippingEdge {
public:
ClippingEdge(Math::Vec4<float24> coeffs, Math::Vec4<float24> bias = Math::Vec4<float24>(
float24::FromFloat32(0), float24::FromFloat32(0),
float24::FromFloat32(0), float24::FromFloat32(0)))
ClippingEdge(Common::Vec4<float24> coeffs,
Common::Vec4<float24> bias = Common::Vec4<float24>(float24::FromFloat32(0),
float24::FromFloat32(0),
float24::FromFloat32(0),
float24::FromFloat32(0)))
: coeffs(coeffs), bias(bias) {}
bool IsInside(const Vertex& vertex) const {
return Math::Dot(vertex.pos + bias, coeffs) >= float24::FromFloat32(0);
return Common::Dot(vertex.pos + bias, coeffs) >= float24::FromFloat32(0);
}
bool IsOutSide(const Vertex& vertex) const {
@ -37,8 +39,8 @@ public:
}
Vertex GetIntersection(const Vertex& v0, const Vertex& v1) const {
float24 dp = Math::Dot(v0.pos + bias, coeffs);
float24 dp_prev = Math::Dot(v1.pos + bias, coeffs);
float24 dp = Common::Dot(v0.pos + bias, coeffs);
float24 dp_prev = Common::Dot(v1.pos + bias, coeffs);
float24 factor = dp_prev / (dp_prev - dp);
return Vertex::Lerp(factor, v0, v1);
@ -46,8 +48,8 @@ public:
private:
float24 pos;
Math::Vec4<float24> coeffs;
Math::Vec4<float24> bias;
Common::Vec4<float24> coeffs;
Common::Vec4<float24> bias;
};
static void InitScreenCoordinates(Vertex& vtx) {
@ -95,7 +97,7 @@ void ProcessTriangle(const OutputVertex& v0, const OutputVertex& v1, const Outpu
static_vector<Vertex, MAX_VERTICES> buffer_b;
auto FlipQuaternionIfOpposite = [](auto& a, const auto& b) {
if (Math::Dot(a, b) < float24::Zero())
if (Common::Dot(a, b) < float24::Zero())
a = a * float24::FromFloat32(-1.0f);
};
@ -114,13 +116,14 @@ void ProcessTriangle(const OutputVertex& v0, const OutputVertex& v1, const Outpu
static const float24 f0 = float24::FromFloat32(0.0);
static const float24 f1 = float24::FromFloat32(1.0);
static const std::array<ClippingEdge, 7> clipping_edges = {{
{Math::MakeVec(-f1, f0, f0, f1)}, // x = +w
{Math::MakeVec(f1, f0, f0, f1)}, // x = -w
{Math::MakeVec(f0, -f1, f0, f1)}, // y = +w
{Math::MakeVec(f0, f1, f0, f1)}, // y = -w
{Math::MakeVec(f0, f0, -f1, f0)}, // z = 0
{Math::MakeVec(f0, f0, f1, f1)}, // z = -w
{Math::MakeVec(f0, f0, f0, f1), Math::Vec4<float24>(f0, f0, f0, EPSILON)}, // w = EPSILON
{Common::MakeVec(-f1, f0, f0, f1)}, // x = +w
{Common::MakeVec(f1, f0, f0, f1)}, // x = -w
{Common::MakeVec(f0, -f1, f0, f1)}, // y = +w
{Common::MakeVec(f0, f1, f0, f1)}, // y = -w
{Common::MakeVec(f0, f0, -f1, f0)}, // z = 0
{Common::MakeVec(f0, f0, f1, f1)}, // z = -w
{Common::MakeVec(f0, f0, f0, f1),
Common::Vec4<float24>(f0, f0, f0, EPSILON)}, // w = EPSILON
}};
// Simple implementation of the Sutherland-Hodgman clipping algorithm.

20
src/video_core/swrasterizer/framebuffer.cpp
View file

@ -18,7 +18,7 @@
namespace Pica::Rasterizer {
void DrawPixel(int x, int y, const Math::Vec4<u8>& color) {
void DrawPixel(int x, int y, const Common::Vec4<u8>& color) {
const auto& framebuffer = g_state.regs.framebuffer.framebuffer;
const PAddr addr = framebuffer.GetColorBufferPhysicalAddress();
@ -61,7 +61,7 @@ void DrawPixel(int x, int y, const Math::Vec4<u8>& color) {
}
}
const Math::Vec4<u8> GetPixel(int x, int y) {
const Common::Vec4<u8> GetPixel(int x, int y) {
const auto& framebuffer = g_state.regs.framebuffer.framebuffer;
const PAddr addr = framebuffer.GetColorBufferPhysicalAddress();
@ -257,10 +257,12 @@ u8 PerformStencilAction(FramebufferRegs::StencilAction action, u8 old_stencil, u
}
}
Math::Vec4<u8> EvaluateBlendEquation(const Math::Vec4<u8>& src, const Math::Vec4<u8>& srcfactor,
const Math::Vec4<u8>& dest, const Math::Vec4<u8>& destfactor,
FramebufferRegs::BlendEquation equation) {
Math::Vec4<int> result;
Common::Vec4<u8> EvaluateBlendEquation(const Common::Vec4<u8>& src,
const Common::Vec4<u8>& srcfactor,
const Common::Vec4<u8>& dest,
const Common::Vec4<u8>& destfactor,
FramebufferRegs::BlendEquation equation) {
Common::Vec4<int> result;
auto src_result = (src * srcfactor).Cast<int>();
auto dst_result = (dest * destfactor).Cast<int>();
@ -299,8 +301,8 @@ Math::Vec4<u8> EvaluateBlendEquation(const Math::Vec4<u8>& src, const Math::Vec4
UNIMPLEMENTED();
}
return Math::Vec4<u8>(std::clamp(result.r(), 0, 255), std::clamp(result.g(), 0, 255),
std::clamp(result.b(), 0, 255), std::clamp(result.a(), 0, 255));
return Common::Vec4<u8>(std::clamp(result.r(), 0, 255), std::clamp(result.g(), 0, 255),
std::clamp(result.b(), 0, 255), std::clamp(result.a(), 0, 255));
};
u8 LogicOp(u8 src, u8 dest, FramebufferRegs::LogicOp op) {
@ -359,7 +361,7 @@ u8 LogicOp(u8 src, u8 dest, FramebufferRegs::LogicOp op) {
// Decode/Encode for shadow map format. It is similar to D24S8 format, but the depth field is in
// big-endian
static const Math::Vec2<u32> DecodeD24S8Shadow(const u8* bytes) {
static const Common::Vec2<u32> DecodeD24S8Shadow(const u8* bytes) {
return {static_cast<u32>((bytes[0] << 16) | (bytes[1] << 8) | bytes[2]), bytes[3]};
}

12
src/video_core/swrasterizer/framebuffer.h
View file

@ -10,17 +10,19 @@
namespace Pica::Rasterizer {
void DrawPixel(int x, int y, const Math::Vec4<u8>& color);
const Math::Vec4<u8> GetPixel(int x, int y);
void DrawPixel(int x, int y, const Common::Vec4<u8>& color);
const Common::Vec4<u8> GetPixel(int x, int y);
u32 GetDepth(int x, int y);
u8 GetStencil(int x, int y);
void SetDepth(int x, int y, u32 value);
void SetStencil(int x, int y, u8 value);
u8 PerformStencilAction(FramebufferRegs::StencilAction action, u8 old_stencil, u8 ref);
Math::Vec4<u8> EvaluateBlendEquation(const Math::Vec4<u8>& src, const Math::Vec4<u8>& srcfactor,
const Math::Vec4<u8>& dest, const Math::Vec4<u8>& destfactor,
FramebufferRegs::BlendEquation equation);
Common::Vec4<u8> EvaluateBlendEquation(const Common::Vec4<u8>& src,
const Common::Vec4<u8>& srcfactor,
const Common::Vec4<u8>& dest,
const Common::Vec4<u8>& destfactor,
FramebufferRegs::BlendEquation equation);
u8 LogicOp(u8 src, u8 dest, FramebufferRegs::LogicOp op);

100
src/video_core/swrasterizer/lighting.cpp
View file

@ -20,63 +20,63 @@ static float LookupLightingLut(const Pica::State::Lighting& lighting, std::size_
return lut_value + lut_diff * delta;
}
std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
std::tuple<Common::Vec4<u8>, Common::Vec4<u8>> ComputeFragmentsColors(
const Pica::LightingRegs& lighting, const Pica::State::Lighting& lighting_state,
const Math::Quaternion<float>& normquat, const Math::Vec3<float>& view,
const Math::Vec4<u8> (&texture_color)[4]) {
const Common::Quaternion<float>& normquat, const Common::Vec3<float>& view,
const Common::Vec4<u8> (&texture_color)[4]) {
Math::Vec4<float> shadow;
Common::Vec4<float> shadow;
if (lighting.config0.enable_shadow) {
shadow = texture_color[lighting.config0.shadow_selector].Cast<float>() / 255.0f;
if (lighting.config0.shadow_invert) {
shadow = Math::MakeVec(1.0f, 1.0f, 1.0f, 1.0f) - shadow;
shadow = Common::MakeVec(1.0f, 1.0f, 1.0f, 1.0f) - shadow;
}
} else {
shadow = Math::MakeVec(1.0f, 1.0f, 1.0f, 1.0f);
shadow = Common::MakeVec(1.0f, 1.0f, 1.0f, 1.0f);
}
Math::Vec3<float> surface_normal;
Math::Vec3<float> surface_tangent;
Common::Vec3<float> surface_normal;
Common::Vec3<float> surface_tangent;
if (lighting.config0.bump_mode != LightingRegs::LightingBumpMode::None) {
Math::Vec3<float> perturbation =
Common::Vec3<float> perturbation =
texture_color[lighting.config0.bump_selector].xyz().Cast<float>() / 127.5f -
Math::MakeVec(1.0f, 1.0f, 1.0f);
Common::MakeVec(1.0f, 1.0f, 1.0f);
if (lighting.config0.bump_mode == LightingRegs::LightingBumpMode::NormalMap) {
if (!lighting.config0.disable_bump_renorm) {
const float z_square = 1 - perturbation.xy().Length2();
perturbation.z = std::sqrt(std::max(z_square, 0.0f));
}
surface_normal = perturbation;
surface_tangent = Math::MakeVec(1.0f, 0.0f, 0.0f);
surface_tangent = Common::MakeVec(1.0f, 0.0f, 0.0f);
} else if (lighting.config0.bump_mode == LightingRegs::LightingBumpMode::TangentMap) {
surface_normal = Math::MakeVec(0.0f, 0.0f, 1.0f);
surface_normal = Common::MakeVec(0.0f, 0.0f, 1.0f);
surface_tangent = perturbation;
} else {
LOG_ERROR(HW_GPU, "Unknown bump mode {}",
static_cast<u32>(lighting.config0.bump_mode.Value()));
}
} else {
surface_normal = Math::MakeVec(0.0f, 0.0f, 1.0f);
surface_tangent = Math::MakeVec(1.0f, 0.0f, 0.0f);
surface_normal = Common::MakeVec(0.0f, 0.0f, 1.0f);
surface_tangent = Common::MakeVec(1.0f, 0.0f, 0.0f);
}
// Use the normalized the quaternion when performing the rotation
auto normal = Math::QuaternionRotate(normquat, surface_normal);
auto tangent = Math::QuaternionRotate(normquat, surface_tangent);
auto normal = Common::QuaternionRotate(normquat, surface_normal);
auto tangent = Common::QuaternionRotate(normquat, surface_tangent);
Math::Vec4<float> diffuse_sum = {0.0f, 0.0f, 0.0f, 1.0f};
Math::Vec4<float> specular_sum = {0.0f, 0.0f, 0.0f, 1.0f};
Common::Vec4<float> diffuse_sum = {0.0f, 0.0f, 0.0f, 1.0f};
Common::Vec4<float> specular_sum = {0.0f, 0.0f, 0.0f, 1.0f};
for (unsigned light_index = 0; light_index <= lighting.max_light_index; ++light_index) {
unsigned num = lighting.light_enable.GetNum(light_index);
const auto& light_config = lighting.light[num];
Math::Vec3<float> refl_value = {};
Math::Vec3<float> position = {float16::FromRaw(light_config.x).ToFloat32(),
float16::FromRaw(light_config.y).ToFloat32(),
float16::FromRaw(light_config.z).ToFloat32()};
Math::Vec3<float> light_vector;
Common::Vec3<float> refl_value = {};
Common::Vec3<float> position = {float16::FromRaw(light_config.x).ToFloat32(),
float16::FromRaw(light_config.y).ToFloat32(),
float16::FromRaw(light_config.z).ToFloat32()};
Common::Vec3<float> light_vector;
if (light_config.config.directional)
light_vector = position;
@ -85,8 +85,8 @@ std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
light_vector.Normalize();
Math::Vec3<float> norm_view = view.Normalized();
Math::Vec3<float> half_vector = norm_view + light_vector;
Common::Vec3<float> norm_view = view.Normalized();
Common::Vec3<float> half_vector = norm_view + light_vector;
float dist_atten = 1.0f;
if (!lighting.IsDistAttenDisabled(num)) {
@ -111,33 +111,33 @@ std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
switch (input) {
case LightingRegs::LightingLutInput::NH:
result = Math::Dot(normal, half_vector.Normalized());
result = Common::Dot(normal, half_vector.Normalized());
break;
case LightingRegs::LightingLutInput::VH:
result = Math::Dot(norm_view, half_vector.Normalized());
result = Common::Dot(norm_view, half_vector.Normalized());
break;
case LightingRegs::LightingLutInput::NV:
result = Math::Dot(normal, norm_view);
result = Common::Dot(normal, norm_view);
break;
case LightingRegs::LightingLutInput::LN:
result = Math::Dot(light_vector, normal);
result = Common::Dot(light_vector, normal);
break;
case LightingRegs::LightingLutInput::SP: {
Math::Vec3<s32> spot_dir{light_config.spot_x.Value(), light_config.spot_y.Value(),
light_config.spot_z.Value()};
result = Math::Dot(light_vector, spot_dir.Cast<float>() / 2047.0f);
Common::Vec3<s32> spot_dir{light_config.spot_x.Value(), light_config.spot_y.Value(),
light_config.spot_z.Value()};
result = Common::Dot(light_vector, spot_dir.Cast<float>() / 2047.0f);
break;
}
case LightingRegs::LightingLutInput::CP:
if (lighting.config0.config == LightingRegs::LightingConfig::Config7) {
const Math::Vec3<float> norm_half_vector = half_vector.Normalized();
const Math::Vec3<float> half_vector_proj =
norm_half_vector - normal * Math::Dot(normal, norm_half_vector);
result = Math::Dot(half_vector_proj, tangent);
const Common::Vec3<float> norm_half_vector = half_vector.Normalized();
const Common::Vec3<float> half_vector_proj =
norm_half_vector - normal * Common::Dot(normal, norm_half_vector);
result = Common::Dot(half_vector_proj, tangent);
} else {
result = 0.0f;
}
@ -192,7 +192,7 @@ std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
lighting.lut_scale.d0, LightingRegs::LightingSampler::Distribution0);
}
Math::Vec3<float> specular_0 = d0_lut_value * light_config.specular_0.ToVec3f();
Common::Vec3<float> specular_0 = d0_lut_value * light_config.specular_0.ToVec3f();
// If enabled, lookup ReflectRed value, otherwise, 1.0 is used
if (lighting.config1.disable_lut_rr == 0 &&
@ -237,7 +237,7 @@ std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
lighting.lut_scale.d1, LightingRegs::LightingSampler::Distribution1);
}
Math::Vec3<float> specular_1 =
Common::Vec3<float> specular_1 =
d1_lut_value * refl_value * light_config.specular_1.ToVec3f();
// Fresnel
@ -261,7 +261,7 @@ std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
}
}
auto dot_product = Math::Dot(light_vector, normal);
auto dot_product = Common::Dot(light_vector, normal);
if (light_config.config.two_sided_diffuse)
dot_product = std::abs(dot_product);
else
@ -297,8 +297,8 @@ std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
}
}
diffuse_sum += Math::MakeVec(diffuse, 0.0f);
specular_sum += Math::MakeVec(specular, 0.0f);
diffuse_sum += Common::MakeVec(diffuse, 0.0f);
specular_sum += Common::MakeVec(specular, 0.0f);
}
if (lighting.config0.shadow_alpha) {
@ -314,17 +314,17 @@ std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
}
}
diffuse_sum += Math::MakeVec(lighting.global_ambient.ToVec3f(), 0.0f);
diffuse_sum += Common::MakeVec(lighting.global_ambient.ToVec3f(), 0.0f);
auto diffuse = Math::MakeVec<float>(std::clamp(diffuse_sum.x, 0.0f, 1.0f) * 255,
std::clamp(diffuse_sum.y, 0.0f, 1.0f) * 255,
std::clamp(diffuse_sum.z, 0.0f, 1.0f) * 255,
std::clamp(diffuse_sum.w, 0.0f, 1.0f) * 255)
auto diffuse = Common::MakeVec<float>(std::clamp(diffuse_sum.x, 0.0f, 1.0f) * 255,
std::clamp(diffuse_sum.y, 0.0f, 1.0f) * 255,
std::clamp(diffuse_sum.z, 0.0f, 1.0f) * 255,
std::clamp(diffuse_sum.w, 0.0f, 1.0f) * 255)
.Cast<u8>();
auto specular = Math::MakeVec<float>(std::clamp(specular_sum.x, 0.0f, 1.0f) * 255,
std::clamp(specular_sum.y, 0.0f, 1.0f) * 255,
std::clamp(specular_sum.z, 0.0f, 1.0f) * 255,
std::clamp(specular_sum.w, 0.0f, 1.0f) * 255)
auto specular = Common::MakeVec<float>(std::clamp(specular_sum.x, 0.0f, 1.0f) * 255,
std::clamp(specular_sum.y, 0.0f, 1.0f) * 255,
std::clamp(specular_sum.z, 0.0f, 1.0f) * 255,
std::clamp(specular_sum.w, 0.0f, 1.0f) * 255)
.Cast<u8>();
return std::make_tuple(diffuse, specular);
}

6
src/video_core/swrasterizer/lighting.h
View file

@ -11,9 +11,9 @@
namespace Pica {
std::tuple<Math::Vec4<u8>, Math::Vec4<u8>> ComputeFragmentsColors(
std::tuple<Common::Vec4<u8>, Common::Vec4<u8>> ComputeFragmentsColors(
const Pica::LightingRegs& lighting, const Pica::State::Lighting& lighting_state,
const Math::Quaternion<float>& normquat, const Math::Vec3<float>& view,
const Math::Vec4<u8> (&texture_color)[4]);
const Common::Quaternion<float>& normquat, const Common::Vec3<float>& view,
const Common::Vec4<u8> (&texture_color)[4]);
} // namespace Pica

8
src/video_core/swrasterizer/proctex.cpp
View file

@ -63,7 +63,7 @@ static float NoiseCoef(float u, float v, TexturingRegs regs, State::ProcTex stat
const float g3 = NoiseRand2D(x_int + 1, y_int + 1) * (x_frac + y_frac - 2);
const float x_noise = LookupLUT(state.noise_table, x_frac);
const float y_noise = LookupLUT(state.noise_table, y_frac);
return Math::BilinearInterp(g0, g1, g2, g3, x_noise, y_noise);
return Common::BilinearInterp(g0, g1, g2, g3, x_noise, y_noise);
}
static float GetShiftOffset(float v, ProcTexShift mode, ProcTexClamp clamp_mode) {
@ -154,7 +154,7 @@ float CombineAndMap(float u, float v, ProcTexCombiner combiner,
return LookupLUT(map_table, f);
}
Math::Vec4<u8> ProcTex(float u, float v, TexturingRegs regs, State::ProcTex state) {
Common::Vec4<u8> ProcTex(float u, float v, TexturingRegs regs, State::ProcTex state) {
u = std::abs(u);
v = std::abs(v);
@ -187,7 +187,7 @@ Math::Vec4<u8> ProcTex(float u, float v, TexturingRegs regs, State::ProcTex stat
const u32 offset = regs.proctex_lut_offset.level0;
const u32 width = regs.proctex_lut.width;
const float index = offset + (lut_coord * (width - 1));
Math::Vec4<u8> final_color;
Common::Vec4<u8> final_color;
// TODO(wwylele): implement mipmap
switch (regs.proctex_lut.filter) {
case ProcTexFilter::Linear:
@ -212,7 +212,7 @@ Math::Vec4<u8> ProcTex(float u, float v, TexturingRegs regs, State::ProcTex stat
// uses the output of CombineAndMap directly instead.
const float final_alpha =
CombineAndMap(u, v, regs.proctex.alpha_combiner, state.alpha_map_table);
return Math::MakeVec<u8>(final_color.rgb(), static_cast<u8>(final_alpha * 255));
return Common::MakeVec<u8>(final_color.rgb(), static_cast<u8>(final_alpha * 255));
} else {
return final_color;
}

2
src/video_core/swrasterizer/proctex.h
View file

@ -9,6 +9,6 @@
namespace Pica::Rasterizer {
/// Generates procedural texture color for the given coordinates
Math::Vec4<u8> ProcTex(float u, float v, TexturingRegs regs, State::ProcTex state);
Common::Vec4<u8> ProcTex(float u, float v, TexturingRegs regs, State::ProcTex state);
} // namespace Pica::Rasterizer

133
src/video_core/swrasterizer/rasterizer.cpp
View file

@ -64,12 +64,12 @@ private:
*
* @todo define orientation concretely.
*/
static int SignedArea(const Math::Vec2<Fix12P4>& vtx1, const Math::Vec2<Fix12P4>& vtx2,
const Math::Vec2<Fix12P4>& vtx3) {
const auto vec1 = Math::MakeVec(vtx2 - vtx1, 0);
const auto vec2 = Math::MakeVec(vtx3 - vtx1, 0);
static int SignedArea(const Common::Vec2<Fix12P4>& vtx1, const Common::Vec2<Fix12P4>& vtx2,
const Common::Vec2<Fix12P4>& vtx3) {
const auto vec1 = Common::MakeVec(vtx2 - vtx1, 0);
const auto vec2 = Common::MakeVec(vtx3 - vtx1, 0);
// TODO: There is a very small chance this will overflow for sizeof(int) == 4
return Math::Cross(vec1, vec2).z;
return Common::Cross(vec1, vec2).z;
};
/// Convert a 3D vector for cube map coordinates to 2D texture coordinates along with the face name
@ -134,13 +134,13 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
// triangle borders. Is it that the correct solution, though?
return Fix12P4(static_cast<unsigned short>(round(flt.ToFloat32() * 16.0f)));
};
static auto ScreenToRasterizerCoordinates = [](const Math::Vec3<float24>& vec) {
return Math::Vec3<Fix12P4>{FloatToFix(vec.x), FloatToFix(vec.y), FloatToFix(vec.z)};
static auto ScreenToRasterizerCoordinates = [](const Common::Vec3<float24>& vec) {
return Common::Vec3<Fix12P4>{FloatToFix(vec.x), FloatToFix(vec.y), FloatToFix(vec.z)};
};
Math::Vec3<Fix12P4> vtxpos[3]{ScreenToRasterizerCoordinates(v0.screenpos),
ScreenToRasterizerCoordinates(v1.screenpos),
ScreenToRasterizerCoordinates(v2.screenpos)};
Common::Vec3<Fix12P4> vtxpos[3]{ScreenToRasterizerCoordinates(v0.screenpos),
ScreenToRasterizerCoordinates(v1.screenpos),
ScreenToRasterizerCoordinates(v2.screenpos)};
if (regs.rasterizer.cull_mode == RasterizerRegs::CullMode::KeepAll) {
// Make sure we always end up with a triangle wound counter-clockwise
@ -189,9 +189,9 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
// drawn. Pixels on any other triangle border are drawn. This is implemented with three bias
// values which are added to the barycentric coordinates w0, w1 and w2, respectively.
// NOTE: These are the PSP filling rules. Not sure if the 3DS uses the same ones...
auto IsRightSideOrFlatBottomEdge = [](const Math::Vec2<Fix12P4>& vtx,
const Math::Vec2<Fix12P4>& line1,
const Math::Vec2<Fix12P4>& line2) {
auto IsRightSideOrFlatBottomEdge = [](const Common::Vec2<Fix12P4>& vtx,
const Common::Vec2<Fix12P4>& line1,
const Common::Vec2<Fix12P4>& line2) {
if (line1.y == line2.y) {
// just check if vertex is above us => bottom line parallel to x-axis
return vtx.y < line1.y;
@ -210,7 +210,7 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
int bias2 =
IsRightSideOrFlatBottomEdge(vtxpos[2].xy(), vtxpos[0].xy(), vtxpos[1].xy()) ? -1 : 0;
auto w_inverse = Math::MakeVec(v0.pos.w, v1.pos.w, v2.pos.w);
auto w_inverse = Common::MakeVec(v0.pos.w, v1.pos.w, v2.pos.w);
auto textures = regs.texturing.GetTextures();
auto tev_stages = regs.texturing.GetTevStages();
@ -243,11 +243,11 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
continue;
auto baricentric_coordinates =
Math::MakeVec(float24::FromFloat32(static_cast<float>(w0)),
float24::FromFloat32(static_cast<float>(w1)),
float24::FromFloat32(static_cast<float>(w2)));
Common::MakeVec(float24::FromFloat32(static_cast<float>(w0)),
float24::FromFloat32(static_cast<float>(w1)),
float24::FromFloat32(static_cast<float>(w2)));
float24 interpolated_w_inverse =
float24::FromFloat32(1.0f) / Math::Dot(w_inverse, baricentric_coordinates);
float24::FromFloat32(1.0f) / Common::Dot(w_inverse, baricentric_coordinates);
// interpolated_z = z / w
float interpolated_z_over_w =
@ -288,12 +288,13 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
//
// The generalization to three vertices is straightforward in baricentric coordinates.
auto GetInterpolatedAttribute = [&](float24 attr0, float24 attr1, float24 attr2) {
auto attr_over_w = Math::MakeVec(attr0, attr1, attr2);
float24 interpolated_attr_over_w = Math::Dot(attr_over_w, baricentric_coordinates);
auto attr_over_w = Common::MakeVec(attr0, attr1, attr2);
float24 interpolated_attr_over_w =
Common::Dot(attr_over_w, baricentric_coordinates);
return interpolated_attr_over_w * interpolated_w_inverse;
};
Math::Vec4<u8> primary_color{
Common::Vec4<u8> primary_color{
static_cast<u8>(round(
GetInterpolatedAttribute(v0.color.r(), v1.color.r(), v2.color.r()).ToFloat32() *
255)),
@ -308,7 +309,7 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
255)),
};
Math::Vec2<float24> uv[3];
Common::Vec2<float24> uv[3];
uv[0].u() = GetInterpolatedAttribute(v0.tc0.u(), v1.tc0.u(), v2.tc0.u());
uv[0].v() = GetInterpolatedAttribute(v0.tc0.v(), v1.tc0.v(), v2.tc0.v());
uv[1].u() = GetInterpolatedAttribute(v0.tc1.u(), v1.tc1.u(), v2.tc1.u());
@ -316,7 +317,7 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
uv[2].u() = GetInterpolatedAttribute(v0.tc2.u(), v1.tc2.u(), v2.tc2.u());
uv[2].v() = GetInterpolatedAttribute(v0.tc2.v(), v1.tc2.v(), v2.tc2.v());
Math::Vec4<u8> texture_color[4]{};
Common::Vec4<u8> texture_color[4]{};
for (int i = 0; i < 3; ++i) {
const auto& texture = textures[i];
if (!texture.enabled)
@ -391,9 +392,10 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
if (use_border_s || use_border_t) {
auto border_color = texture.config.border_color;
texture_color[i] = Math::MakeVec(border_color.r.Value(), border_color.g.Value(),
border_color.b.Value(), border_color.a.Value())
.Cast<u8>();
texture_color[i] =
Common::MakeVec(border_color.r.Value(), border_color.g.Value(),
border_color.b.Value(), border_color.a.Value())
.Cast<u8>();
} else {
// Textures are laid out from bottom to top, hence we invert the t coordinate.
// NOTE: This may not be the right place for the inversion.
@ -442,21 +444,21 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
// 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;
Math::Vec4<u8> combiner_buffer = {0, 0, 0, 0};
Math::Vec4<u8> next_combiner_buffer =
Math::MakeVec(regs.texturing.tev_combiner_buffer_color.r.Value(),
regs.texturing.tev_combiner_buffer_color.g.Value(),
regs.texturing.tev_combiner_buffer_color.b.Value(),
regs.texturing.tev_combiner_buffer_color.a.Value())
Common::Vec4<u8> combiner_output;
Common::Vec4<u8> combiner_buffer = {0, 0, 0, 0};
Common::Vec4<u8> next_combiner_buffer =
Common::MakeVec(regs.texturing.tev_combiner_buffer_color.r.Value(),
regs.texturing.tev_combiner_buffer_color.g.Value(),
regs.texturing.tev_combiner_buffer_color.b.Value(),
regs.texturing.tev_combiner_buffer_color.a.Value())
.Cast<u8>();
Math::Vec4<u8> primary_fragment_color = {0, 0, 0, 0};
Math::Vec4<u8> secondary_fragment_color = {0, 0, 0, 0};
Common::Vec4<u8> primary_fragment_color = {0, 0, 0, 0};
Common::Vec4<u8> secondary_fragment_color = {0, 0, 0, 0};
if (!g_state.regs.lighting.disable) {
Math::Quaternion<float> normquat =
Math::Quaternion<float>{
Common::Quaternion<float> normquat =
Common::Quaternion<float>{
{GetInterpolatedAttribute(v0.quat.x, v1.quat.x, v2.quat.x).ToFloat32(),
GetInterpolatedAttribute(v0.quat.y, v1.quat.y, v2.quat.y).ToFloat32(),
GetInterpolatedAttribute(v0.quat.z, v1.quat.z, v2.quat.z).ToFloat32()},
@ -464,7 +466,7 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
}
.Normalized();
Math::Vec3<float> view{
Common::Vec3<float> view{
GetInterpolatedAttribute(v0.view.x, v1.view.x, v2.view.x).ToFloat32(),
GetInterpolatedAttribute(v0.view.y, v1.view.y, v2.view.y).ToFloat32(),
GetInterpolatedAttribute(v0.view.z, v1.view.z, v2.view.z).ToFloat32(),
@ -478,7 +480,7 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
const auto& tev_stage = tev_stages[tev_stage_index];
using Source = TexturingRegs::TevStageConfig::Source;
auto GetSource = [&](Source source) -> Math::Vec4<u8> {
auto GetSource = [&](Source source) -> Common::Vec4<u8> {
switch (source) {
case Source::PrimaryColor:
return primary_color;
@ -505,8 +507,8 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
return combiner_buffer;
case Source::Constant:
return Math::MakeVec(tev_stage.const_r.Value(), tev_stage.const_g.Value(),
tev_stage.const_b.Value(), tev_stage.const_a.Value())
return Common::MakeVec(tev_stage.const_r.Value(), tev_stage.const_g.Value(),
tev_stage.const_b.Value(), tev_stage.const_a.Value())
.Cast<u8>();
case Source::Previous:
@ -524,7 +526,7 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
// 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] = {
Common::Vec3<u8> color_result[3] = {
GetColorModifier(tev_stage.color_modifier1, GetSource(tev_stage.color_source1)),
GetColorModifier(tev_stage.color_modifier2, GetSource(tev_stage.color_source2)),
GetColorModifier(tev_stage.color_modifier3, GetSource(tev_stage.color_source3)),
@ -631,10 +633,11 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
// store the depth etc. Using float for now until we know more
// about Pica datatypes
if (regs.texturing.fog_mode == TexturingRegs::FogMode::Fog) {
const Math::Vec3<u8> fog_color = Math::MakeVec(regs.texturing.fog_color.r.Value(),
regs.texturing.fog_color.g.Value(),
regs.texturing.fog_color.b.Value())
.Cast<u8>();
const Common::Vec3<u8> fog_color =
Common::MakeVec(regs.texturing.fog_color.r.Value(),
regs.texturing.fog_color.g.Value(),
regs.texturing.fog_color.b.Value())
.Cast<u8>();
// Get index into fog LUT
float fog_index;
@ -778,7 +781,7 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
UpdateStencil(stencil_test.action_depth_pass);
auto dest = GetPixel(x >> 4, y >> 4);
Math::Vec4<u8> blend_output = combiner_output;
Common::Vec4<u8> blend_output = combiner_output;
if (output_merger.alphablend_enable) {
auto params = output_merger.alpha_blending;
@ -787,11 +790,11 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
FramebufferRegs::BlendFactor factor) -> u8 {
DEBUG_ASSERT(channel < 4);
const Math::Vec4<u8> blend_const =
Math::MakeVec(output_merger.blend_const.r.Value(),
output_merger.blend_const.g.Value(),
output_merger.blend_const.b.Value(),
output_merger.blend_const.a.Value())
const Common::Vec4<u8> blend_const =
Common::MakeVec(output_merger.blend_const.r.Value(),
output_merger.blend_const.g.Value(),
output_merger.blend_const.b.Value(),
output_merger.blend_const.a.Value())
.Cast<u8>();
switch (factor) {
@ -852,15 +855,15 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
return combiner_output[channel];
};
auto srcfactor = Math::MakeVec(LookupFactor(0, params.factor_source_rgb),
LookupFactor(1, params.factor_source_rgb),
LookupFactor(2, params.factor_source_rgb),
LookupFactor(3, params.factor_source_a));
auto srcfactor = Common::MakeVec(LookupFactor(0, params.factor_source_rgb),
LookupFactor(1, params.factor_source_rgb),
LookupFactor(2, params.factor_source_rgb),
LookupFactor(3, params.factor_source_a));
auto dstfactor = Math::MakeVec(LookupFactor(0, params.factor_dest_rgb),
LookupFactor(1, params.factor_dest_rgb),
LookupFactor(2, params.factor_dest_rgb),
LookupFactor(3, params.factor_dest_a));
auto dstfactor = Common::MakeVec(LookupFactor(0, params.factor_dest_rgb),
LookupFactor(1, params.factor_dest_rgb),
LookupFactor(2, params.factor_dest_rgb),
LookupFactor(3, params.factor_dest_a));
blend_output = EvaluateBlendEquation(combiner_output, srcfactor, dest, dstfactor,
params.blend_equation_rgb);
@ -869,13 +872,13 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve
.a();
} else {
blend_output =
Math::MakeVec(LogicOp(combiner_output.r(), dest.r(), output_merger.logic_op),
LogicOp(combiner_output.g(), dest.g(), output_merger.logic_op),
LogicOp(combiner_output.b(), dest.b(), output_merger.logic_op),
LogicOp(combiner_output.a(), dest.a(), output_merger.logic_op));
Common::MakeVec(LogicOp(combiner_output.r(), dest.r(), output_merger.logic_op),
LogicOp(combiner_output.g(), dest.g(), output_merger.logic_op),
LogicOp(combiner_output.b(), dest.b(), output_merger.logic_op),
LogicOp(combiner_output.a(), dest.a(), output_merger.logic_op));
}
const Math::Vec4<u8> result = {
const Common::Vec4<u8> result = {
output_merger.red_enable ? blend_output.r() : dest.r(),
output_merger.green_enable ? blend_output.g() : dest.g(),
output_merger.blue_enable ? blend_output.b() : dest.b(),

2
src/video_core/swrasterizer/rasterizer.h
View file

@ -13,7 +13,7 @@ struct Vertex : Shader::OutputVertex {
// Attributes used to store intermediate results
// position after perspective divide
Math::Vec3<float24> screenpos;
Common::Vec3<float24> screenpos;
// Linear interpolation
// factor: 0=this, 1=vtx

22
src/video_core/swrasterizer/texturing.cpp
View file

@ -51,8 +51,8 @@ int GetWrappedTexCoord(TexturingRegs::TextureConfig::WrapMode mode, int val, uns
}
};
Math::Vec3<u8> GetColorModifier(TevStageConfig::ColorModifier factor,
const Math::Vec4<u8>& values) {
Common::Vec3<u8> GetColorModifier(TevStageConfig::ColorModifier factor,
const Common::Vec4<u8>& values) {
using ColorModifier = TevStageConfig::ColorModifier;
switch (factor) {
@ -60,37 +60,37 @@ Math::Vec3<u8> GetColorModifier(TevStageConfig::ColorModifier factor,
return values.rgb();
case ColorModifier::OneMinusSourceColor:
return (Math::Vec3<u8>(255, 255, 255) - values.rgb()).Cast<u8>();
return (Common::Vec3<u8>(255, 255, 255) - values.rgb()).Cast<u8>();
case ColorModifier::SourceAlpha:
return values.aaa();
case ColorModifier::OneMinusSourceAlpha:
return (Math::Vec3<u8>(255, 255, 255) - values.aaa()).Cast<u8>();
return (Common::Vec3<u8>(255, 255, 255) - values.aaa()).Cast<u8>();
case ColorModifier::SourceRed:
return values.rrr();
case ColorModifier::OneMinusSourceRed:
return (Math::Vec3<u8>(255, 255, 255) - values.rrr()).Cast<u8>();
return (Common::Vec3<u8>(255, 255, 255) - values.rrr()).Cast<u8>();
case ColorModifier::SourceGreen:
return values.ggg();
case ColorModifier::OneMinusSourceGreen:
return (Math::Vec3<u8>(255, 255, 255) - values.ggg()).Cast<u8>();
return (Common::Vec3<u8>(255, 255, 255) - values.ggg()).Cast<u8>();
case ColorModifier::SourceBlue:
return values.bbb();
case ColorModifier::OneMinusSourceBlue:
return (Math::Vec3<u8>(255, 255, 255) - values.bbb()).Cast<u8>();
return (Common::Vec3<u8>(255, 255, 255) - values.bbb()).Cast<u8>();
}
UNREACHABLE();
};
u8 GetAlphaModifier(TevStageConfig::AlphaModifier factor, const Math::Vec4<u8>& values) {
u8 GetAlphaModifier(TevStageConfig::AlphaModifier factor, const Common::Vec4<u8>& values) {
using AlphaModifier = TevStageConfig::AlphaModifier;
switch (factor) {
@ -122,7 +122,7 @@ u8 GetAlphaModifier(TevStageConfig::AlphaModifier factor, const Math::Vec4<u8>&
UNREACHABLE();
};
Math::Vec3<u8> ColorCombine(TevStageConfig::Operation op, const Math::Vec3<u8> input[3]) {
Common::Vec3<u8> ColorCombine(TevStageConfig::Operation op, const Common::Vec3<u8> input[3]) {
using Operation = TevStageConfig::Operation;
switch (op) {
@ -144,7 +144,7 @@ Math::Vec3<u8> ColorCombine(TevStageConfig::Operation op, const Math::Vec3<u8> i
// TODO(bunnei): Verify that the color conversion from (float) 0.5f to
// (byte) 128 is correct
auto result =
input[0].Cast<int>() + input[1].Cast<int>() - Math::MakeVec<int>(128, 128, 128);
input[0].Cast<int>() + input[1].Cast<int>() - Common::MakeVec<int>(128, 128, 128);
result.r() = std::clamp<int>(result.r(), 0, 255);
result.g() = std::clamp<int>(result.g(), 0, 255);
result.b() = std::clamp<int>(result.b(), 0, 255);
@ -153,7 +153,7 @@ Math::Vec3<u8> ColorCombine(TevStageConfig::Operation op, const Math::Vec3<u8> i
case Operation::Lerp:
return ((input[0] * input[2] +
input[1] * (Math::MakeVec<u8>(255, 255, 255) - input[2]).Cast<u8>()) /
input[1] * (Common::MakeVec<u8>(255, 255, 255) - input[2]).Cast<u8>()) /
255)
.Cast<u8>();

10
src/video_core/swrasterizer/texturing.h
View file

@ -12,14 +12,14 @@ namespace Pica::Rasterizer {
int GetWrappedTexCoord(TexturingRegs::TextureConfig::WrapMode mode, int val, unsigned size);
Math::Vec3<u8> GetColorModifier(TexturingRegs::TevStageConfig::ColorModifier factor,
const Math::Vec4<u8>& values);
Common::Vec3<u8> GetColorModifier(TexturingRegs::TevStageConfig::ColorModifier factor,
const Common::Vec4<u8>& values);
u8 GetAlphaModifier(TexturingRegs::TevStageConfig::AlphaModifier factor,
const Math::Vec4<u8>& values);
const Common::Vec4<u8>& values);
Math::Vec3<u8> ColorCombine(TexturingRegs::TevStageConfig::Operation op,
const Math::Vec3<u8> input[3]);
Common::Vec3<u8> ColorCombine(TexturingRegs::TevStageConfig::Operation op,
const Common::Vec3<u8> input[3]);
u8 AlphaCombine(TexturingRegs::TevStageConfig::Operation op, const std::array<u8, 3>& input);