Implement sRGB texture sampling.
Previously sRGB data was converted to linear space on upload. This
caused a loss of precision. This change performs the conversion after
texel lookup. Note that we had a code path for performing the
conversion after filtering, but that leads to failures in dEQP and
unacceptable darkening between texels.
Also, glTexSubImage calls can update sRGB textures using a format/type
combination with no indication of the color space, which caused an
unintentional conversion on upload. Likewise we were missing support
for an A2B10G10R10UI implementation format.
Change-Id: Ib10845f628fb2d1849e88d7a9350868cdec32fa2
Reviewed-on: https://swiftshader-review.googlesource.com/15068
Reviewed-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
diff --git a/src/Shader/SamplerCore.cpp b/src/Shader/SamplerCore.cpp
index a50c1e8..8aebbd8 100644
--- a/src/Shader/SamplerCore.cpp
+++ b/src/Shader/SamplerCore.cpp
@@ -133,46 +133,23 @@
if(fixed12 && !hasFloatTexture())
{
- if(has16bitTextureFormat())
+ if(state.textureFormat == FORMAT_R5G6B5)
{
- switch(state.textureFormat)
- {
- case FORMAT_R5G6B5:
- if(state.sRGB)
- {
- sRGBtoLinear16_5_12(c.x);
- sRGBtoLinear16_6_12(c.y);
- sRGBtoLinear16_5_12(c.z);
- }
- else
- {
- c.x = MulHigh(As<UShort4>(c.x), UShort4(0x10000000 / 0xF800));
- c.y = MulHigh(As<UShort4>(c.y), UShort4(0x10000000 / 0xFC00));
- c.z = MulHigh(As<UShort4>(c.z), UShort4(0x10000000 / 0xF800));
- }
- break;
- default:
- ASSERT(false);
- }
+ c.x = MulHigh(As<UShort4>(c.x), UShort4(0x10000000 / 0xF800));
+ c.y = MulHigh(As<UShort4>(c.y), UShort4(0x10000000 / 0xFC00));
+ c.z = MulHigh(As<UShort4>(c.z), UShort4(0x10000000 / 0xF800));
}
else
{
for(int component = 0; component < textureComponentCount(); component++)
{
- if(state.sRGB && isRGBComponent(component))
+ if(hasUnsignedTextureComponent(component))
{
- sRGBtoLinear16_8_12(c[component]); // FIXME: Perform linearization at surface level for read-only textures
+ c[component] = As<UShort4>(c[component]) >> 4;
}
else
{
- if(hasUnsignedTextureComponent(component))
- {
- c[component] = As<UShort4>(c[component]) >> 4;
- }
- else
- {
- c[component] = c[component] >> 3;
- }
+ c[component] = c[component] >> 3;
}
}
}
@@ -316,8 +293,8 @@
}
else
{
- // FIXME: YUV and sRGB are not supported by the floating point path
- bool forceFloatFiltering = state.highPrecisionFiltering && !state.sRGB && !hasYuvFormat() && (state.textureFilter != FILTER_POINT);
+ // FIXME: YUV is not supported by the floating point path
+ bool forceFloatFiltering = state.highPrecisionFiltering && !hasYuvFormat() && (state.textureFilter != FILTER_POINT);
bool seamlessCube = (state.addressingModeU == ADDRESSING_SEAMLESS);
if(hasFloatTexture() || hasUnnormalizedIntegerTexture() || forceFloatFiltering || seamlessCube) // FIXME: Mostly identical to integer sampling
{
@@ -380,52 +357,23 @@
{
Vector4s cs = sampleTexture(texture, u, v, w, q, bias, dsx, dsy, offset, function, false);
- if(has16bitTextureFormat())
+ if(state.textureFormat == FORMAT_R5G6B5)
{
- switch(state.textureFormat)
- {
- case FORMAT_R5G6B5:
- if(state.sRGB)
- {
- sRGBtoLinear16_5_12(cs.x);
- sRGBtoLinear16_6_12(cs.y);
- sRGBtoLinear16_5_12(cs.z);
-
- convertSigned12(c.x, cs.x);
- convertSigned12(c.y, cs.y);
- convertSigned12(c.z, cs.z);
- }
- else
- {
- c.x = Float4(As<UShort4>(cs.x)) * Float4(1.0f / 0xF800);
- c.y = Float4(As<UShort4>(cs.y)) * Float4(1.0f / 0xFC00);
- c.z = Float4(As<UShort4>(cs.z)) * Float4(1.0f / 0xF800);
- }
- break;
- default:
- ASSERT(false);
- }
+ c.x = Float4(As<UShort4>(cs.x)) * Float4(1.0f / 0xF800);
+ c.y = Float4(As<UShort4>(cs.y)) * Float4(1.0f / 0xFC00);
+ c.z = Float4(As<UShort4>(cs.z)) * Float4(1.0f / 0xF800);
}
else
{
for(int component = 0; component < textureComponentCount(); component++)
{
- // Normalized integer formats
- if(state.sRGB && isRGBComponent(component))
+ if(hasUnsignedTextureComponent(component))
{
- sRGBtoLinear16_8_12(cs[component]); // FIXME: Perform linearization at surface level for read-only textures
- convertSigned12(c[component], cs[component]);
+ convertUnsigned16(c[component], cs[component]);
}
else
{
- if(hasUnsignedTextureComponent(component))
- {
- convertUnsigned16(c[component], cs[component]);
- }
- else
- {
- convertSigned15(c[component], cs[component]);
- }
+ convertSigned15(c[component], cs[component]);
}
}
}
@@ -2036,6 +1984,26 @@
}
else ASSERT(false);
+ if(state.sRGB)
+ {
+ if(state.textureFormat == FORMAT_R5G6B5)
+ {
+ sRGBtoLinear16_5_16(c.x);
+ sRGBtoLinear16_6_16(c.y);
+ sRGBtoLinear16_5_16(c.z);
+ }
+ else
+ {
+ for(int i = 0; i < textureComponentCount(); i++)
+ {
+ if(isRGBComponent(i))
+ {
+ sRGBtoLinear16_8_16(c[i]);
+ }
+ }
+ }
+ }
+
return c;
}
@@ -2238,7 +2206,7 @@
bool isInteger = Surface::isNonNormalizedInteger(state.textureFormat);
int componentCount = textureComponentCount();
- for(int n = 0; n < componentCount; ++n)
+ for(int n = 0; n < componentCount; n++)
{
if(hasUnsignedTextureComponent(n))
{
@@ -2555,11 +2523,11 @@
cf = Float4(As<UShort4>(cs)) * Float4(1.0f / 0xFFFF);
}
- void SamplerCore::sRGBtoLinear16_8_12(Short4 &c)
+ void SamplerCore::sRGBtoLinear16_8_16(Short4 &c)
{
c = As<UShort4>(c) >> 8;
- Pointer<Byte> LUT = Pointer<Byte>(constants + OFFSET(Constants,sRGBtoLinear8_12));
+ Pointer<Byte> LUT = Pointer<Byte>(constants + OFFSET(Constants,sRGBtoLinear8_16));
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 0))), 0);
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 1))), 1);
@@ -2567,11 +2535,11 @@
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 3))), 3);
}
- void SamplerCore::sRGBtoLinear16_6_12(Short4 &c)
+ void SamplerCore::sRGBtoLinear16_6_16(Short4 &c)
{
c = As<UShort4>(c) >> 10;
- Pointer<Byte> LUT = Pointer<Byte>(constants + OFFSET(Constants,sRGBtoLinear6_12));
+ Pointer<Byte> LUT = Pointer<Byte>(constants + OFFSET(Constants,sRGBtoLinear6_16));
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 0))), 0);
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 1))), 1);
@@ -2579,11 +2547,11 @@
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 3))), 3);
}
- void SamplerCore::sRGBtoLinear16_5_12(Short4 &c)
+ void SamplerCore::sRGBtoLinear16_5_16(Short4 &c)
{
c = As<UShort4>(c) >> 11;
- Pointer<Byte> LUT = Pointer<Byte>(constants + OFFSET(Constants,sRGBtoLinear5_12));
+ Pointer<Byte> LUT = Pointer<Byte>(constants + OFFSET(Constants,sRGBtoLinear5_16));
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 0))), 0);
c = Insert(c, *Pointer<Short>(LUT + 2 * Int(Extract(c, 1))), 1);
