mirror of
https://github.com/MaSzyna-EU07/maszyna.git
synced 2026-11-19 01:46:01 +01:00
fix(render): alpha to coverage only where it is on, full coverage on flat alpha
Review of the alpha to coverage change for alpha tested surfaces: - An alpha which is flat over the pixel (fwidth 0) and equal to the opacity threshold got a coverage of 0.5, so e.g. opacity 1 on an opaque or DXT1 texture turned half see-through with multisampling. Without an alpha edge in the pixel the plain test decides now: step(opacity, alpha). - The shaders took the coverage path whenever they were built with multisampling, also where alpha to coverage is off (the reflections pass) or does nothing (the single-sampled model previews), and grew the cut-outs there by about half a pixel. model_ubo gets an alpha_to_coverage flag, set together with the GL state, and the shaders keep the plain alpha test without it. The previews no longer turn alpha to coverage on. The flag takes 4 of the 12 padding bytes after shadow_tone, so the std140 layout and the size of the block stay the same. - The shader validator never set iMultisampling, so it never compiled the alpha to coverage branches. Its "features on" configuration uses the default of 2 now, on desktop GL and GLES.
This commit is contained in:
@@ -65,6 +65,7 @@ void gl::glsl_common_setup()
|
||||
float fog_density;
|
||||
float alpha_mult;
|
||||
float shadow_tone;
|
||||
float alpha_to_coverage;
|
||||
};
|
||||
|
||||
layout (std140) uniform scene_ubo
|
||||
@@ -117,12 +118,19 @@ void gl::glsl_common_setup()
|
||||
// multisampling; otherwise the alpha test keeps or drops whole pixels and its edges stay
|
||||
// jagged whatever the multisampling. the alpha is sharpened to go from 0 to 1 over about
|
||||
// a pixel around the test threshold, so the edge keeps the place and sharpness the alpha
|
||||
// test gave it. computed before any discard, which leaves the derivatives undefined
|
||||
// test gave it. computed before any discard, which leaves the derivatives undefined.
|
||||
// alpha_to_coverage of model_ubo tells whether alpha to coverage is on for this draw; where
|
||||
// it isn't (reflections, previews) the shaders keep the plain alpha test
|
||||
float alpha_coverage(float alpha)
|
||||
{
|
||||
#if ALPHA_TO_COVERAGE_ENABLED
|
||||
if (opacity > 0.0)
|
||||
return clamp((alpha - opacity) / max(fwidth(alpha), 0.0001) + 0.5, 0.0, 1.0);
|
||||
if (opacity > 0.0 && alpha_to_coverage > 0.0)
|
||||
{
|
||||
// a flat alpha has no edge in the pixel and gets the plain test, so an alpha equal
|
||||
// to the threshold stays fully covered
|
||||
float fw = fwidth(alpha);
|
||||
return (fw > 0.0 ? clamp((alpha - opacity) / fw + 0.5, 0.0, 1.0) : step(opacity, alpha));
|
||||
}
|
||||
#endif
|
||||
return alpha;
|
||||
}
|
||||
@@ -133,7 +141,7 @@ void gl::glsl_common_setup()
|
||||
bool alpha_test_fail(float alpha, float coverage)
|
||||
{
|
||||
#if ALPHA_TO_COVERAGE_ENABLED
|
||||
if (opacity > 0.0)
|
||||
if (opacity > 0.0 && alpha_to_coverage > 0.0)
|
||||
return coverage <= 0.0;
|
||||
#endif
|
||||
return (opacity >= 0.0 ? (alpha < opacity) : (alpha >= -opacity));
|
||||
|
||||
Reference in New Issue
Block a user