RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-27 04:57:45 +00:00
Code Issues Activity Actions

LibGfx: Avoid float->double->float when converting from linear to gamma

Browse source 
Benchmark shows 5x speedup (from 644ms to 110ms).
This commit is contained in:
Oleg Sikorskiy 2021-04-04 16:19:47 +03:00 • committed by Andreas Kling
parent c0eacf4cc1
commit b34f194168
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
Userland/Libraries/LibGfx/Gamma.h
View file

@ -106,7 +106,7 @@ inline float linear_to_gamma(float x)
constexpr float a = 0.00279491; constexpr float a = 0.00279491;
constexpr float b = 1.15907984; constexpr float b = 1.15907984;
float c = (b / sqrt(1 + a)) - 1; float c = (b / sqrt(1 + a)) - 1;
return ((b / __builtin_sqrt(x + a)) - c) * x; return ((b / __builtin_sqrtf(x + a)) - c) * x;
} }
// Linearize v1 and v2, lerp them by mix factor, then convert back. // Linearize v1 and v2, lerp them by mix factor, then convert back.