Revision as of 10:21, 28 December 2021 by Vegard (talk | contribs) (add link)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Quick rules of thumb, since I can never remember which is which:

  • RGB = linear space
  • sRGB = gamma space
  • most image files store sRGB
  • sRGB is unfit for blending, blurring, etc.

The corollary is:

  • Image files loaded as textures should be converted from sRGB to RGB (but some decoders do this for you!)
  • Since alpha is linear, alpha premultiplication should happen in linear space, i.e. after converting to RGB [1]
  • The shader responsible for putting the final pixels on the screen should have a gamma correction step doing:
const float gamma = 2.2;

out vec4 color;

void main()
    // ...

    color = pow(color, 1. / gamma);

Note, however, that OpenGL may also perform this step for you on any on-screen framebuffers (i.e. the default framebuffer) if you do glEnable(GL_FRAMEBUFFER_SRGB).

See also