Mip mapping uses an incorrect downscaling algorithm

[HybridDog]

missing gamma correction: http://www.ericbrasseur.org/gamma.html?i=1
on mipmaps: https://paroj.github.io/gltut/Texturing/Tut16%20Mipmaps%20and%20Linearity.html
I noticed that mipmapping ignores gamma when downscaling, you can test it by using the gamma dalai lama as some texture:

Note that you likely need to watch the image in 1:1 size because browsers also use incorrect downscaling.

Here's another example where you can see the impact: A black white chess board pattern looks (and is) brighter than the texture downscaled by the mipmapping algorithm:

Please take care to view this screenshot in 1:1, too.

[numberZero]

Correctly, gamma should be applied after rendering (or as last operation in fragment shader). But currently it is effectively applied on meshgen, that is, on a per-vertex basis.

[SmallJoker]

@HybridDog How is this issue different to #6846? Both are reporting the suboptimal mipmap downscaling algorithm.

[HybridDog]

This issue is about missing gamma correction and the other one is about the alpha channel problem.

Both are reporting the suboptimal mipmap downscaling algorithm.

l think so, too.

[paramat]

Duplicate of #6846