Soufiane KHIAT

@soufianekhiat.bsky.social

Lead Rendering Software Engineer At @EA - @CriterionGames, ex @Unity3d, ex @UbisoftMTL All expressed opinions are my own

Compressing with BC1 with custom RGB ({R,G,B,white point} xy-coordinate, "gamma") seem to be profitable Bench: sRGB -> Compute custom RGB -> Compress BC1 -> Uncompress -> Inverse Custom RGB to sRGB -> Measure Error vs Original Result: +2% to -18% size of the BC1 compressed.

Bild

Illustration of "ideas are free" Discussing with @troy_s, now I'm convience that all algorithms to "process" "out of gamut" are wrong, working on it, but not trivial, here an illustration of all iterations without explaination!

BildBildBildBild

Do you want an 340000x, 2500x, 96x in your code? I was playing with @blosc.org in fact that is simple: Data -> Filter -> Codec (compressed) -> Codec_inv -> Filter_inv -> Data. What if we can do operations in Filter-Space without decoding it? Let's do a simple box-blur 1/11

A chaque vague de chaleur, ils disent "c'est normal, c'est l'été". Au mois de mai, ce n'est pas l'été : totalement impossibles (en mai) avant, désormais possibles dans un monde à 2°C, et probablement courantes dans une France à +4°C, elles sont plus précoces, plus intenses, plus longues.

On comptabilise 66 vagues de chaleur en France depuis 1900 : toutes situées après la mi-juin. Sauf probablemen celle qui arrive. 
Graphique de Serge Zaka.

Qaws (قوس) Arabic for *arc*, *bow*, *curve*. A dependency-free C11 library for creating, evaluating, sampling, traversing, and inspecting parametric curves and surfaces in 2D and 3D. A dependency-free C11 library for parametric curves and surfaces in 2D/3D. 1/13

Bild

Introducing Alwan (ألوان): A pure C colour science library 100+ RGB spaces, 30+ colour spaces, ACES 1.x & 2.0, spectral processing, CAMs, ... Zero dependencies. Unit tested against colour-science f32 or f64 floats. Looking for reviewers before public release (DM for repo). 1/15

I've been reading about color science for my book and I wanted to verify the claim that Oklab gives better linear gradients than the otherwise excellent "CAM16-UCS" color space. And yes, its gradient obviously too desaturated in this test. Oklab's ancestor "IPT" also works fine.

A blue-to-white gradient in the sRGB, CIELAB, CAM16-UCS, Oklab, IPT, and Optimized IPT color spaces. The CIELAB and CAM16-UCS results are clearly off. The gamma-space sRGB interpolation is also slightly too purple.