Niklas Haas

Hmm. I tried various other approaches but I'm always hitting the bottleneck that LittleCMS does not seem to support black point compensation with relative colorimetric intent at all. I might...

The author of Little-CMS [thinks your ICC profile is wrong](https://github.com/mm2/Little-CMS/issues/328#issuecomment-1207492591), which is why you're getting such bad results. I'm not sure what to do with this information. I'm still waiting...

> Sorry for interjecting but how does one get that log output for ICC information/generation of cube? I'm just using some hacky debug code ```c diff --git a/src/shaders/icc.c b/src/shaders/icc.c index...

@aufkrawall something I noticed in my own testing is that enabling `--debanding` raises the black point in a strange and noticeable way - an encoded value of `16` (i.e. reference...

@aufkrawall based on your OP settings, I believe you are applying the 3DLUT incorrectly. A correct application would be `--lut=bt1886nocal.cube --lut-type=conversion`. This instructs mpv to use this LUT (instead of...

In your OP settings you are using `--image-lut` instead of `--lut`, too. `--lut-type` only affects `--lut`, not `--image-lut` (that would be `--image-lut-type`) Edit: Although, if the rest of the color...

Random thought: mpv assumes untagged BT.1886 sources have a contrast of `1000:1` by default. So we actually do tone-mapping of the black point. Might be related, but might be not....

Not currently. I remember looking into this but I don't remember exactly what the conclusion was. Some things are unfortunately very hard to support as properties because DoVi records are...

From my impression of the IRC debugging, and the mention of extremely low sample values in the loop (as low as 4.06264450777050965202e-39), I think this article is relevant: https://en.wikipedia.org/wiki/Subnormal_number#Performance_issues I...

What do you mean by "far more annoying framedrop" when using `display-resample`? That shouldn't be dropping *any* frames!