Malcolm Bechard

The descriptor only has an entry for binding 0, which is a float texture. There is nothing specified for binding 1, so it's left as-is in the Metal argument table....

Here is the shader that is used beforehand `fragment main0_out main0(main0_in in [[stage_in]], texture2d uTDCurveTexture [[texture(0)]], texture2d uTDBandTexture [[texture(1)]], sampler uTDCurveTextureSmplr [[sampler(0)]], sampler uTDBandTextureSmplr [[sampler(1)]])` The shader with this declaration...

This does not happen in 1.3.275, but occurs in everything after that.

Great, thanks!

Thanks for raising this. It seems to me the high level summary should be: Matching should be done by location if locations are specified, and by name if they are...

Is this fixed by? https://github.com/shader-slang/slang/issues/4026 Has that been brought into MoltenVK?

It's exactly the values from the AP1 columns in the table in: https://en.wikipedia.org/wiki/Academy_Color_Encoding_System#ACES_Color_Spaces So redx = 71300, redy = 29300, so: 29300 > 100000 - 71300, and the check fails....

I think here is the Official doc page for it: https://docs.acescentral.com/specifications/acescg/#color-space The python link does work if you change the branch to main, but ya the link is busted pointing...

> But you still haven't answered my question; what CIEXYZ endpoints are you using? So far I've not used CIEXYZ endpoints directly in my work incorporating colorspace awareness, and only...

Thanks so much for the detailed response! Yeah with ACES only really defining EXR and MXF as formats to hold the data, it seems like there is value is supporting...