Fix sRGB exponent in Metal rendering

[exavi]

This tiny change should 'fix' the sRGB encoding used on the metal backend.

I noticed this while using the MaterialXViewer in a CI setup when generating images where no single pixel matched against reference values (not even the background color :D)

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: exavi / name: exavi (e56a0c0)

[jstone-lucasfilm]

Thanks for this PR, @exavi, and since this code goes all the way back to the original integration of Metal support in 2023, I'm CC'ing @Morteeza, @lkerley, and @dgovil for their domain expertise.

[doug-walker]

The proposed change seems correct, the exponent for the sRGB formula should use 2.4 rather than 2.2.

[dgovil]

I agree this looks like the change is correct, as far as I understand gamma approximations anyway. I will ask Lee to check me on that but I think its a good catch

[ld-kerley]

Looks good to me too - I trust Doug for color stuff more than myself :).

I am a little surprised we're not just handling this conversion natively in Metal, but thats a possible improvement for the future.

[jstone-lucasfilm]

Great, thanks to all of our reviewers!

@exavi Before we can merge this work, we'll need you to address the EasyCLA errors, which you can do by following the steps listed here:

#2846 (comment)

[exavi]

Awesome! I’ll sort it out when I get back to the office!