Feature: LMS Color space

[EdenEast]

You have cvd filters that use the lms color space but there is no lms color space that can be converted to.

Also looking at the calc_cvd_matrix where are you getting the matrix that converts xyz to lms?

coloraide/tools/calc_cvd_matrices.py

Lines 15 to 19 in c33df67

	M = [
	[0.15514, 0.54312, -0.03286],
	[-0.15514, 0.45684, 0.03286],
	[0, 0, 0.01608]
	]

[facelessuser]

Smith & Pokorny (1975) 2-deg cone fundamentals. This is the transition matrix defined by Smith and Pokorny.

[facelessuser]

@EdenEast If you really want to convert to LMS using this, you can actually use a plugin to do so. Here it is working live: https://facelessuser.github.io/coloraide/playground/?source=https%3A%2F%2Fgist.githubusercontent.com%2Ffacelessuser%2Fafd14b9425b996c6f91b30a8312488bc%2Fraw%2F88f5b978870cefaf4d75b5be425f0aaae96589ba%2Flms.py

If you 'd like to see the source, just click edit. The gist is also here: https://gist.github.com/facelessuser/afd14b9425b996c6f91b30a8312488bc

[EdenEast]

Thanks for you answers, they have helped me with my understanding of this space. So if my understanding is correct, you are using Smith and Pokorny which requires that the srgb-linear to xyz has been Judd-Vos corrected instead of the standard cie. Does that mean that the base xyz is not the standard defined by cie1931 but is instead Judd-Vos corrected?

[facelessuser]

Actually no...we are using CIE XYZ...I believe we are ignoring the Judd-Vos corrections 🤦🏻. So, as far as CVD is concerned, we have a bug and need to apply that correction.

I've applied the correction to the gist. Live version here: https://facelessuser.github.io/coloraide/playground/?source=https%3A%2F%2Fgist.githubusercontent.com%2Ffacelessuser%2Fafd14b9425b996c6f91b30a8312488bc%2Fraw%2F3f1cca6d10acabfccfd00cfcac0bfb94c6121a77%2Flms.py

We are using the correction matrix found here: https://vision.psychol.cam.ac.uk/jdmollon/papers/colourmaps.pdf

I'll need to update related matrices in ColorAide. I'm glad you were asking these questions. I knew I had read that, but overlooked it in the implementation.

[facelessuser]

Keep in mind, we are mainly targeting LMS with this approach mainly for CVD as previous studies mainly used this as well.

[facelessuser]

Related fix #283