I don't know where they get the bit count from, it seems to be based upon an understanding of the cone's operation that I've never seen justified in any of thes

I don't know where they get the bit count from, it seems to be based upon an understanding of the cone's operation that I've never seen justified in any of these articles. Assuming that the cones of different frequency range have approximately the same degree of sensitivity, I'd be thinking that the number is about n^(4/3) where n is the number of colours perceivable by trichromats. But even if that's true - which I doubt - the additional colours are in unlikely to have a useful distribution, and will be centered around the frequency response curve of the fourth cone (I believe it's generally in the yellow spectrum; somewhere between red and green).