Do opponent process theories help physicalism about color?

Byrne & Hilbert (B&H) give some excellent replies to the objections to realism about color. However, the particular form of realism they propose, based on opponent processing, prompts several challenges. Why characterize a color by its tendency to produce an intermediate brain signal, rather than in terms of the final effect – either a perception or a neural substrate for it? At the level of the retina, and even of the cortex, there are processes that partly parallel the structure of color experience; but the correspondence is not exact. Must we assume that there is any place in the brain where an exact structural correspondence is found? At the level of psychophysical functioning, there is indeed opponency; but it is not clear that this gives us the kind of type-reduction that B&H want.