On the other end with yellow+blue I can make lots of different shades of green, but for that they want you to buy the Berlin series. I want to play Yellow A and Blue B flat.That's what I call chromatic choice...
Absolutely. Plus, getting in and mixing your own colours is very often the fun part (and finger painting is arguably one of the better metaphors how I personally approach composition
)
But there's also a point at which the colour metaphor breaks down, in that hearing involves a hierarchy of perceptual and cognitive dimensions, different from visual perception of colour. 'Blending' of sound occurs in our minds in multiple ways - harmonically, texturally (as in foreground and background textural layers etc), as well as a lower level of perceptual blending of individual overtones. You can get sounds to blend onto a single layer, or into a single chord without actually invoking the lower level perceptual effects in which the overruns are fused into a single perceptual process.
So, for instance, in a complex 4 part voice leading where perceptual independence of the lines is paramount, if you have two different instruments playing a single line, then even small differences in, for instance, the vibrato might conceivably break the perceptual independence of that line, causing your mind to scramble to try to interpret a 5 part voice leading, and since 4 part voice leading already has human perception push right to it's edge (*) his could break the voice leading for all but the most attentive listeners (**). Or maybe, for a listener failing to grasp a 5 part voice leading instead, this out of sync vibrato might then just break the voice leading altogether. It's a bit like the collapse of the wave function in quantum mechanism (if that's a helpful metaphor).
So a lot of the time when I really enjoy the finger-paining like process of mixing colours into sounds, its about them much easier colours on textural layers a lot more often than it is about separate perceptual streams. Sometime its possible to blend within a signal perceptual layers, but as the number of independent parts in the voice leading goes up, it gets a lot harder, very quickly.
Anyway, I am not nearly a sophisticated enough orchestrator to predict how amazing or not amazing this perceptual fusion based approach to orchestration of Ark4 will be. Just saying that there's a good theoretical basis on which to argue that it might be a really fun thing to try.
Hopefully I'll get the chance to play with it someday.
(*) I've always suspected that by they time he got to 5 and 6 part counterpoint, Bach was really just showing off. Modern perceptual science suggests that my suspicion was not unreasonably, in that 3 independent perceptual sound streams is normal for the human mind. 4 is entirely possible, but much hard to achieve. 5 is apparently not impossible, but probably going to take Bach-like powers of voice leading and an extremely talented and attentive listener. 6, well I'm pretty sure that's just showing off.
(**) All of this assumes we're orchestrating for an audience of humans of course. If you're orchestrating for, say, dolphins, then I'd expect their perceptual capacities would open up entirely new possibilities - and difficulties - in orchestration. Not that there's been a lot written on how to orchestrate for dolphins yet, so far as I know.