The issue i was running into was that changing the samplerate does so for all tracks.
So pitch-shifting with a keyboard press would have to be “trigger track, pitch shift the track”. But this means that while you could play multiple tracks, their pitch shifts are locked to the same shifts. Its also the chipmunk style pitch shifting that changes the duration of the sample.
Whats really made me question whether to go through the trouble of doing this is, im thinking the end result wouldnt necessarily be better than just triggering samples that are pre-pitch shifted.
The better way to do this might be to write some middleware that takes a sample, creates pitch shifted versions, and dumps them into banks to trigger for the ini (we already have middleware made to dump banks of drum samples, so this proof of concept works). Then, you have the arduino logic load in the banks as you play them. Theres a lot going against this idea too, since you effectively have 16 notes of buffer to play with, but its possible that you could use the 16 available tracks like 16 note poly with the usual voice drop logic.
If this last method is going to work, it will take some planning to get it manageable.