Trying to design a polyphonic voice selector (like analog polysynths had)

[slikmanp]

Hey everyone,

I've been trying to come up with some sort of key scanner or voice selector to recreate the way early polysynths would have worked. I want to try to mimic the way an analog poly would have had a subtly different sound in each voice. Also, I can't thing of any way to make an arpeggiator if you don't know all of the notes you're playing at the same time.

So far I've been able to freeze and isolate the note information inside the MIDI Multi Note for each note that I play using a bufferizer to hold the number and various comparison modules to dictate which output of the multi note gets which signal.

It kind of works but breaks down as soon as you start releasing and re-triggering notes. It also totally fails if you try to play a chord.

I'm wondering if anyone has tried to tackle this problem?

As far as I can tell I would either need:
- a multi note module in which I could define exactly what's going on in each voice layer (like the piecewise circuit).
- the ability to define I/O by which voice layer they are interacting with (input/output 0 = v1, input/output 1 = v2, etc).
- some way of making a voice layer self aware (voice 4 "knows" that it's voice 4 and can make comparison's based on that knowledge).

I'm pretty sure that all of these would require digging into the source code which is something that I'm not up to.

Any thoughts/suggestions?

Thanks,

Stuart.