Unison Detune?

[CutRock]

I was wondering if someone here might be able to provide me some insight into how detune is usually handled.

I'm sort of in the process of mapping out what modules I want to make for my first project, and I'm sort of hung up on how to handle unison detune. I'm already planning on having a keyboard routed into a MCU to allocate voices to different oscillators, but I want to be able to stack and detune oscillators as well. Essentially, my question is: Is detune handled by the MCU, giving the oscillators a distribution of detuned pitches, or is there some sort of analog circuitry involved (and if this is the case, I would really appreciate a link to an example).

Thanks

[edit] Also - is portamento usually handled by a digital voice allocation chip?

[lordavon]

the short answer is, it depends.

based on my personal experience, if you're using analogue oscillators, then you get detune naturally. If you're using digital you'll have to detune them that way.

As for the portamento, again it depends, some synths use a single CPU (PPG for example), so there's no seperate voice allocation chip.
Also if you're using analogue oscillators, you need only a VCSlew generator to give portamento.

If you're building your own synth, it's probably best to break down into small parts and decide where you feel things are best placed.

Paul

[JovianPyx]

In a digital realm, I handle portamento by passing the pitch data through a single pole lowpass IIR filter. This method can require the use of wide arithmetic to make sure the pitch stays accurate. Too narrow an arithmetic width can result in quantization of the pitch which makes it sound out of tune.