[sdiy] Generating a large number of CV outputs

[Tom Wiltshire]

> On 12 Dec 2023, at 14:51, David Kantowitz via Synth-diy <synth-diy at synth-diy.org> wrote:
> 
> For a practical converter, as Tom wrote, you really need to not use the
> top/bottom ~10% and rescale.

Yes. For audio, this would be a good idea, and this is really pretty easy to do. After all, for a signal that mostly hangs around the centre line, we only rarely go to the far extremes, and we can simply turn the volume down to avoid them completely.
For my StompLFO firmware, which uses a PDM output, I was able to go closer to the extremes, since the signal bandwidth (a ramp wave at 25Hz is the worst) is so far away from the clock frequency (2MHz, iirc) that the filtering can deal with the rest.

I was interested to see that the bit-reversed PWM has the same property - the top and bottom become the worst cases, and the centre is the best, quite unlike standard PWM.

> Someone else asked about errors from edge shape & jitter.  I'm guessing
> jitter doesn't matter much at the expected oversampling rates of 64 to 256.
> Especially if that jitter is random.

For the NCO-carry case, the jitter very much *isn't* random, but creates a long pattern (and hence a low frequency). Adding genuine random jitter to wobble the pulses about a bit and break up the repetition can help spread the noise out more. This is all related to "NCO frequency spurs".

Tom