Chuckk Hubbard wrote:
Instead of adding more portamento for finer pitch resolution, would it be possible to do different registers separately?
I don't get it. I didn't mean to suggest adding portamento. That's a drawback of PWM: you can't switch quickly between levels and have a pure DC level at the same time: the voltage ramps up or down.
Since I intend to compose the entire piece before passing it through the Moog, and my patch handles all the tempos, it would not be a problem to record the audio in segments and line them up in Pro Tools. I suppose the wandering tuning of the synth itself will be an issue.
You could use pd to tune the synth: count the frequency and compare it to what it should be, change the CV. The Oberheim Xpander does that to overcome the limitations of its 8-bit main DAC.
I like your very last suggestion. I realized through some preliminary investigations with a multimeter that my soundcard removes DC pretty quickly. I guess your point about the 22kHz constant frequency is that this would not happen.
It still happens but it's easier to filter out the 22kHz than the 22Hz or so that PWM would need. You have to add the DC back in by passing the signal through a diode somewhere, the easiest circuit looks like this but it will be too noisy: diode in --->|----+-------out | = capacitor | gnd Basically the diode forces all the voltage to be positive relative to the average voltage held on the capacitor. With a high enough input frequency and a large enough capacitor you get a steady DC voltage. The capacitor stores charge, the larger it is the slower the output changes. If it's too small the input frequency passes through, which you don't want.
For the op-amp idea, would a signal not be sufficient? I thought [sig~ .4], for instance, would output a constant value of .4?
Probably that would work. The actual voltage you get from a soundcard is probably too low by a factor of ten so you need to amplify it somehow, usually with an op-amp.
Another thing I didn't mention is that the only computer in the same room as the Moog is a university Mac, which I won't be able to tear open. I don't yet own a working laptop, so if anything like that needs to be done, it will be me bringing my PC to school. I think this will still be easier for me than the PWM squarewave, which I still don't understand.
The PWM idea is simple: it's to average the wave. If it's 25% on, 75% off you get an average level of 25% full-scale. To get the average level you low-pass filter it at a very low frequency, which then limits how fast you can change the level. And if you're sampling at 44100Hz and need 1000 different levels you have to use a frequency of 44.1 Hz for the PWM frequency.
Martin