Harmonic Bias Source
2000-04-15 by Tkacs, Ken
It's true that to work with "true" additive, you need to be able to specify the phases of the partials, but I'm figuring you'll never get that close with analog. Still I think you can get damned close (the ear isn't real sensitive to phase in general), *provided* that any sounds you mix in to be pseudo-partials of your fundamental oscillator are sines. That's why you can't make a complex divider-based module for this---you get square waves instead of sines and the ear will always hear them as separate pitches because each has its own harmonic series. You could devise some kind of divider-based pseudo-additive module if you could find a good square-to-sine wavehaping circuit (hopefully cheap & simple because you'd need a bunch of them). In all my wanderings, I've only found one square-to-sine converter (usually sines are derived from the more similar triangle waves), and that was in a data sheet for a shift register... wasn't a real clean sine, I think, and it required two chips and 16 precision resistors per converter, which starts to get to be a lot.