No idea what this would sound like...

[earthtonesaudio]

But I want to try:

http://www.audioscientific.com/28_Simple_Sine_wave_generator_has_no_filters_.jpg

Replace the input clock signal with clock whose frequency is modulated by an audio signal.   Maybe tweak the resistor values a bit.

Oh, and if you wanted a better sample rate, you could use more of the outputs from the counter, increase the input frequency, and add additional XOR gates.


EDIT: and perhaps something added to the output to make the gain track the input signal envelope.

[Taylor]

Well, maybe I'm missing something here, but it will sound like a sine wave, no? All sines sound the same, or they aren't sines.

The rest, regarding audio tracking, is a great idea, but as you probably know with audio-tracking oscillators, it's not quite so simple as it seems to get it to work right.

[tommy.genes]

Basically, you feed that circuit a square wave (clock), and it gives you a sine wave at ¹/₃₂ the frequency of the input square wave - five octaves below the original. To use this as a "fundamental extractor," you'd have to first convert your instrument signal into a clean square wave at 32x the fundamental frequency. Those may be the details in which the Devil is hiding in this case, but that doesn't mean you shouldn't try it to see what happens. Please report your results if you do.

-- T. G. --

[earthtonesaudio]

Haha, yes I suppose it would sound rather like a sine...

I think I was a little unclear on my first post, I meant to replace the clock with a high-frequency FM signal, the carrier being HF and the modulating source being the audio.  But having thought about it a little more, it's not much different from a sine FM oscillator. 

This may have some utility as a carrier generator for a ring modulator though.

[synthmonger]

The output waveform is a stepped sine wave. Basically a rounded pyramid waveform. Your best bet is using a fundamental extractor that is driving a CD4046 and 4017 setup as a signal multiplier that then clocks the supposed 'sine wave generator'. Google around for the PLL (4046 + 4017) multiplier. I can't remember where I found it.

[R.G.]

The output waveform is a stepped sine wave. Basically a rounded pyramid waveform. Your best bet is using a fundamental extractor that is driving a CD4046 and 4017 setup as a signal multiplier that then clocks the supposed 'sine wave generator'. Google around for the PLL (4046 + 4017) multiplier. I can't remember where I found it.

It's a standard trick. Phase lock with multiply then feed a dividing waveform generator. The circuit for the sine generation is at GEO, kiped from The CMOS Cookbook.

The real trick with this is to get a good signal to lock to. The CD4046 can be fussy about input signals, so some conditioning may be needed.