Casio CZ/ VZ/ FZ - Pro Series

>I didn't think it was a rumor, but a bona fide explanation of how
>the CZ synth engine worked.  I don't know if it was Casio slinging
>the bull in order to avoid a suit, but it seemed to make sense the
>way the synth engine supposedly worked--harmonic content determined
>not by frequency modulation of one wave by another but by altering
>the phase of the wave based on the rate at which waves were read out
>of ROM:

I'm actually fairly sure FM synthesis works like this too:

From what I understand (and this is based on some speculation but it seems 
right to me), it's really pretty basic stuff.

Imagine you have a digital oscillator that consists of a phase register, a 
pitch register (increments the phase register by a certain amount every 
cycle), and a sinewave lookup table (actually just a quarter-sine that gets 
flipped around 3 times).  The phase register feeds into the lookup table and 
the output is a sine wave: sin(f*t) (I'm ignoring the amplifier).  If you 
added an external signal from another oscillator to the value of the phase 
register, you'd have phase modulation, as in the DX7: sin(f*t + x).  I'm 
fairly sure all the CZ series does is use something slightly more fancy as 
the modulating signal (well, except the resonant waveforms, which use some 
kind of funky amplitude modulated sine wave sync to mimic resonance).  I 
think the DX7 only really uses one hardware oscillator for all operators and 
switches between them every clock cycle.