[sdiy] dumb wavetable question

[Tom Wiltshire]

David,

For me personally, definitely (1). A wavetable oscillator can have an arbitrary waveshape. That's a *massive* improvement over an analogue VCO, or it would be if it had no downside. In practice, it's more of a tradeoff than that.

The analogue VCO and wavetable idea has been tried in a few places. Neil J mentioned one, and the Jomox Sunsyn included "RCOs" (Ramp-Controlled Oscillators) which are the idea that Olivier mentioned, where a ramp drives a ADC which drives a wavetable.

In many ways I agree that wavetable playback is trivial, at least in theory. The fun starts when you get to practical implementation, and start banging up against all the tradeoffs and differing approaches which have kept us all busy these past few days.

Regards,
Tom


On 5 Feb 2011, at 21:38, David G. Dixon wrote:

> All this talk about wavetables (which, I must confess, I've only skimmed for
> lack of understanding most of what is being said) has left me with a few
> questions.
> 
> What is the big draw of wavetable synthesis?  Is it:
> 
> 1) to gain access to an infinitude of waveshape possibilities?
> 2) to overcome certain limitations inherent to analog VCOs?
> 3) to render MIDI/computer control more convenient?
> 4) all of the above?
> 5) some other reason?
> 
> It seems like the most complicated thing is to get the wavetable to play
> exactly once within one period of the desired oscillation frequency, and the
> wavetable part itself is in fact pretty trivial.  This leads to my real (and
> potentially stupid) question:
> 
> Why not clock the wavetable with an analog clock driven by a conventional
> expo converter?  Of course, the tracking would only be as good as the
> exponential converter, but the wavetable could now be clocked with a plain
> old binary counter, and the frequency would always be equal to the analog
> clock frequency divided by the number of wavetable samples.  Since it's
> analog, it could be glided and detuned smoothly, just like any other VCO.
> Of course, you'd still need a MIDI-to-CV converter to control it digitally,
> but that could be built in.
> 
> You'd need a current-controlled clock capable of top speeds of about 5MHz
> (assuming wavetables with 256 samples) to cover the audio range.  Is this
> doable?
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy