multiple-type Digisound VCDO thing?

[WeAreAs1 at aol.com]

a-davidh at microsoft.com (David Halliday) wrote:

<< I am asuming that you are taking a VC oscillator, running it through some
kind of address generator and clocking address lines of an EPROM - you have
waveshapes loaded in EPROM which are D/A'd and sent out. >>

Yes, that is correct.

<< Just off the top of my head, you could probably run an adder circuit
between the address lines - cycle the other adder input with a slowly varying
input from a VC Digital Low Freq Osc...  This would give you a varying phase
shift between the two ( or more ) signals. >>

Yeah, you could, but at that point you would be spending more money and
effort than you would have by simply using another 3340 as a second
voltage-controlled clock, don't you think?  Besides, it's the flexibility and
elasticity of voltage control that gives the VCDO its main appeal.  The adder
circuit idea wouldn't be nearly as flexible as plain ol' voltage control,
IMO.

<< You could carry this even further by having the DLFO clock another EPROM
with various vaveshapes in it... >>

DLFO?  You've lost me here.  Do you mean VC clock?  The clock runs at a
fairly high frequency (probably the sampling rate of the EPROM data), and is
what's used to clock the address generator.  Are you talking about using an
LFO to switch thru different wave addresses, as can be done on a Wave PPG?
 This, under voltage control, would be fun..  If this is what you mean, how
would you accomplish this?

<< Another thing is that you could use a much bigger EPROM and manually
switch the upper address bits to select different waveshapes. >>

Yep.  If the basic wave data length is short enough, you could fit quite a
few of these waves in a *big* EPROM.  Each wave would have its own manually
selected "meta-address".  Furthermore, I think it would be very interesting
if you could devise some kind of clocked/toggled meta-address, which would
switch between two user-selected meta-addresses at the wave half-cycle point.
 (Yes, I really do think this would sound good, or at least weird.  In a good
way.)  Kind of like a digital half-wave rectfier, except that the "rectified
half of the wave could be another completely different wave.

I don't think such logic would be too difficult to construct, although I'm
not quite sure how well it might integrate into the existing Digisound
circuit - I'll have to look closer at the schemo.  I have a lot of experience
with modifying the old Oberheim DMX drum machine voice cards to do this kind
of stuff, and those voices were kind of similar to the VCDO idea (except that
they were triggered one-shot ROM players, not continuously looped wave
players).

I'd still like to know about:

<< is there an accepted standard for VCDO wavetable sampling rate, source
wave pitch, and sample-loop byte length?  What are these specs on the
Digisound? >>