[sdiy] Multi-voice architecture without CV parameters: mechanical replication of settings?

[cheater cheater]

On Tue, Apr 6, 2010 at 13:05,  <jhaible at debitel.net> wrote:
>> Welp, the technology has progressed, we have CV and other stuff now -
>> but the thing is, with adding CV capability we have also changed the
>> sound. I guess this is for people who like the old sound :-) For one
>> thing the new SEM specifically kept everything non-CV-able.
>
> I always loved the concept of the Oberheim Four-Voice-Synthesizer, even though I
> never owned an original one. The ability to play 4 slightly different sounds
> from one keyboard is fascinating, especially with that amount of control
> (different key assignment strategies that can be changed on the fly).

Yep, that's very endearing!

> Can't see any benefit in mechanical parameter tracking or the like. Either ditch
> the tracking altogether, or use VCAs and control them with a sum of a local
> parameter and a global parameter, each. Lots of pots, lots of fun.

Well, sometimes you might want to play your polysynth like a
'standard' polysynth - in that case swap the panels and you've got a
completely different instrument! A different user interface *does*
inspire new sounds, sometimes even very different from what you would
do 'normally' imo :-)

Colin,
> I don't see any major reason why you could't use one or two fully floating
> VCRs to replace any given pot.
> That should allow you to make a voice voltage controllable without affecting
> the sound.

I thought so as well, and I asked Tom Oberheim about this, and he said
he's specifically left out voltage control on the SEM to improve the
quality of the sound. (Of course, the quality of sound is a very
subjective matter). He said that he tried going down the VC path and
does not recommend it. Here it is, from the mouth of the SEM guy.

After some thought it is undeniable that two different circuits *will*
at some points sound different, it can be difficult to notice in some
instances and easier in others. Something like filter resonance is
where I would be looking for an easy to spot difference between VC
resistor and real resistor. And a little difference can go a long way
in synths, where even different *layout* of a pcb can make a
difference to a VCO. Besides, if it's possible, and someone wants to
do this in their time, then why not? ;-)

Cheers
D.