adsr and lfo abolish

[John Speth]

On Wednesday, April 07, 1999 8:36 PM, Gene Zumchak [SMTP:zumchak at cerg.com] 
wrote:
>     Digital solutions have not eliminated analog approaches because they
> tend to be too perfect, too canned.  One FM synthesis keyboard sounds
> exactly the same as another of the same model.  But two Moogs never
> sound alike because of the idiots turning the pots and the impreciseness
> of the analog circuits (although they are getting better).  I remember
> working with avant garde musicians in Moog's original studio in his shop
> in Trumansburg, NY in the late sixties.  Sometimes by sheer luck they
> would come up with some truly amazing sounds.  Every patch, every knob
> location was carefully documented.  A week later when the patch was
> duplicated, the sound was nowhere to be found.  It was lost forever.
> That was and is the romance of analog synths.
>
>     Still, as a digital (microcomputer) person,  I have the urge to
> control the analog pieces with digital.  There is a lot of interesting

We consistently hear about how non-varying CPU controlled instruments 
sound.  So I'd like to field a question relating to this...

Isn't there a way to model the variations of analog circuitry in a 
sufficiently powered and programmed CPU to "inject" it into the sound 
creation process in real time.  It seems to me that analog component 
variations can be characterized.  Then when the waveforms are calculated, 
the program could vary whatever parameters are needed to make it sound 
analog.

Has anybody tried this?  Is there any good reasons why this can't work?

JJS