SV: Re: [sdiy] analogue physical modelling organ dream

[karl dalen]

--- "JH." <jhaible at debitel.net> skrev:

> > I also have quite difficult to get the resonator to track
> > KBD CV, im thinking of using a small micro to handle this.
> 
> Yes, unless everything (all the filters included) is perfectly tracking
> 1V/oct,
> we'll always have tuning problems.

I recon that 16 bits of DAC resolution would be enough
for the limited frequency range of a typical bore/tube based
instrument. Even string based could be done with 16 bits,
if 16 bits are very good for the average modular with VCO
ranges of 10hz to 18-20Khz it must do for the more limited 
range of PYMO's!.....I suppose!
 
> OTOH, some of the filters should not be tracking, in order to change
> the tone across the keyboard range.

On the clavia nord modular pages you can find some indepth
descriptions of the tuning problem the digital folks have
with the delay based resonator. 

> On a mechanical instrument, you just place a finger hole slightly
> different from where the "ideal" location would be, for instance.

Exactly, this is a delicate issue since this little missplacememnt
directly translates to the tiny changes of the PYMO model wich makes
it delicate to play, just as the real mechanical instrument.
> 
> So you need an autotune routine which plays _all_ notes of the
> PM instrument (at some nominal excitation level), and creates
> a lookup table to correct the CV accordingly.
> 
> Can't be too difficult.

Extreamly easy to do!

> Personally, I can't do this, as I'm not
> into uP programming. If you, or anybody else, would implement
> this and make it available, that would be a major breakthru for
> synth-diy, IMO.

But i wonder wheter this are applieable to all PYMO models?
I have a bad feeling that each model needs its own table lookup
for tuning, this isnt a problem when it comes to MCU memory per see,
but the routine to find the accurate notes for each model could be a 
razz, also what if the model becomes complex!?

At the moment i know to little about these issues!
 
> Cause as you wrote, you can get the most astounding sounds,
> and the most sensitive (to controllers) instruments by just
> combining delays, filters and nonlinearities - if it wasn't
> for the tuning issues.

However,you could do the tapped delay type of tuning,
by tapp off the delay at several points would equal 
the same thing as the diffrent holes in the bore 
in a flute for instance. Wich then starts the question
of how" exact" would the tapps be in the delay steps!?

Also there are no BBD's that alow this!!

KD