David Schuyeteneer wrote:

> we should develop a way to
> DESCRIBE
> the properties, behavior and characteristics of a soundsignal and
> transmitting
> THOSE instead of the mass-data of the signal itself.......

It is such a good idea, it has already been developed in the same labs
as Csound (based on Csound)! ;)  There was a pointer to the site on
this list just the other day.

> Simple synthesis methods would be not such a problem....but...what if one
> wants
> to transmit complex signals like sampled sounds ????  Detailed
> interpolation algorithms on FFT windows ??? maybe with sublevel deviation
> description between
> successive FFT frames, to enhance the flat interpolation...

I believe this is the basis of the original MPEG audio compression.
I might be telling you stuff you already know, so forgive me, but
FFT is performed on frames of audio, and stuff stripped out which
it estimates the human ear cannot perceive - eg. less resolution
on extreme frequences, masking and so on.

Good thinking anyway, and join the club of people who've had seemingly
amazing new ideas only to find they've been done already ;)

Thanks!
Ben.