ot: Voltera-Wiener nonlinear system analysis was: RE: [sdiy] ot: rotating speaker simulation or stupid approach

[Czech Martin]

Thank you for bringing this back into my attention!

It is also called the Voltera-Wiener method.
It seem to have been developed in 1970 or so, which
is a bad thing because this smells like really heavy
math.

When I understood the possibility of "correlation" based
Linear Time Invariant (LTI) system analysis (you feed in
noise and do a correlation of inout and output signal)
some years ago (thanx to Magnus, Juergen and others!!)
the next step was obvious:

-will this be extendable to time variant systems
 (which I think is possible in a straightforward way,
 but with almost absurd computational effort, it is not
 only me, but some serious papers exist since long)
-and what will happen if nonlinear effects come into play
 (I had no answer to this, now at least I have a name for
  it)

A "google" on "input output wiener volterra"
gives a lot of hits, one of them:

http://www.win.tue.nl/macsi-net/Events/Nijmeijer.pdf

So far I've not found an understandable (for my poor eng. head)
version on the net. Perhaps I find a good book.
I'm very curious to learn more about this.

I would be very greatfull for any hints to basic texts about
this topic.

Does it open a way for emulating analog nonlinear time variant circuits?
Certainly a step forward if you can describe what's really
happening.

So far it seems to me, that emulations of nonlinear systems
(e.g. guitar amps) are too simple minded and fumbling,
I never heard that any manufacturer has tried it with
Voltera-Wiener ideas. So no suprise that most emulations
fail.

m.c.

-----Original Message-----
From: Tom May [mailto:tom at tommay.net]
Sent: Donnerstag, 3. Juli 2003 19:50
To: Sdiy (E-mail)
Subject: Re: [sdiy] ot: rotating speaker simulation or stupid approach


"Czech Martin" <Martin.Czech at micronas.com> writes:

> The sound of the real Leslie has also distortion
> components and currently I have absolutely no clue
> how to measure or understand these nonlinearities from
> a black box view.

I don't know anything about this really, but it has been mentioned
here before that Volterra series are used to model nonlinearities
based on blackbox measurements.  I used to have a pointer to a
document that described how to do this, but unfortunately I can no
longer locate it.

Tom.