[music-dsp] DSP vs analog circuits

[Theo]

Hi James et al,

Please excuse my ignorance,
but could someone explain to me why using feedback is so difficult?
Something tells me that the problem may be caused by sequenced updating of
the variables.

To explain my way of thinking:
Switched capacitor circuits are sampling systems too, however using feedback
causes no trouble.
The switched cap. circuit is "updated" as a whole during the on time of the
capacitor.
In DSPs however its more natural to update all variables one after an other.
The obvious but overlooked solution may be to switch to a parallel updating
algorithm.
A similar solution worked for NN structures1) so why not for DSP?
Probably I am completely off here, so I hope someone will set me straight.

Regards,
Theo

1) see M. Drossaers. Little Linguistic Creatures. Neslia Paniculata,
Enschede, 1995.



From: James Chandler Jr <jchandjr at bellsouth.net>

> >>> Umm, guys... they allready do this....Digital filtering...  Active
> filters
> > are made with opamps, caps and resistors mostly.  They can do that in
> > digital quite efficiently and more effectively than most analogue
filters
> > now adays.
>
> Hi, Jared
>
> Yeah, there are usually equivalent ways to get analog tasks done in DSP,
> though the strategies are often quite different.
>
> Was just pointing out that opamp (or feedback in general) is a very useful
> trick that is pretty difficult in DSP.
>
> EQ is one of many cases... Using a bank of parallel analog filters in the
> feedback loop of an opamp, it is quite easy to get symmetrical constant-Q
> Boost-Cut behavior.
>
> In the closest DSP EQ equivalent of a parallel filter bank, the Cut
> percentage bandwidth is narrower than the Boost percentage bandwidth.
>
> One can get symmetrical Boost-Cut DSP EQ behavior with a bank of series
> peaking filters, however. Just a different approach to the problem.
>
> Some DSP equivalents seem more problematic. A really useful analog
feedback
> technique, for steeply supressing harmonics on a signal-- Make the
> equivalent of a backend-sensing compressor, except you substitute a steep
> tracking lowpass filter for the usual wide-band tracking gain element.
>
> I've tried this tracking filter trick in DSP, and the result is not as
> satisfactory (at best marginally useful). A smarter programmer could
> probably get it to work better (G).
>
> James Chandler Jr
> jchandjr at bellsouth.net
> http://www.mp3.com/mikrodotz
>
>
>
>
>
> dupswapdrop -- the music-dsp mailing list and website: subscription info,
> FAQ, source code archive, list archive, book reviews, dsp links
> http://shoko.calarts.edu/musicdsp/
>


dupswapdrop -- the music-dsp mailing list and website: subscription info,
FAQ, source code archive, list archive, book reviews, dsp links
http://shoko.calarts.edu/musicdsp/