[chuck-users] Filters blowing up: any news?

[Kassen]

Dan;

2010/3/2 Daniel Trueman <dtrueman at princeton.edu>

> i have no idea if this will help, but i've been meaning to bring the old gQ
> filter stuff to ChucK for sometime. gQ is a really nice filter:
>
>
That's a good idea in any case. More filters would be better. I'd still like
a a state-variable filter with three outputs. Such filters are spectacularly
useful in modular synthesis yet often under-appreciated in the digital
context.


> http://www.music.princeton.edu/~dan/gQpage/gQ.html
>
> it might be more stable. maybe not. it usually sounds real nice. in any
> case, here it is, and it seems to work ok. probably should make a class out
> of it...
>
>
Sounds great! I'd say, BTW, that unless this is implemented as a new object
with three outputs it could also be made a part of the old plan to extend
LPF, BPF and HPF to be able to switch to different topographies.

I have to say though that the plain "convenience" filters like BPF are -from
memory- 2nd order Butterworth filters. That's hardly a exotic type of
filter. At the very least they should be stable for static settings with a
reasonable Q. For what I understand of DSP there are two ways to get into
the current situation; you can make a typo or reasoning error in the
function that sets the coefficients or there can be instabilities induced by
rounding due to finite word-length. These can be checked for, though of
course at some computational price.

These, IMHO, need to be fixed, as something is plainly buggy there. How much
time would this take from somebody who is fluent in C/C++ (needed to
evaluate the rounding possibility) and not too scared of the Z-plane stuff?
I'd look into it myself but I'd be out of my league in both fields.

Yours,
Kas.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cs.princeton.edu/pipermail/chuck-users/attachments/20100302/0c826c1d/attachment.htm>