[sdiy] filters with independent resonance 'location'

[jhaible at t-online.de]

Jeffrey Pontius schrieb:
> In general, a typical lowpass or highpass filter has the 
> resonance set 
> around the cutoff frequency.  Are there filters 
> available, and if not 
> available could one be reasonable constructed, that would 
> allow the 
> resonance 'location' to be set independently of the 
> cutoff frequency? 
> 
> This is certainly not a new idea, but I have not seen or 
> read about such a 
> filter.

Normally such a filter will be "composed" from two individual
filters, in parallel or in series connection.

Connecting a BPF and a LPF in parallel is the most straightforward
method: Roland SH-5 and Synton Syrinx are two examples.
(On a modular, try a 4pole LPF and a 2pole BPF into a mixer.)

An example for a series connection would be a resonant phaser
and a LPF. A resonant phaser with two stages has one resonant
peak (like a BPF), but unlike a BPF, frequencies far from the center
frequency are not much attenuated: the asymptode is unity gain.
(more phaser stages result in more peaks.)

Another parallel connection example is the Moog Voyager dual LPF
configuration. I have recently built something like that (with the
second LPF switchable for direct comparison), and it's a nice
effect, too.

With all this, keep in mind that you never just add amplitudes
in parallel connection. Phase relation between the two partial
filters can create deep notches where you would just expect
a flat region between peak and cutoff.

(There are filter types like Chebycheff which have some ripple
in their passband by definition, but these filters are normally
built from individual 2pole or 3pole filters as well.)

JH.