[sdiy] Filter slopes

[Rutger Vlek]

Constructing band-pass filters isn't rocket science, but you do need to
work from the builings blocks typically available. I'm currently working
with Sound Semiconductor on a datasheet update for their SSI2164 part and
proposed to include all basic building blocks you need to construct a
low/band/high/all-pass filter. Basically, if you cascade two low-pass
filter blocks (6dB/oct slope each) you arrive at 12dB/oct. The same holds
for high-pass filters. Cascading both low and high pass blocks gives you
band pass, which you could use for your purpose of instrument emulation.

As was already mentioned, pole mixing is a great alternative way to obtain
various band-pass filter responses from what's basically a well layed-out
low-pass filter core. In addition to the Polaris and Olivier Gillet
implementation I'd like to mention something I designed over the last
years: http://riversynths.com/products/dendrites. The Dendrites eurorack
module goes a little further than the Polaris, in the sense that it
includes various resonance characters (basically different types of
resonance paths inspired by various famous filters).

If you need slopes lower than 6dB/oct, you could have a look at the
principle of spectral tilting, which I implemented (under CV control) in
this module http://riversynths.com/products/marble. Basically it's done by
a CV-controlled weighted sum of the output of parallel all-pass filters.

If you need slopes above 6dB/oct, but not a multiple thereof, you could
have a look at the Jurgen Haible "variable slope VCF". In this filter you
can switch each core (6dB/oct) between HP/LP mode, but you can also spread
the poles apart (different frequencies). This way you can construct slopes
that approach something in between 6dB and 12dB for instances. Musically,
I've never found this level of detail to be relevant, but in your case in
might be.

Finally, a very different approach to acoustical instrument simulation was
taken by the "woodwind wave shaper", which you can find on the site of
Jurgen Bergfors (bergfotron). I once prototyped this, but couldn't get it
running. It's a waveshaper, rather than a filter, but the demo on Jurgen's
site was rather convincing I think (in similar vein to the extremely
awesome demo of the stand-alone polymoog resonator demo circulating on
youtube used for emulating acoustic instruments as well).

Regards,

Rutger



2018-06-11 11:33 GMT+02:00 Tom Wiltshire <tom at electricdruid.net>:

>
> > On 11 Jun 2018, at 09:32, Neil Johnson <neil.johnson71 at gmail.com> wrote:
> >
> > For a practical implementation hunt out the Oberheim multimode filter
> circuit:
> > http://www.synfo.nl/servicemanuals/Oberheim/OBERHEIM_MATRIX-12_SERVICE_
> MANUAL.pdf
> >
> > See page 4.
>
> While the Oberheim is definitely the original source of this idea,
> Olivier’s or David’s 2164 implementation of this idea is more practical, in
> my view. The benefit of the inverting filter stages and the well buffered
> output from each integrator makes it as simple as just adding a bunch of
> resistors for each output you want. This can either be switched, as in
> Olivier’s original Shruthi "4-pole mission” filter, or you can implement a
> separate mixer for each required output and have them available
> simultaneously.
>
> Another simple possibility is to do the same thing using the
> AS3320/CEM3320. Again, this includes the buffers, making life much simpler,
> and in the case of this chip, you get VC-resonance for free too, unlike the
> 2164.
>
> These pole-mixing filters are ideal for creating asymmetrical-slope
> bandpass responses though, that’s for sure.
>
> Tom
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20180611/8ecde420/attachment.htm>