[sdiy] Polymoog resonator question

Mattias Rickardsson mr at analogue.org
Fri Oct 12 11:36:44 CEST 2018

On Thu, 11 Oct 2018 at 16:45, <rburnett at richieburnett.co.uk> wrote:

> On 2018-10-11 10:52, Mattias Rickardsson wrote:
> > Generally speaking in a multi-band parametric eq or resonator like
> > this I'd definitely want the ranges to overlap more or less - having
> > 60-300 & 300-1500 & 1500-7500 without overlap sounds like a
> > questionable idea from Moog's side - but on the other hand there can
> > be some problems with huge amplitudes when peaks collide.
> I don't think you will get really huge amplitudes when you tune two
> resonators to the same amplitude.  This is a parallel filter bank.  It's
> not like a parametric EQ where the audio is fed through each resonator
> one by one in series.

Ah, good point! My bad, I wasn't aware that it was a parallel processing

> Parametric EQ:  Signal gets boosted by say 10dB in narrow range around
> f1.  Then, signal gets 10dB boosted in narrow range around f2.  Then
> signal gets 10dB boosted in narrow range around f3.  When f1=f2=f3 the
> gain in this frequency area is 30dB.
> Moog parallel resonators:  Signal gets divided into three equal parts.
> Each part gets signal boosted by 10dB in a narrow range around f1, f2,
> or f3.  Then the three portions of the original signal get combined
> again.  If you split a signal into 3 parts, do the same thing to all
> three parts and then combine them back together, isn't it the same thing
> as applying the processing one to the whole signal?

There will be a gain increase in the second case but I don't think it
> will be anywhere near as profound as what you would get with a true
> parametric EQ when all bands are tuned to the same frequency and full
> boost applied.

My first quick guess is that 10 dB boost means that you added 216 % to the
original signal (since 10 dB boost means a 3.16 factor), and if you do that
with three parallel bands you add 3 * 216 % = 648 % to the original signal,
totalling 748 %. The factor 7.48 means 17 dB boost. This should be compared
to your 30 dB boost for the parametric eq case.

Perhaps of more interest are the deep notches that you get *between* the
> peak frequencies of the resonators where the skirts of the responses
> meet each other with opposite phase, and annihilate each other!  Now
> that's what the phase inversion switch on the middle band is all about
> influencing...

That's another detail to dive into, for sure. :-)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20181012/99e5edf5/attachment.html>

More information about the Synth-diy mailing list