Ach! NO! inversion does not happen at the input!
I guess you can look at the topologies and discover for yourself.

I'm having a bad week.
x
Ed


_-_-_-_-_-_-_-^-_-_-_-_-_-_-_

For Lone Shark releases, Pure Data software and published Research, go to http://sharktracks.co.uk 


On Tuesday, 24 April 2018, 14:33:13 GMT+1, Ed Kelly <morph_2016@yahoo.co.uk> wrote:


Correct me if I'm wrong (and I'm sure someone will) but going back to analog electronics...

If you examine the topology of a standard analog filter design, such as the Thomas Henry VCF1 (http://www.birthofasynth.com/Thomas_Henry/pdf/VCF-1/Sheet_0002.pdf) you will see that the signal path from input to output goes through 5 inversion stages before being fed back into the input to generate resonance. Therefore it is inverted.

Since most (resonant) filters consist of an even number of stages (and therefore an even number of poles and/or zeroes) you should probably invert the feedback. With an odd number you should probably not invert feedback. My guess is that vcf~ implements an even-number (probably 4-pole) resonant filter, a la Moog. You can also try mvcf~ from my library (ekext - http://sharktracks.co.uk/puredata/ekext-0.1.8.tar.gz) which is another attempt at emulating this topology, or there are lots of others I'm sure you know about.

This might seem counter-intuitive, but the image below shows how this works with regards to vcf~ - the output of vcf~ is NOT inverted, whereas the output of an analog VCF should be.

Odd-number staged filters are very rare in the analog world, and most (i.e. 18dB per octave) were created to implement the (mythical 3-pole topology) filter of the Roland TB303. However, according to the schematics of the TB303 it is a four-pole diode-ladder VCF using transistors as diodes, with extra resonance artifacts revealed by Tim Stinchcombe on his excellent website. Tim Stinchcombe - TB-303 Diode Ladder Filter model




Meanwhile, if you want to go deeper with regards to 4-pole filter topologies (and lowpass/highpass configurations) it is worth checking out a paper published in 1985 - http://electronotes.netfirms.com/EN85VCF.PDF

The easiest way to think about it, from this paper, is that you think about each stage as an inversion, and the input itself is one of the stages, so for a 4-pole filter there are 4+1 inversion stages. But in the digital world I guess you'll have to test input and output at a relatively low frequency to the cutoff to determine inversion principles for the enhancement of resonance.

I hope this helps. It gave me a break from mixing!
Ed
_-_-_-_-_-_-_-^-_-_-_-_-_-_-_

For Lone Shark releases, Pure Data software and published Research, go to http://sharktracks.co.uk 


On Tuesday, 24 April 2018, 06:44:19 GMT+1, William Huston <williamahuston@gmail.com> wrote:


What's the best or "correct" way to add resonance to a [vcf~]?
  • Should the feedback be delayed?
  • Should the feedback be inverted?
  • Should the feedback be from the outlet I'm using?
    • Or should the feedback always come from the BP or HiPass side?
Thanks!
BH

--
William Huston:  WilliamAHuston@gmail.com
Binghamton NY

Public Service Mapping / Videography / Research
Blog -- Facebook -- Twitter  -- Youtube
Document collections: VirtualPipelines -- BHDCSDimockArchive
Please support my work: TinyURL.com/DonateToBillHuston


_______________________________________________
Pd-list@lists.iem.at mailing list
UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list