| Author |
Message |
varice

Joined: Dec 29, 2004 Posts: 626 Location: Northeastern shore of Toledo Bend
Audio files: 16
G2 patch files: 46
|
Posted: Sun Jan 25, 2009 11:34 pm Post subject:
About the Classic Filter dB/Oct Settings |
 |
|
It is obvious that Clavia intended to emulate the famous classic Moog 24dB/octave transistor ladder low pass filter with the Classic filter module set to 24dB/Oct. But, I wonder about the Classic 12dB/Oct and 18dB/Oct settings.
Is the Classic 18dB/Oct setting intended to emulate the (in)famous classic Roland TB-303 18dB/octave diode ladder low pass filter?
Is the Classic 12dB/Oct setting intended to emulate the (in)famous classic Oberheim SEM 12dB/octave low pass filter? [EDIT: how silly of me, of course I meant to refer to the 12dB/octave filter sounds of the latter Oberheim synths, not the earlier SEM filter sound]
Or, are these just additional slope option settings for the Moog ladder emulation? _________________ varice Last edited by varice on Wed Jan 28, 2009 8:10 pm; edited 1 time in total |
|
|
Back to top
|
|
 |
drapdap

Joined: Oct 11, 2004 Posts: 186 Location: bp
Audio files: 1
|
Posted: Mon Jan 26, 2009 5:33 am Post subject:
Re: About the Classic Filter dB/Oct Settings |
 |
|
| varice wrote: | | It is obvious that Clavia intended to emulate the famous classic Moog 24dB/octave transistor ladder low pass filter... |
is it? For me it doesn't really much sound like that.
Mind you, i have rare encounters with that filter, plus i've never heard one backed with digital oscillators,
(apart from the unorthodox formant clone i'm using with the engine, but that's another conversation...)
Or didn't knew its that, but still...
But Where's that Omph close to the bottom?
Arturia MM had that sound much much closer than Clavia did.
So my bet is it's named Classic as a 24db ladder filter is a Classic in a modular, a 24db Highpass it isn't. Well, not for the Moog anoraks.
So it's a digital implementation of an analog filter.
And then there are the poles, ok, but isn't a diode ladder supposed to sounds different than a transistor ladder?
i -for one- think it's not just the number of poles used that make a filter.
Strange question you've got there...
My question is how do you implement a Steiner filter in the G2, with hp, bp, and lp inputs?
Where is Rob Hordijk to take us out of the fog?
Varice, did you read the paper he wrote on filters?
That is an excellent read...
róbert |
|
|
Back to top
|
|
 |
Tim Kleinert

Joined: Mar 12, 2004 Posts: 990 Location: Zürich, Switzerland
Audio files: 6
G2 patch files: 201
|
Posted: Tue Jan 27, 2009 10:44 am Post subject:
|
 |
|
I once skimmed a paper that described modelling the Moog transistor ladder filter as a string of four individual 6dB poles with nonlinearities inbetween. This approach makes it easy to optionally tap the 12dB and 18dB as well. Hence I assume that Clavia took this approach. Rob mentioned that the Classic filter module indeed adds the first odd harmonic to the signal if driven hard. He exploits this nicely in a DIY filter design that uses the Classic filter in a feedback structure, thus adding the whole odd harmonic series in decreasing magnitude to the signal.
No, the Classic Filter in 18dB mode doesn't sound like the trademark Roland TB filter at all. That filter has a very quirky design to keep the resonance boost in check, which gives it its characteristic sound.
And the SEM wasn't a pure lowpass filter. It could sweep continuously from LP via BP to HP. IIRC, it was the OBX, the OBXa and the OB8 that had 12dB lowpass only, but the first one was discrete circuitry while the latter two were based on CEM chips -which makes it even harder to find similarities. Matrix12/Xpander: 15 filter types -totally different ballgame. _________________
 |
|
|
Back to top
|
|
 |
iPassenger

Joined: Jan 27, 2007 Posts: 1051 Location: Sheffield, UK
Audio files: 5
G2 patch files: 78
|
Posted: Wed Jan 28, 2009 2:17 am Post subject:
|
 |
|
| tim wrote: | | I once skimmed a paper that described modelling the Moog transistor ladder filter as a string of four individual 6dB poles with nonlinearities inbetween. This approach makes it easy to optionally tap the 12dB and 18dB as well. Hence I assume that Clavia took this approach. Rob mentioned that the Classic filter module indeed adds the first odd harmonic to the signal if driven hard. He exploits this nicely in a DIY filter design that uses the Classic filter in a feedback structure, thus adding the whole odd harmonic series in decreasing magnitude to the signal. |
You have a link to this filter? or was it one of the ones in his g2 guide?
Drifting OT: seem to remember you can get some pretty unstable and interesting results by creating your own feedback path for the Nord filter. _________________ iP (Ross)
- http://ipassenger.bandcamp.com
- http://soundcloud.com/ipassenger |
|
|
Back to top
|
|
 |
Tim Kleinert

Joined: Mar 12, 2004 Posts: 990 Location: Zürich, Switzerland
Audio files: 6
G2 patch files: 201
|
Posted: Wed Jan 28, 2009 3:28 am Post subject:
|
 |
|
| iPassenger wrote: | | You have a link to this filter? |
It's a paper by Annti Huovilainen:
http://dafx04.na.infn.it/WebProc/Proc/P_061.pdf
Huovilainen is the author of the Smartelectronix ASynth softsynth plugin, which is based on this design. Sound clips here:
http://www.acoustics.hut.fi/publications/papers/dafx2004-moog/
...I must say, this sounds rather good. And AFAIK, it's freeware.
The problem with all types of digital nonlinear feedback systems (like filters) is that they require copious oversampling on order to offer uncompromised quality within the surrounding system frequency. Since the G2 doesn't offer oversampling for DIY filter circuits, those have to be bandlimited to approx. 12kHz in order not to become unstable. Rob does this with his intergrating mixers. This process however makes the resonance increasingly duller as cutoff frequency sweeps close to and past 12kHz, where it peters out completely -which is the compromise I'm referring to.
| iPassenger wrote: | | or was it one of the ones in his g2 guide? |
Rob talks about the instability issues and how to combat them with integrating mixers. Not so much about the specific digital Moog design.
| iPassenger wrote: | | Drifting OT: seem to remember you can get some pretty unstable and interesting results by creating your own feedback path for the Nord filter. |
Oh absolutely. Feedbacking the quadrature inverted filter output to the resonance modulation input and opening that just a bit makes the filter produce characteristic odd harmonic soft clipping. And the whole thing remains stable too, no need for integration. Very cheap. It's one of my favs.
Oops, I drifted off topic too. Well it sort of is related to the Classic Filter module, which is inspired by the Moog filter, so I hope nobody minds. _________________
 |
|
|
Back to top
|
|
 |
varice

Joined: Dec 29, 2004 Posts: 626 Location: Northeastern shore of Toledo Bend
Audio files: 16
G2 patch files: 46
|
Posted: Wed Jan 28, 2009 11:24 pm Post subject:
|
 |
|
Hey guys!
róbert, Ross, and Tim - I thank you guys very much for your replies to my very (overly?) simplistic post. I really do appreciate your discussion and opinions!
Tim - thanks for correcting me about which model of the Oberheim synths had that classic 12dB/Oct filter sound... (not the old SEM, of course!)
What I did not mention in my topic question - is that my Moog MF-101 LPF also has a 2-pole (12dB/octave) setting, so I can easily do an A/B comparison of the G2 Classic filter 12dB/Oct setting vs. my moogerfooger filter!
Reference to my G2X/moogerfooger LPF Frankenstein Monster:
http://electro-music.com/forum/topic-12266.html
So, I already have my own opinion about the G2 Classic filter dB/Oct settings that I will post here soon...
But anyway, would anyone else care to comment about the G2 Classic Filter? _________________ varice |
|
|
Back to top
|
|
 |
varice

Joined: Dec 29, 2004 Posts: 626 Location: Northeastern shore of Toledo Bend
Audio files: 16
G2 patch files: 46
|
Posted: Fri Jan 30, 2009 1:17 am Post subject:
|
 |
|
| drapdap wrote: | | Mind you, i have rare encounters with that filter, plus i've never heard one backed with digital oscillators... Or didn't knew its that, but still... |
You can hear a real Moog LPF driven by the digital G2 oscillators in the sound clips that I posted in my topic about using the moogerfooger filter with my G2X.
| drapdap wrote: | | Arturia MM had that sound much much closer than Clavia did. |
Yes, Arturia did a good job of emulating the Moog filter. I don't like to use softsynths though.
| drapdap wrote: | | Where is Rob Hordijk to take us out of the fog? Varice, did you read the paper he wrote on filters? That is an excellent read... |
I have only read what he has posted on this site and his G2 Tutorial. Did he write something else? Do you have a link to it?
| tim wrote: | I once skimmed a paper that described modelling the Moog transistor ladder filter as a string of four individual 6dB poles with nonlinearities inbetween. This approach makes it easy to optionally tap the 12dB and 18dB as well. Hence I assume that Clavia took this approach...
No, the Classic Filter in 18dB mode doesn't sound like the trademark Roland TB filter at all... |
After doing an A/B comparison of the sound of the 12dB/Oct settings of the G2 Classic filter and my moogerfooger filter, I have concluded that they sound very similar. So, I believe that the Classic 12dB/Oct and 18dB/Oct settings are only just optional slope settings of the 24dB/Oct Moog ladder emulation.
While working on this, I read a paper "Analysis of the Moog Transistor Ladder and Derivative Filters" by Tim Stinchcombe and also checked out his website. One interesting thing that I learned is that the Roland TB-303 diode ladder filter is actually a 4 pole design, even though Roland claimed that it is an 18dB/octave filter. _________________ varice |
|
|
Back to top
|
|
 |
iPassenger

Joined: Jan 27, 2007 Posts: 1051 Location: Sheffield, UK
Audio files: 5
G2 patch files: 78
|
Posted: Fri Jan 30, 2009 2:37 am Post subject:
|
 |
|
| varice wrote: | | One interesting thing that I learned is that the Roland TB-303 diode ladder filter is actually a 4 pole design, even though Roland claimed that it is an 18dB/octave filter. |
Isn't it just that the Res loop uses all 4 filters but the filter itself only uses 3?
Can anyone explain in relatively simple terms what the diode ladder arrangement imparts on the signal as opposed to the transister ladder?
- My sub-GCSE standard electronic knowledge thinks that the diode should sound rougher (or more distorted) due to their blocking/one-way nature but I am sure that is juvenile way of looking at it.
From a manufacturing perspective I presume it would be cheaper and easier to use diode design? (again see above for severely limited to know-how) _________________ iP (Ross)
- http://ipassenger.bandcamp.com
- http://soundcloud.com/ipassenger |
|
|
Back to top
|
|
 |
Blue Hell
Site Admin

Joined: Apr 03, 2004 Posts: 19596 Location: The Netherlands, Enschede
Audio files: 116
G2 patch files: 317
|
Posted: Fri Jan 30, 2009 9:35 am Post subject:
|
 |
|
| varice wrote: | I have only read what he has posted on this site and his G2 Tutorial. Did he write something else? Do you have a link to it?
|
Some links : http://electro-music.com/forum/topic-12871.html - maybe its in there. _________________ Jan |
|
|
Back to top
|
|
 |
varice

Joined: Dec 29, 2004 Posts: 626 Location: Northeastern shore of Toledo Bend
Audio files: 16
G2 patch files: 46
|
Posted: Fri Jan 30, 2009 10:54 am Post subject:
|
 |
|
| iPassenger wrote: | | varice wrote: | | One interesting thing that I learned is that the Roland TB-303 diode ladder filter is actually a 4 pole design, even though Roland claimed that it is an 18dB/octave filter. |
Isn't it just that the Res loop uses all 4 filters but the filter itself only uses 3? |
According to Stinchcombe's analysis, the core design has 4 poles that affect the frequency response even when the feedback (resonance) is set to zero.
| iPassenger wrote: | | Can anyone explain in relatively simple terms what the diode ladder arrangement imparts on the signal as opposed to the transister ladder? |
According to Stinchcombe, the transistor ladder has more isolation between each pole than the diode ladder. One result of this is that the diode ladder begins to cut into the passband sooner and has a much more gentle corner than the transistor ladder. See Figure 31 and the associated text in his paper.
http://www.timstinchcombe.co.uk/synth/Moog_ladder_tf.pdf
| iPassenger wrote: | | From a manufacturing perspective I presume it would be cheaper and easier to use diode design? |
Actually, the TB-303 has transistors in the filter poles, but only the base-emitter junction is used as a simple diode! It appears that the main reason for using diodes was to try to prevent infringement of the Moog transistor ladder patent. _________________ varice Last edited by varice on Fri Jan 30, 2009 11:24 am; edited 1 time in total |
|
|
Back to top
|
|
 |
varice

Joined: Dec 29, 2004 Posts: 626 Location: Northeastern shore of Toledo Bend
Audio files: 16
G2 patch files: 46
|
Posted: Fri Jan 30, 2009 10:57 am Post subject:
|
 |
|
| Blue Hell wrote: | | varice wrote: | I have only read what he has posted on this site and his G2 Tutorial. Did he write something else? Do you have a link to it?
|
Some links : http://electro-music.com/forum/topic-12871.html - maybe its in there. |
Thanks Jan. I will take a look. _________________ varice |
|
|
Back to top
|
|
 |
iPassenger

Joined: Jan 27, 2007 Posts: 1051 Location: Sheffield, UK
Audio files: 5
G2 patch files: 78
|
Posted: Wed Feb 04, 2009 7:45 am Post subject:
|
 |
|
| tim wrote: | Feedbacking the quadrature inverted filter output to the resonance modulation input and opening that just a bit makes the filter produce characteristic odd harmonic soft clipping. And the whole thing remains stable too, no need for integration. Very cheap. It's one of my favs.
|
Cheers for the great tip! Been playing around with filter out self-ringmod (same as quadrature right?) controlling the filter res level.. Sounds great on the Nord LP filter, don't think I inverted it though. _________________ iP (Ross)
- http://ipassenger.bandcamp.com
- http://soundcloud.com/ipassenger |
|
|
Back to top
|
|
 |
iPassenger

Joined: Jan 27, 2007 Posts: 1051 Location: Sheffield, UK
Audio files: 5
G2 patch files: 78
|
|
|
Back to top
|
|
 |
egw

Joined: Feb 01, 2003 Posts: 1501 Location: Asheville NC
G2 patch files: 8
|
Posted: Thu Feb 05, 2009 9:59 am Post subject:
|
 |
|
| tim wrote: |
Oh absolutely. Feedbacking the quadrature inverted filter output to the resonance modulation input and opening that just a bit makes the filter produce characteristic odd harmonic soft clipping. And the whole thing remains stable too, no need for integration. Very cheap. It's one of my favs.
|
Can you explain what you mean by "quadrature inverted filter output?" _________________ www.gregwaltzer.com |
|
|
Back to top
|
|
 |
varice

Joined: Dec 29, 2004 Posts: 626 Location: Northeastern shore of Toledo Bend
Audio files: 16
G2 patch files: 46
|
Posted: Fri Feb 06, 2009 1:05 am Post subject:
|
 |
|
| egw wrote: | | tim wrote: |
Oh absolutely. Feedbacking the quadrature inverted filter output to the resonance modulation input and opening that just a bit makes the filter produce characteristic odd harmonic soft clipping. And the whole thing remains stable too, no need for integration. Very cheap. It's one of my favs.
|
Can you explain what you mean by "quadrature inverted filter output?" |
Perhaps Tim is referring to using level (four quadrant) multiplier modules in the filter resonance feedback as described by Rob Hordijk in the Resonant filter section of his G2 Tutorial. _________________ varice |
|
|
Back to top
|
|
 |
|