KVR Audio

deastman · Post by **deastman** » Sun Sep 25, 2016 7:52 pm

Self oscillating filters produce sine waves, not square waves. If you're getting a self oscillating square wave, it is because the gain is being overdriven somewhere after the filter in the signal chain, and the sine is being clipped to a square.

Architeuthis · Post by **Architeuthis** » Mon Sep 26, 2016 2:47 am

Yes, but I'm even having trouble finding a digital filter where it will self oscillate and become square without cheating (i.e. adding arbitrary distortion). The problem with cheating is that you don't get the "disappearing oscillation" effect. And if you can't get that effect, there's even more that you're not getting and therefore losing out on important iconic analog style sound... well, it's iconic to me, anyway.

Ok, updated list so far that exhibit this behavior:

Thomas White Buchla LoPass Gate (nothing to do with MS-20)
Ian Fritz Threeler (MS-20 inspired)
MOTM-420 (there's good reason to believe this is a very accurate replica of the MS-20)

Also, listening to demos on youtube of the MS-20, I hear hints of this phenomenon but I cannot verify it with 100% certainty. But it's pretty safe to say that the MS-20 behaved like this as well (once you plugged its output back into its input). Mystran, I'm talking to you! Here, just take a listen: https://youtu.be/yn5FVZGPWLM?t=1m14s

valhallasound · Post by **valhallasound** » Mon Sep 26, 2016 4:06 am

MS20 Mini is $449 nowadays. It has 2 MS-20 filters in it, AND the External Signal Processor, which is horrible at tracking the input signal, but is awesome at turning anything into a wall of distortion. I love digital stuff (I make my living from designing digital stuff), and I feel that the MS20 Mini is a great way to augment a computer full of tasty DSP.

Sean Costello

Architeuthis · Post by **Architeuthis** » Mon Sep 26, 2016 5:42 am

For privacy reasons I will not reveal the source of information, but CASE CLOSED. What I am after is possible in the digital realm. That's all I needed to know.

Yes Sean, it is nice to have some analog in your workflow for that extra something something. Analog synths just work. No dealing with anti-aliasing, samplerate, all the those digital limitations. There's just a few things I'd like to steal from analog synths for the digital world so that I can just have those things ready to go and compose with on the fly. My analog modular is impractical for efficient composition so I'm now trying to replace it by studying it and understanding what I like about analog and recreating that digitally.

Architeuthis · Post by **Architeuthis** » Mon Sep 26, 2016 6:21 am

I think people get too caught up in the details. Example: the digital Moog recreations that attempt to model the circuits to get those precise arbitrary details. For me, that is all exhausting and overly-detailed. I understand people hear great synth patches from an analog Moog synthesizer and then want that sound, and want it EXACTLY. I used to think this was very important, getting the exact timbres, the exact bassy-ness, the exact high frequencies, the exact waveshapes, all these things that are incredibly arbitrary. But I've come to learn, as I study my analog synth, that the sounds I like, the sounds I am after have nothing to do with the detail, it has all to do with behavior. Yes, some detail is nice to get that extra phatness or crispness, but that's less important if you can't even get the basic sound you want. Random fluctuation of voltage due to resistors and heat and such, it's not the reason analog sounds better.

Now, if you want the holy grail of screaming distortion, don't look for an overdriven filter. What you want are oscillators with the correct features. I learned that true scream is created with feedback-sync and FM (IIRC, it's two oscillators syncing with each other and applying FM somewhere). Filters weren't made to scream and so I claim that filters are not really the perfect tool for the job. Once in a while I can only be satisfied by the true master of scream, the simple oscillator (well, two oscillators interacting).

demo: http://www.elanhickler.com/_/analog_bio ... eation.mp3

It's a more aggressive scream than a filter can ever provide while still maintaining musicality. I hope to have this feature in the product I'm working on, Chaosfly.

Anyway, these are just my opinions and ramblings. I just know there are people out there like me searching for that sound they heard in some music, desperate to have it in their own music, not realizing that they could be totally looking in the wrong place, using the wrong tools, and being only half-satisfied.

urosh · Post by **urosh** » Mon Sep 26, 2016 12:33 pm

Architeuthis wrote: Thomas White Buchla LoPass Gate (nothing to do with MS-20)

Actually, LPGs (and derivates) are based on Sallen-Key topology, same as MS10/MS20mk1, but using LDRs as control elements. However, for two phenomenons you've described, critical element beside topology is resonance limiting, which is pretty much the same in all 3 filters you've mentioned. This behaviour is definitely not common for all analog filters.

Architeuthis · Post by **Architeuthis** » Mon Sep 26, 2016 1:00 pm

urosh wrote:This behaviour is definitely not common for all analog filters.

If you can push the resonance to the point of becoming square I would think said behavior is inevitable. No?

Urs · Post by **Urs** » Mon Sep 26, 2016 1:41 pm

Architeuthis wrote:
urosh wrote:This behaviour is definitely not common for all analog filters.
If you can push the resonance to the point of becoming square I would think said behavior is inevitable. No?

There is no filter where the self oscillation can be pushed to a square. There are only filters where the resonance can be so loud, it clips an electronic part that sits behind the filter.

urosh · Post by **urosh** » Mon Sep 26, 2016 3:18 pm

Architeuthis wrote:If you can push the resonance to the point of becoming square I would think said behavior is inevitable. No?

You can't "push the resonance to the point of becoming square" in most topologies, not even close (and in none you can get distinct clipped waveform, even Korg-ish). For example, no matter how much feedback you apply in transistor ladder VCF, output will be sinewave-ish. Same with typical SVF or quad OTA filters. I've encountered behaviour you've mention (sort of squarewave-ish resonance and reso around zero crossing when overloaded) only in specific circumstances: (quasi)Sallen-Key with diode limiter in feedback path.
Harmonic locking is another issue, lot of filters tend to do that, including some digital . Btw, I can get all these effects with Diva, maybe just not as dramatic as Threeler (but that's one very special filter). I use a lot self-oscillating MS20 filter in Diva for bass drum exactly for that filtered squarewave tone.

xoxos · Post by **xoxos** » Thu Oct 13, 2016 2:30 pm

enough money *will* make it square! amitabha. it's like ross perot does dsp.

staiff · Post by **staiff** » Wed Oct 19, 2016 4:02 pm

hello !
MS20 filter is a (pretty close) MOOG ladder filter with 24 db resonnant filter.
i know "minisynth" from racafx use this ladder filter, modulation matrix for combining the 2 oscillators & 1 combined ADSR enveloper (VCA+filter) like the JUNO 106.
the rackAFX code is more or less diffrent than WDL so you need to "ajust" it

Has anyone created a good ms-20 filter emulation?