Has anyone created a good ms-20 filter emulation?

DSP, Plugin and Host development discussion.
RELATED
PRODUCTS

Post

Has anyone created a good ms-20 filter emulation? I just need to answer one question: Has anyone created a digital filter that exhibits the "disappearing oscillation" effect? This: https://youtu.be/15oSTHittCM?t=3m48s

...if there's an official term for this let me know. :tu:

This is not unique to ms-20 filters, but is probably a behavior in all analog filters that can be overdriven to the point of having square resonance.

Post

Korg MS20 vsti can be used as a filter, you should try it

Cheers,
David

Post

I'm not seeing a demo download link for that.

Post

Architeuthis wrote: This is not unique to ms-20 filters, but is probably a behavior in all analog filters that can be overdriven to the point of having square resonance.
It's not really universal to "all analog filters" at all and on the other hand it's not particularly hard to recreate in software (and there are plenty of plugins that do it, for MS-20 emulation or otherwise).

It happens when you have a resonance feedback path with low-pass type response (at least until where it clips, such that the low-frequencies can push the cutoff-frequency components against the saturation to lower the effective feedback for part of the wave-form), this path needs to clip somewhat earlier than rest of the filter or you won't get the "resonance only middle" effect and then you also need to drive rest of the filter with high enough input amplitude that the resonance path can't keep up with the levels.. and finally you need enough resonance feedback that the oscillation can regenerate fast as soon as it comes back into the signal range where it's not squeezed by the saturation anymore.

Other than that there's nothing particularly magical about it.

Post

I really need to see an example in digital software, I've been looking for this behavior and can't find it even in Urs/u-he stuff. If it's not particularly difficult to create in software then I don't know why I've not seen it.

Mystran, you're not quite addressing the phenomenon; it's not simply that the resonance regenerates fast "as soon as it comes back into signal range". If it was as simple as that we would always see partial oscillations at the point around said threshold. Partial oscillations occur with lesser overdrive of course, but not when the resonance becomes completely square. I've seen this behavior in soft sync... such as U-he ACE variable sync. Well, soft sync is not a good enough analog of the behavior

Post

I don't know about this phenomenon but checkout The Drop by Cytomic. It emulates the MS 20 rev1 and rev2 in a lot of detail.

Post

Ok let's try it...

...Cytomic The Drop doesn't do it. The resonance never becomes square... unless I'm not using it correctly? I tried the LP DRIVE knob, tried changing the input level (although squaring off the resonance should only require turning up the LP RES knob, but that didn't work).

Post

Hmmm ... Have you checked the Korg Legacy Collection - http://www.korg.com/us/products/softwar ... page_1.php

Post

Juha_p, I could not find a trial version.

Post

Architeuthis wrote:Juha_p, I could not find a trial version.
OK, I didn't check that. So they don't have "try before buy" strategy anymore or do you need to create an account to get the demo software links?

Post

The filter in the video is labeled as "Ian Fritz Threeler (MS20 inspired)" so it's almost certainly not the actual MS20 circuit. Without a schematic it's impossible to say how close or far it might be.

I tried messing with the model parameters of an actual MS20 type filter (and it's probably not the best model out there, but I'd be willing to call it "commercial quality" anyway and it matches fairly closely what you already found in plugins), and it appears that you need to boost the resonance feedback level somewhat from the actual circuit maximum and then modify the feedback saturation to clip somewhat harder (or rather pass less signal when in saturation) compared to an actual MS20 r2 (and the output of the r1 version looks completely different).

In total, I spent about 5 minutes on replicating this behavior in software, it's not hard or anything if you already have the overall circuit modeled. Unfortunately, I don't have this type of filter as an analog circuit right now, so I'd have to build one to try, but I suspect you could get it from that as well by swapping a resistor or two from what the MS20 schematics specify. That said, I don't see how you would ever get this from an unmodified circuit.

So .. most likely what you find in plugins is the "correct behavior" for MS20 and what you see in the video is a slightly different filter circuit with different parameters and therefore different behavior.

YMMV.

Post

mystran wrote:The filter in the video is labeled as "Ian Fritz Threeler (MS20 inspired)" so it's almost certainly not the actual MS20 circuit. Without a schematic it's impossible to say how close or far it might be.
It's 3 single pole filters in series that can be configured as HP or LP each. The feedback goes back into the capacitor of the first stage. So It's much like a Sallen-Key with an extra stage in-between. Resonance self osc is afaik not at teh corner frequency and the whole things is designed to be heavily overdriven.

Post

Architeuthis wrote: probably a behavior in all analog filters that can be overdriven to the point of having square resonance.
Not, not a all, and in fact, it's quite the opposite : many synths (from late 70s, at least) had filters tuned so that they won't reach self-oscillation, or stay quite smooth when self-oscillating.
Juno/Jupiter, ARP, Teisco, Yamaha CS series among many others don't exhibit such a behaviour.

Also, the filter in this video sounds good, it's "ms20 style" (raw, dirty, overdriven, characterful), but still not really accurate imho, so it shouldn't be used as a comparison point.

Post

Please stop with the MS-20 comments. I'm not interested in what is accurate, correct, common, historical, intentional, etc. The fact is that two, possibly FIVE filters I have owned at one point exhibit this behavior, 3 of which are MS-20 based. But MS-20 or not, this is a common behavior in analog filters (at least as far as MY experience goes!). Another filter that does this is Thomas White Buchla LoPass Gate. Here's a demonstration:

https://youtu.be/iJOYLeFCGnE?t=1m22s

This is a very important behavior to recreating the sound of an analog filter. If you disagree, I just... I have no words for you. This sound is the very reason I fell in love with analog filters and I am only now connecting two and two together... i.e. why I like the sound VS analog filter behavior VS why I don't like digital recreations.

Edit: The reason for this thread is so that I can either figure out how to recreate this behavior myself in software or get someone else to do it (or find an already-made product). Furthermore, I am convinced that analog synthesizers are not special, and that digital synths can practically replace analog if we can recreate all desirable behaviors.

Post

...and I bet anything that I'm not alone, that this behavior is desired by a lot of potential customers and is a worthy endeavor.

Post Reply

Return to “DSP and Plugin Development”