Jun 30, 2013 at 11:35pm


Hi friends!

Do you any of you have experience working with vactrol filters….or, does anyone have any ideas on recreating some of the behaviors of the LEDs and LDRs therein?



Jul 1, 2013 at 12:07am

I do, I love a good vactrol jobbie. I have a few filters with them, too.

The trick is emulating that ping. You can do that with the slide~ object, among others. Set the slide_down parameter to match the discharge of the vactrol – it should still be very fast on the slide_up.

Jul 1, 2013 at 7:26pm

Perhaps one you’d like to add to the toolbox?

Jul 1, 2013 at 9:32pm

Great question! I’m curious about this too!

I hear that Aalto can do some vactrol-esque sounds. I’d be curious to see how they do it.

I bet you could get similar plucky, long-tail sounds with physical modeling but I don’t know enough to really be of help.

You can fairly easily model the CV response of a vactrol as mentioned above but given the complexity of analog filters, I think there is room to explore a “vactrol-like” sound from a holistic perspective too, indirectly modeling that sound.

Jul 1, 2013 at 11:23pm

Thanks for getting back, everyone. Wetterberg, you are talking about… lowpass filtering (via slide~), e.g., an adsr~ and using that to modulate…. a filter cut off? Could you please tell me or show me what you mean when you say “that ping”?

Jul 2, 2013 at 1:41am

Peter et al, I should have been more specific; I meant physical analog filters that use vactrols. Low Pass Gates, as they’re called. I have some diy units and a MakeNoise QMMG, if you want to look them up.

And yes, it could be modelled holistically, but honestly, apart from doing resonance-squeals, the main attraction and difference between normal and optocoupler-controlled filters is the ping-sensation, which is just rubbery goodness all day long.

What I mean by the ping is that the cv control of a lowpass gate is done through the vactrol sections, and those have a slew to them; so if you feed it a quick pulse you get a fast-attack, slow release thing coming out, and that is what controls the filter. It is reminiscent of the things you hear in phys.modelling, but on traditional synthesis sounds it’s really a thing to behold.
On top of that a true lowpass gate has both vca and vcf properties, so there’s some tricky stuff in modelling that, too.

I personally believe that it could be modelled sufficiently with a control-addon to an already existing max/gen~ algorithm, but I’d have to do some testing – it’s certainly an interesting area, that I’d love to help out with! A joint forum project, perhaps?

Jul 2, 2013 at 4:42pm

And don’t forget a randomly chosen leakage amount…

Jul 2, 2013 at 5:13pm

absolutely… it adds punch. Or “vintageness”, I forget which. ;)

Jul 12, 2013 at 12:30am

So, I thought I’d post some sort of starting point for this forum project…. mhmm, Wetterberg:<code>

– Pasted Max Patch, click to expand. –

p.s. Thanks for the filter model, Peter, it makes max a lot better sometimes.

Jul 21, 2013 at 6:57pm

You can actually still buy vtlc3s (the optoisolator in question) and the datasheet linked from this product page illustrates exactly the characteristic slew you’re after:

Would LOVE to see a good gen-based vactrol model. There’s really not a heck of a lot going on under the hood of the actual Buchla LPG iirc (see for example, so this would be a good, broadly useful, and (unless I am overlooking something massive) relatively simple starting point.

Jan 4, 2014 at 7:31am

In my effort assemble a modular in software and therefore not have to buy one (it’s not going well), I found both this thread and a paper describing a low pass gate emulated in Max:

A Digital Model of the Buchla Low Pass Gate
Julian Parker and Stefano D’Angelo

along with audio samples, an MXF file, and source:

Jan 4, 2014 at 8:30am

That is really good! The Buchla LPG isn’t resonant, so I’m wondering why all that extra jazz is in there, it looks more like the makenoise implementation.

Jan 4, 2014 at 5:29pm

I think slide~ is a great place to start if you want that vactrol sound. I’ve a lot of different hardware vactrols here and it’s worth remembering that even two vacs from the same batch all have different rise and fall times so you there is no real definitive model. Don’t get to lost in trying to model something to the nth degree and make some sounds that you like using the vactrol concept.

Here is a patch I made a couple of years back when I was hankering after a 303, this patch scratched the itch and it uses vacotrol style envelopes for Filter and Volume.



– Pasted Max Patch, click to expand. –


Jan 5, 2014 at 12:23am

Fantstic patch John, as the old Chinese restaurant joke goes, it’s rubbery…

Jan 30, 2014 at 4:33am

Coming to this a bit late, but that’s my paper linked above by Manresa. Glad somebody noticed it :)

Wetterberg – The circuit modelled in the paper is closest to Thomas White’s DIY version of the LPG, which is in turn almost identical to the original Buchla LPG circuits with the addition of an amplifier to control the resonance level in ‘lowpass’ mode. The original LPG has resonance too in ‘lowpass’ mode, but it’s just at a fixed value. The amplifier simplifies down to just a scaling in the model, so it didn’t make much sense to not include it.

The vactrol model in the paper is actually not so different to using e.g. slide~. The difference is that there is some extra feedback that gives it a more exponential shape. I’ve done a little bit of work on a lower-level physical model of vactrols in general, which would be able to capture all the behaviour including response to light leakage, long-term memory effects etc. Hopefully I’ll get a chance to complete that in the future.

Jan 30, 2014 at 8:57am

Hi Bacchus, I actually downloaded your code once I bumped into your article, quite by accident. I was looking into ways of getting more organic-sounding ( to me at least) envelopes since I’ve been thinking about getting away from traditional ADSR.

I would definitely welcome you sharing your implementations with the community. :-)

best wishes


You must be logged in to reply to this topic.