Mutable Instruments Plaits



Clouds is the last module which does not use separate ADC channels for pots and for CV inputs (For example in Braids, Tides, Grids, Clouds… the pots are just generating offsets mixed with the CV, and the sum is read by a single ADC channel).

Now I’m using a separate ADC channel for the CV inputs, and for the pots. Benefits:

  • You can use heavy low pass filtering on the pot readouts to get a very stable and smooth knob response.
  • You can do all kinds of tweaks in software to the pot’s control law.
  • You can assign “shift” functions to pots independently of the CV input.
  • Attenuverting is done in software, and you can tweak the attenuverter’s response to have a very flat plateau at the center.
  • Calibration is easier!


Looks great!


It certainly looks the goods. A few observations:

  • now we know why STM32F3 support libraries appeared in the MI repository on GitHub about two years ago: Plaits almost certainly uses an STM32F373 processor

  • the 16-bit precision that the delta low-noise ADCs built in to the STM32F373 is very welcome and will make this the module of choice for those into microtonal music and alternative tunings. It also matches the 16-bit pitch CV output precision of Yarns, and also Ornament & Crime, and the Tubbutec uTune (there may be other quantisers or pitch CV sources that use 16-bit DACs, but I don’t know of them since most manufacturers don’t bother to publish technical specs like Mutable does).

  • -3V to +6V pitch CV input range happens to match Ornament & Crime’s -3.5V to +6.5V output range perfectly!

  • 48kHz 16-bit outputs, not 96kHz at 16-bits as on Braids. But better bandwidth-limited synthesis algorithms probably mean that it sounds better than Braids, despite this.

    • Edit: Olivier has subsequently posted on MW regarding this:

    • Many algorithms in Braids did not use proper band-limited synthesis, and thus benefited (a bit) from running at a higher sample rate. Also, Braids used an instrumentation DAC with a separate reconstruction filter - in this kind of configuration, it helps to run at 2x or 4x the ideal sample rate (which would be 40kHz given that we don’t hear above 20kHz). Hence: 96kHz. Plaits uses band-limited synthesis pretty much everywhere, so there’s no need for oversampling. And it uses a really good audio DAC (PCM5100A) with a good interpolation filter.

  • DC-coupled outputs means Plaits can be used as an LFO, unlike Braids that wasn’t useful below about 1Hz due to it AC-coupled output.

  • I like the name! I suggested “Silver Plaits” to Patrick Dowling a few years ago for a hyper-Braids module he was working on (it never saw the light of day due to competing priorities, I think), as a bad pun on “s’il vous plaît” as well as being cognate to Braids.

  • the price is refreshing! Modules shouldn’t cost the thick part of a grand, as so many these days do.


I’m intrigued by the LPC stuff. I’ve been poking around with LPC synthesis myself, @pichenettes . Can I send you a pvt message with a question about formant control, please?


Yes, just great!
Thank you Olivier.


Why not create a separate thread about this?


Ok, but I’m not sure anyone else will have the answer :slight_smile:


I was admittedly sketical about how on Earth one could upend Braids, which has been invaluable to me as I started and continue my modular journey. This looks very exciting and, importantly, complementary to my growing system. It makes me want to pick up Ripples and Warps, but likely Plaits first, especially if a gamble was taken on production.

I am so excited.
EDIT: Just saw the price. I love you MI!


But why should I keep my reply private?


I’m really intrigued about the normalised release envelope. Such a great idea.

EDIT: Was exited but a bit tipsy when i wrote the first version.


So what are the built-in phonemes and/or words in speech/vowel mode? And can the module say its own name (which would be a meta mode)? And if so, does it say it with American pronunciation (“plates”, omg!) or the proper pronunciation (“plats” as in “cats” or “flats”)?


Ok, fair point.


While Pichenettes alone might be able to give you an answer, it might be very valuable for the community to be able to read the answer. Also, it will stay there for future reference, in case the question should come up again in the future.


so, the internal decay curve is hardwired to the three attenuverter on timbre, fm and morph?
this would open up a huge amount of animation with just one trigger, crazy! (if so: love it!)


Good points. I made a new thread (see above).


This is correct and yes, it’s particularly amazing when using the percussion models


And that’s the easter egg :slight_smile:


Do the mini knobs have a dead zone for the middle position to turn them off completely?


Ah you just answered my question about a dead zone to avoid unwanted modulation from the mini knobs. That’s a really interesting approach. What kind of curve do you use?


Cubic works well and has been my go-to curve for a while, but this time it’s something different: