Interview with Olivier Gillet
Man, machine, computers and hardware, Yair Etziony from False Industries talks to Olivier Gillet the mastermind behind “Mutable Instruments” a company that releases some of the most cutting edge modules in “Eurorack” scene right now. This is the first interview in a series.
Did it ever happen to you that a track or a song, inspire you to make a specific module? And if so, can you specify?
While I was almost done with the development process of Elements, I
saw a performance of Pierre Bastien, a guy who builds
”electro-mechanical” instruments. What I found very interesting is
that only half of what you hear in his music is the musical signal, the notes themselves – the other half is noise from the mechanical device – friction, rubbing, clunky mechanical parts in action.
It reminded me that the sound of a guitar is not just the vibration of the string, but also the noise of the hands moving on the neck or tapping on the body, or even the humming of the performer…
sound engineers will tell you it’s something that shouldn’t be
captured, while some will tell you it’s what makes a recording
”breath”. I was fascinated by all these “scrtchchc” and “pfrrlrlr” which
almost sounded like percussions in themselves, filling the blanks
between the notes like the ghost notes of a drummer. So it pushed me
to rework the exciter section of Elements to add these kinds of noise
sources – which I obtained by removing everything pitched/harmonic
from recordings of me scratching, rubbing, hitting things.
Another example is the music of “Disasterpeace” (the soundtracks for Fez and It follows), which often have super wide chords played by unfiltered waveforms.
That’s a very interesting sound – it doesn’t sound “chiptune” at all because of the large polyphony, and actually reminds me more of string machines. I made a string synth easter egg in Rings with that kind of sound – not an actual vintage-sounding emulation, but instead something brighter that would fill every corner of the spectrum.
Reading about your background (which is more software oriented), would you please elaborate on your decision to make hardware?
It’s for sure harder and more risky than software, what did you had in mind?
Around summer 2009 I got interested in electronics and wanted to learn it seriously. It felt like a “hole” in my education, which was very focused on signal processing.
In France signal processing exists as a theoretical subject on its own, independent of electrical engineering, so I did not have to sit through classes about transistor circuits or microcontrollers or FPGAs or RF… and I came to regret it.
There was a bunch of other factors that pushed me towards it at that time…
I was getting tired of my day job in very large scale data processing and wanted to go back to “small” things, at the megahertz and kilobyte scale, get back to the feeling of understanding absolutely everything going on in a machine, a feeling we had 15 years earlier in the demo-scene. And that’s the time at which the “Arduino” boards were getting popular, so it looked like a good platform for learning.
It didn’t feel “hard” or “risky” at all because at the time it was just a hobby, it was absolutely not my goal to turn it into a business, I just wanted to try building a synth for fun.
It took 2 years between
these first experiments and the creation of Mutable Instruments as a
proper company, and 2 more years before I started selling my first
factory-made modules but none of this was really planned.
Yes, hardware is rather unforgiving, but I still prefer that to
software. Actually it’s not true, I love writing software… but for
my own hardware!
My issue is not with software itself, but with modern computing environments – mobile or desktop computers.
I’m not really in good terms with modern operating systems – I find their layers of abstraction cumbersome, I don’t like the rapid cycles of updates that can make things obsolete, and I really despise app-stores and control on software installation and distribution. My other issue is that people just don’t pay attention to software; it’s not a field in which you can make something and expect it to be in use in 5 years.
People are not going to invest time exploring or learning what you made, because they know that next year there’ll be an upgrade. When I see the mobile app market where it’s all about getting featured in the right place, and navigating iOS and Android’s minefields – it doesn’t look less risky or less hard than hardware.
I still remember that modular synthesis was something that is “Not sexy” very big boxes, that were associated with old school and outdated approach to electronic music, why do you think this changed so much lately?
One word: Eurorack. It is not backed by Buchla or Moog, so there is no heritage we should inconsciouly stick to, no rule about how things should sound or even look. It is relatively easy to make modules – no wiring, no exotic parts, no official panel material and font and knob model to exactly match. It became a medium of expression for people with new ideas about synthesis or compositional processes. People who knew their voice wouldn’t be heard if they released a plugin.
The Eurorack market is changing right now, more big companies enter (Roland, Moog for example), and there are more small manufactures than before. Why do you think this is happening?
Modular seems like a very niche market. Where do you see the future of this market in a few years…
Maybe someone at Moog or Roland realized it’s a good platform to release more experimental ideas – things that wouldn’t make sense at a scale of 10,000+ units, but which are worth exploring at the scale of 1,000+ units.
Maybe they thought that this is where a lot of
exploration is taking place, and they want to be seen there, to show
that they care about innovation. I can’t tell if this is driven by
genuine curiosity or trying to get a foothold in a growing market.
As for the next few years…
There’s currently an unhealthy hype surrounding Eurorack modular systems. It seems to me that half of the people who are buying modules at the moment don’t actually need them.
Of course, there’s curiosity, a lassitude from computer environments that I can understand, but also posturing and pathologic gear lust.
This means growth for manufacturers, which (at least in my case) is
difficult to manage:
hire people and you’re in the situation where you become a corporate organization (whose primary function will be to sustain itself and preserve the livelihood of its employees), or stay solo and deal with the surge.
Production volumes reach levels that are too high for local electronics manufacturing companies, you start hearing that at this level, there’s no reason not to outsource to Asia…
Such situations question the “boutique” ethos we strived to keep. So the first consequence is that existing manufacturers will have to lose a part of their soul in this growth, some of them will become larger, more traditional companies – more risk-adverse too.
If the growth continues, we’ll start seeing players like Korg, or even Behringer selling modules too: it might accelerate the growth of some of the existing manufacturers for them to become more competitive, or it can abruptly stop their growth. It’s not clear to me if, and when, the bubble will burst, but I can’t really expect every electronic musician in 2020 to own a modular.
The ”poseurs” will have moved to something else, along with musicians who ”ride the wave” out of curiosity and jump from one technology to the other (VSTis in the 2005, analog in 2010, modular in 2015, whatever it’ll be in 2020…). It’s not clear if the ones left would be just the “revivalists”, or if it’ll remain a platform for exploration.
In any case, the ideas put forward by this generation of module makers will have “trickled down” and will be found in plug-ins, DAWs, desktop units – some of them manufactured by modular synth companies who would have enlarged their scope (as MakeNoise is doing with the 0-coast).
Do you think that the Eurorack modular “scene” of our time, will have an effect on music history? Let’s say like 70’s Moog ? or the Buchla ( i can think of records like “Silver Apples of The Moon for example” Do you think it’s changing or will change the way people make or listen to music ? Or is it just an updated revival of 60-70s synths?
We can look at an original Moog or Buchla system and realize that there was quite nothing like that before. These were revolutionary, indeed.
I don’t think the Eurorack scene is as revolutionary, but still, I wouldn’t say that it’s just a “revival”.
New things are actually being
designed and built – a contact microphone into Rings, or an Euclidean
sequencer triggering a Shapeshifter don’t sound and feel like anything
It’s just not “new” in the same way that what Buchla or Moog did.
This is not a negative judgment by the way; in any field of science there are breakthroughs,
and then, decades of refinement and development of theoretical frameworks.
Maybe to elaborate more on the question above is it a paradigm shift or a re-use of the same paradigm?
We’re just advancing, digging deeper, pushing the boundaries, sorting our ideas, finding the hidden connections between them… within the ”control voltages between signal processing building blocks” developed by the forefathers.
From all your modules that I use, I find ripples to be most different, what inspired you to design this module?
Before making modules, I was selling this DIY hybrid monosynth kit,
which had digital oscillators processed by an analog VCF – there was
actually a bunch of different filter boards for it, with different
flavors of VCFs.
“Crazy digital oscillators into classic analog VCFs “ was kind of my “signature sound”.
So when I started designing my first modules, it seemed absolutely necessary to have an analog VCF in my line. I took the circuit I liked the most among those I designed for the monosynth, and which was the most different from what was available from Doepfer… and adapted it for Eurorack.
Do you use your modules to create music?
I play with them a lot during the development process, even if it’s not with the intention of actually recording music. When I get ”trapped” in a module and that hours and hours pass during which I’m just playing with it, completely absorbed by the sounds it makes, it’s a good sign it’ll be a success. After a module is released, I avoid playing with it, because I am completely scared about finding a flaw, or hear a tiny glitch, or find something frustrating about it that would either make me regret releasing it, or would push me to compulsively tweak it and rework it.
I really need to move on after that.
I also think that there needs to be a sense of “magic” in the
relationship between the musician and the instrument. A sense of
wonder or surprise, of exploration. After having spent six months
tweaking and tuning every aspect of a module, knowing each line of
code and all the details of its algorithms, it’s very difficult for me
to find this “magic” in it anymore. Somehow I have more fun playing
with modules from other brands.
I try not to feel bad about it – once you’ve listened to a track hundreds of time through mixing, mastering… you’re also tired of it and thinking of what to make of all the bits you couldn’t put in it, right? It’s just part of the creative process.
What was the inspiration to create “Braids” the macro oscillator?
Around the end of 2011 I considered making a factory-made version of my DIY hybrid polysynth, Ambika. Contrary to its DIY counterpart, it would use a more powerful ARM processor, so I started writing the code for the oscillators to see how much CPU power was available. Just like on the Shruthi, each oscillator offered a variety of synthesis techniques, with a straight forward control scheme in which the synthesis parameters are “collapsed” to meta-parameters.
I wrote about one model/algorithm a day. When I decided to cancel this project, I realized that all this code would make a good oscillator module. I actually had requests for a Eurorack “Shruthi oscillator”. This seemed all the more necessary that I noticed how some manufacturers were releasing several digital oscillators or FX modules with the same hardware. If the same digital hardware can realize 3 or 4 or even 40 different synthesis techniques, why should they be different modules I thought?
There is something unique about your modules, they don’t fit in west or east cost philosophies,
I know some of that debate is artificial, yet one can see that some other manufacturers are inspired from the west cost or east cost thinking when they build a synth or module.
What’s your take about this debate; do you think it’s still a valid schism in 2016?
There’s no debate or schism here for me, these are just two historical approaches to sound synthesis. Musicians mix and match building blocks from each of them, and of course there are tons of synthesis techniques (especially digital) that were not represented in Moog or Buchla systems and that exist outside of the East/West coast nomenclature.
They are not even mutually exclusive, Serge is a good example of that – it has VCF (so east coast!), but the DUSG is even crazier than the 281e (even more west coast than west coast!).
We should leave that to drunken conversations or history books for the coffee table, it’s no more relevant than Coke vs Pepsi or 808 vs 909 or Beatles vs Rolling Stones.
Do you see your self, creating your own “Buchla Easel”? A musical instrument that is made with specific modules?
Three years ago I had fairly limited ideas on synthesis. I did not
question the (Digital) VCO->VCF->VCA + LFOs + ADSRs + mod matrix
Designing more and more modules made me question that, and realize for example that with digital synthesis techniques a VCF or even a VCA are not necessary. At the moment I don’t have any strong idea about what’s the ideal “recipe” for sound generation I’d like to put in a standalone instrument.
There is something unique about your products, from the small “Hindu” statues, the names they have, how they function and sound. Can you try and explain that?
There are a couple of idiosyncrasies in the way I work. One of them is that I’m not in the least bit interested in reproducing Buchla/Moog modules, and that I’ve already paid my tribute to classic 80s synths with the Shruthi/Ambika.
So there’s a lot of territory with
a big “let’s just not do that again” sign on it.
The second is related to my design process. Design isn’t centered around a specific synthesis technique or building block or circuit, but rather a “goal” or “activity” (“create textures from incoming audio”, “make organic sounds”).
I decide on a set of sounds I’d like to achieve and that fits this overall “goal”, and I try to project this unto 4 or 5 axes which could be interpreted as perceptual dimensions, or words we’d use to describe how we manipulate the sound.
From there, I try to figure out which signal processing techniques
would achieve this result,
and which parameters, or combinations or parameters would correspond to my 4 or 5 dimensions.
The question is:
what kind of sounds or effects can be considered as belonging to the
same family, and then how could we sort or organize this big family. I
also spend a lot of time calibrating the parameters so that good stuff
can be picked all over the course of each potentiometer.
A test I frequently perform during the design phase is to generate 50 sounds by randomly setting all the knobs. A majority of them should be pleasant, and they shouldn’t resemble each other too much. I think this particular approach is the most specific to Mutable Instruments, and the one that makes the modules successful, if somewhat controversial.
Some people would say it is “all sweet spot”, some would emphasize the “musicality” or “beauty” of the sound, while others would argue that there’s some cheating involved here or it’s no more than changing presets because it can’t be made to sound bad. These comments are all an immediate consequence of this choice to articulate the design around the space of possible results, instead of exposing the internals of the algorithms.
Most designers would think “Let’s do an FM synthesis or granular module, let’s look at each parameter in the signal processing algorithm and put a knob and CV input going from the minimum acceptable value to the maximum acceptable value”.
This is how we get modules like the Shapeshifter or Akemie’s Castle where there’s sometimes little correlation between each knob and the musical result (you can get the same result by turning two different knobs, or you can’t predict how things will sound when you turn a knob, or you don’t know which knob to turn to get such or such result, or most combinations of knobs yield screeching noise).
I work the opposite way - my ideal is that the knobs act as some kind of coordinates or ”paths” through a space of related, meaningful sounds. I don’t see that often in other brands’ designs, but the Erbe-verb is kind of similar in design – in that they did not just put a pot/CV for each possible parameter of a reverb, but tried to come up with a small set of important control dimensions, even if these are, behind the scenes, acting on several internal parameters of the algorithm. In parallel to that, there’s the question of “sonic” aesthetics. I try to give my modules a lot of headroom, in the figurative sense – they won’t sound harsh or noisy or distorted unless parameters are set to the extremes. I don’t treat modulations as things that should make sounds go wild, but rather as “gestures” or “expressivity controls”.
Audio rate FM or feedback or self-patching are rarely rewarding strategies with my modules. Patching slow, subtle modulations, smooth random sources, touch plates, sensors, ribbons into their CV inputs is much more interesting, and it’s a priority for me that the modules should handle layering very well (a single sound should not “steal the show” – the spectrum should be sparse and clear enough to allow several voices to be stacked together without getting too muddy).
no wonder that the people who seem to get the most of them are usually
doing layered ambient stuff.
The visual aesthetics are the result of a month-long battle with the
designer who worked on the project, and whom I exposed to all kinds of
things I liked, from classic book covers to Indian textiles.
I wanted the modules to evoke a different kind of time and place, not a 70s research lab. They shouldn’t put you in a “science-y” mindset.
They shouldn’t quote earlier electronic instruments.
In his book “You are not a gadget” Jaron Lanier, define the idea of “Lock downs”, he gives midi as an example for a lock down, do you think in synthesis musical instruments there are also many lockdowns?
Yes, plugins are similar in that they separate audio streams and
control streams – and the control streams are structured as MIDI.
There’s too much “typing” in there, and there are kinds of plug-ins
that we just don’t see because of that. For example this architecture
is not really designed for event processing types of plug-ins, or for
working at the micro sound scale where there’s no clear boundary
between events and signals.
Even environments like “Puredata” have different paths and rules for
audio buffers and the event-driven stuff (messages).
We need a more purpose-neutral representation – a bare signal in time,
and that’s precisely what a control voltage is…
What’s so interesting with Eurorack is actually that it implements what programmers call “duck typing”.
“If it quacks like a duck and walks
like a duck then it’s a duck”.
For example a sequencer module won’t try to check that what it receives on its clock input actually comes from a clock generator: if the signal has sharp rising edges or if it goes over a given voltage threshold, then we’re good, the sequencer will go to the next step.
“If it ticks like a clock then it’s a clock”.
That’s why within a Eurorack system, you can clock your
sequencers by LFOs or VCOs or random sources, send pulse patterns from
a rhythmic sequencers to a VCA CV input, send audio signals to
modulate VCF frequencies, etc. There aren’t many systems offering that
kind of freedom –
even Buchla is “typed” for example, with different kinds of connectors for audio and control signals. Actually, if you think of the most basic thing about music – that it’s something that evolves in time, the control voltage appears like the most obvious mean of interfacing and controlling instruments, or representing musical information.
Another issue resembling a lock-in is the way some tools enforce
things like the 12TET or specific rhythmic divisions.
Finally, but I’m not sure if this fits in the “lock-in category”, but the idea that an envelope should be A, D, S, R ; or that an oscillator should be square, saw, triangle – ideas that synth manufacturers repeat without rethinking… it feels like a lock-in.
Do you think the surge in modular and hardware synths of our time, as a reaction to the success of the DAWs?
Yes. There are many reasons for avoiding DAWs – the overwhelming “anything is possible” feeling they give us, the lack of a tactile element, the fact that it runs on a general purpose computer, the lack of “wonkiness”. Modular synthesizers are a way of addressing these problems, but certainly not the only solution of course.
orchestras” – small networks of “Korg Volcas”, Pocket Operators, pedals,
could be another approach, and there’s also a move towards that.
How do you see the integration of DAWs and Modular in the next few years?
I predict that a big manufacturer of audio interfaces will release a product specifically dedicated to sending / acquiring CVs and triggers, with accurate V/O calibration and handling of Eurorack signal levels. Or maybe that audio interfaces will have a few channels of DC-coupled, calibrated I/O. And Logic or Live will have CV tracks. More synths with CV/Gate too!
You worked in Universities and IT companies, how does it feel to run your own operation?
I like setting my own priorities, chasing stupid projects, obsessing over details and working at my own rate, even if it means doing nothing productive for two months and then rushing for the next two. So it’s much better being on my own.
by Yair Etziony