Mutable Instruments make some of the most popular Eurorack modules we stock at Red Dog Music. From the oscillations of Braids to a unique take on the looping ADSR with Tides, a Mutable Instruments module might be just what your system needs. We caught up with Mutable Instruments’ Olivier Gillet to find out a bit more about what drives the development of a modern module…
First of all, are you able to sum up the philosophy behind Mutable Instruments?
To make it simple: just pouring my heart into the modules. Working only on the things I’ll be very excited about or that I’ll have a lot of fun making. Obsessing over the details that matter to me, making sure everything looks and sounds beautiful.
It’s a mix of being very selfless about a lot of things (to the point of purposely making bad business decisions, almost giving away stuff), but being selfish enough to resist the temptation of giving people what they ask.
Looking at your personal background, where does music fit in with the science and the academia? Can you tell us something of your own influences and musical background? Listening, playing, performing?
During my studies, the criterion I have always used to pick my classes was: “is there any connection with music”; and I naturally picked research projects focusing on music. For many years my main research interest was the algorithmic emulation of aspects of our listening experience, such as telling apart instruments from a mix, picking up the beat/tempo, recognizing changes in instrumentation or structure/build-ups – all these problems combine signal processing, machine learning and musicological insights. Science is a means of transportation, rather than a place in itself, it can be used to explore any field.
In my childhood, the music that got me hooked to electronic/synthetic sounds was british synth pop. Some musical encounters that deeply affected me: the Bristol trip-hop scene in the mid 90s, the em:t label one or two years after that. In the early 2000s, Broadcast and Boards of Canada – but also Éliane Radigue, Terry Riley, lots of early/baroque music, classic hindi film songs… Some bands that have remained companions to my ears throughout the years: Stereolab, the High Llamas, the Sea and Cake, and others in their circles. They are responsible for the music I’ve listened to the most. More recently I’ve become fascinated with the universe of Ghost Box Records. I am not sure if there’s any connection between all this and my work as a module designer – maybe the link is that I like artists who manage to carve their own fragile universes, and wonder – what would the music (or artefacts) from this imaginary place and time look like?
I started making music with trackers and my own (non-realtime) softsynths in my teens (stupid techno/trance), then graduated to a small home-studio with an akai sampler and a few Roland vintage synths (ambient techno of sorts), went back to matlab+trackers (dumb 8-bit stuff), then a combination of lack of money, time, space; and the boom of file-sharing (revealing my own mediocrity compared to whatever I could download) put an end to all of this around 2002 – I kept the gear and played it without trying to work on a “finished product”. I kept following what was going on in software and hardware. When I started building synths in 2009 I really felt that it was what I was meant to do, that I wanted to make “meta-music”, working on the composition and sound generation processes themselves, rather than using them to instantiate a specific piece of music.
How does the inspiration for a Mutable Instruments module come about? Do you see scientific or mathematical techniques and think ‘that would be interesting to do something with in a module’, or do you see a gap in the music technology first and identify the best way to fill it?
Sometimes it’s just scratching a personal itch. Shelves is exactly like that – something I built just because I had a vague idea about how to do it and I wanted to see how interesting it would be. Streams started as a double VCA which looked too similar to a competing product, so I added a MCU and AD/DA converters to it and tried to think of what kind of things this hybrid architecture could do. Some modules have been made for the sole purpose of allowing people to have only Mutable Instruments modules in their case (Yarns, Links, Shades, Peaks…). For Clouds I worked from the concept of “texture synthesis” without really thinking about a specific synthesis technique (the first prototypes used spectral instead of granular methods). Same thing for Elements: I had the vague idea of a “subject-verb-object” approach to synthesis and it turned out that modal synthesis was the best way to do it. I never say “let’s make a granular module”. I tend to think about “activities” or “spaces of sounds” and build a module around that, rather than take the synthesis technique as a starting point and put one knob/CV per parameter.
In parallel, whenever I have a technical idea or find something interesting in a paper, I build a little bit of code – and keep all these organized and ready to be transplanted into something bigger later.
All the Mutable Instruments modules – hardware and software -are open source, what was the reason behind doing this? How do you balance progress through academic freedom with running a business?
Plenty of reasons…
I come from an academic background in which sharing is the default setting. Earlier, I’ve seen in the tech industry the complete waste of time that patents or protection against reverse-engineering or software copy can be.
I have often seen companies, especially in the audio/music instrument space, hype proprietary algorithms which end up being hacks once you look more closely at how they are made… This is kind of lame, so going full open-source is a way of disciplining myself and not over-hyping my products. They are what they are, you can have a look at how they are made, and 99% of it mostly look like this.
I have often been frustrated when using audio gear from other brands – wished the range of this or that parameter was larger, wished the UI behaved a bit differently or the order of items in a menu was different. I want to give (technically minded) musicians the possibility of fixing things they don’t like, or fix for others and share. This is also a way of solving some conflicts regarding design philosophy. For example, to me, internal modulations enabled through menus are a bit too much in a Eurorack module – that’s a line I’ve drawn. But what if some people were actually okay with that and liked very deep menu-driven modules? It would be sad if they suffered from this limitation – it’s just software! So a guy created the “Bees-in-the-Trees” firmware for Braids that added a ridiculous amount of internal modulations.
Graphic design aside, I am responsible for everything in my modules, so there’s nobody to review my code or schematics. Open sourcing the modules helps a bit, it forces me to make my code as clean as possible, and puts me in a position where I can’t sweep too much dust under the carpet… And some of my customers have far more experience with electronics or software than I do, so once in a while they actually spot things I can improve!
I don’t worry that much about hypothetical cheap chinese clones. Beyond the code and schematics there’s capital, know-how, reputation, customer care – people smart enough to figure out all these hard details can very well hire the right guy and make something original. One thing to remember is that whatever Mutable Instruments product is popular on the market at the moment mirrors whatever was going on in my mind 12 to 30 months ago. If shit happens, I’ll just speed up the release of the next big thing. If you’re thinking of blindly cloning something… keep in mind it’s almost already outdated 🙂
Looking at the current range of Mutable Instruments modules – the complex timbres of Braids, the classic filter-to-VCA of Ripples, the Vactrol-modelling VCFA of Streams for example- it seems they would not fit into a straightforward ‘East Coast’ or ‘West Coast’ convention. In the current Eurorack world with such a huge range of modules and people often choosing a diverse range of modules from different manufacturers for their own systems, do you think these types of definition are still relevant?
The east-coast / west-coast split is still relevant when we’re talking about “modules as a functional blocks”. Then it matters into what kind of blocks the signal processing chain is split into… But I think we’ve passed this point now. Nowadays, many modules embody complete compositional processes or “spaces” of related sounds – it’s not just me with Clouds, I see that in the 4ms SMRF or the Make Noise Erbe-verb. So it would make more sense to me to classify modules depending on whether they are exposing the guts of a single operation, or are a curated set of operations capturing a coherent sound or compositional space.
Over the last few years, we’ve seen a huge resurgence in the synthesiser market, and analogue synthesisers in particular. Despite the choice available, the modular market seems to be growing even faster. What do you think is so appealing about modular synthesis at the moment?
We already know about the usual suspects: haptics, the sense of “owning” an instrument optimized for your compositional process… but to me, the most important appeal is the feedback loop between the public and the manufacturers, and what emerged from it.
To manufacturers, the barrier to entry is relatively low. A single person can design all aspects of a module (with a sense of aesthetic coherence), and you don’t need to make thousands of units to break even. So it’s an interesting medium for people who have ideas about synthesis and compositional processes. And it attracts the crowd who wants to experience the bleeding edge. And because there’s such a public very open to new ideas, it’s an additional incentive for manufacturers to try to come up with new ideas.
At the risk of sounding elitist, there’s also a useful “noise gate” effect – because it’s a physical product, it costs money to make a batch of modules – so this prevents the market from being flooded with too many ideas (unlike free VSTs or open-libraries of pd patches for example). I’m not saying all Eurorack modules are amazingly original and high-quality – just that it would have been much, much worse if they had been as easy to make and duplicate as a plug-in or Max/MSP patch. Reversely, because people pay for modules, they tend to be more thorough in their exploration, and try to establish a more deeper connection with the instrument. Versus the “don’t care” attitude we can have with software, especially when we don’t pay for it.
It thus appears to me that Eurorack modules are currently the most efficient way of getting a new idea (about a new family of sounds, a new compositional process – no matter how small it is) into the hands of musicians who will care about it and try to make use of it.
From the point of view of a module manufacturer, does the current Eurorack market give you the flexibility to be able to produce the modules you want to make and find interesting without having to engineer features and specifications to the market?
Yes. That’s why I’m here!
It certainly seems as though the modular world is where the creativity in synthesiser engineering is actually able to make it to market. Do you think any of these technologies and features will end up in a standalone synth from any of the traditional ‘big’ manufacturers?
The oscillator section of recent VA synths are moving away from simple waveforms, and are starting to look like “islands-of-related-timbres” – Nordlead A1, Yamaha Reface CS, the latest Roland System-1 upgrade, Modor NF-1 are all doing that. It seems to me that Braids played a role in this trend.
I expect to see more euclidean pattern generators, wavefolders, slope generators (instead of LFOs/ADSRs) in future hardware synths. If it doesn’t happen fast enough, I’ll take a break from modules and make it happen. Or maybe Yamaha will acquire me.
As we’re in the middle of an analogue revival, it’s interesting to see that you’ve gone digital to really push what a single module is capable of. We hear a lot about (with a lot of it hyperbole) the word length and sample rate for AD/DA when it comes to audio frequencies, does working with control voltages cause any particular challenges with a digital system?
In a fixed-architecture digital synth, hardware or software, modulations are essentially slow and “well-behaved”. With a voltage-controlled digital module, the musician can throw any kind of crap to the CV inputs, not just nice envelopes or LFOs. And some
classic DSP methods do not work well when control parameters are modulated at audio rates – they become unstable, suffer from aliasing, or the computational costs gets very high. Over the past 10 years there has been a lot of interesting research to improve things (for example for digital models of analog filters).
I feel that sometimes, the ability to modulate a control signal at audio rate, is a bit silly – in a physical model of a string, what does it really mean to modulate the stiffness of the string at audio rate? Of course there’s laziness from my side when I say that allowing audio-rate modulation of such or such parameter is a waste of computational resource and that the input CV should just be low-pass filtered… but there’s laziness on the other side, from people to whom the only strategy for building original sounds is to blindly modulate everything at audio-rate.
Back to the issue of CV acquisition, there are also hardware limitations – the AD converters built into various microcontrollers do not have a very high resolution (12-bit with some noise), so you have to intelligently pre-process the signals to get stable, reliable data. Apparently Roland uses audio-grade 24-bit AD on its digital modules, wow. This sounds a bit overkill to me – so many analog modules have a signal to noise ratio that doesn’t exceed 75dB anyway…
In the end, this is what engineering is all about – finding a meaningful trade-off between quality and cost, identifying diminishing returns… Going on the TI website and choosing the most expensive ADC or op-amp money can buy, then brag about the numbers is not engineering.
To many people, the image of the modular synthesiser is still stuck in the 1970s, despite the market continuing to grow and modular synths becoming more visible in the mainstream – Jonny Greenwood and Surgeon’s dj sets spring to mind. Do you think we can expect to see modular synthesisers in more places in the near future?
Not even in the future. They are already everywhere! As the main instrument for live techno sets… in the pedal-board of post-rock guitarists… playing sequences in the intro of an indie-pop song… in the soundtracks of movies… synthesizing sound effects during the production of a video game… or even powering art installations…
Do you see the current Eurorack/modular renaissance as being driven by engineers and musicians who are driven by nostalgia for that 1970s scene, or who see the format as the ideal format to push the synthesiser forward and create sounds at the limit of what is currently possible with analogue and digital technology?
There are several generations of instrument makers, with different goals and motivations. But I think my generation is not interested in recreating the Moog/Buchla canon. We respect it, we’ve been curious about it, but we want to build what’s next.
Finally, can you talk about some of the ‘inspirations’ we can see on the ‘About’ page on the Mutable Instruments site?
Besides synthesizers and music, my other interests include cinema, literature, textiles, and thinking about the future.
Red Dog Music is the UK’s friendliest musical instrument and pro-audio dealer. Between our 5000 square foot Edinburgh shop filled with an incredible range of products, and our London showroom in Clapham specialising in high-end instruments, dj and pro-audio, Red Dog Music has you covered from north to south and from performance to playback.