People & Music Industry

Back in 1996 Hugues Vinet of the IRCAM electronic music institute in Paris, was interviewed by Paul Tingen for an article in Sound On Sound magazine. Now the two are back together discussing the musical and technical innovations of the past 25 years.

Chapters
00:00 - Introduction
01:43 - Hugues Vinet and IRCAM
02:38 - How IRCAM Has Changed
06:02 - Partners and Community
09:03 - Music Styles and Technology
10:56 - Working With Ableton
12:56 - Musical Developments
18:49 - Sound Design
23:05 - Defining Sounds
26:04 - Using Different Tools
28:07 - AI Development

Hugues Vinet Biog
Hugues Vinet is Director of Innovation and Research Means at IRCAM, a world-leading institution associated with the Centre Pompidou in Paris, which is dedicated to music production and research. He has managed all research, development and innovation activities at IRCAM since 1994. He co-founded and ran for several terms the STMS (Science and Technology of Music and Sound) joint lab with the French Ministry of Culture, CNRS and Sorbonne Université.

Hugues previously worked at the Groupe de Recherches Musicales of National Institute of Audiovisual in Paris, where he managed the research and designed the first versions of the award-winning real-time audio processing product, GRM Tools. He has coordinated many collaborative R&D projects, including recently H2020 VERTIGO in charge of the STARTS Residencies program managing 45 residencies of artists with technological research projects throughout Europe. Hugues is currently in charge of the artistic residencies program of the MediaFutures project. He also curates the Vertigo Forum art-science yearly symposium at Centre Pompidou, and participates in various experts bodies in the fields of audio, music, multimedia, information technology and innovation.

Main Ircam site - https://www.ircam.fr
Ircam Lab Software - https://www.ircamlab.com
Free Ircam Technologies - https://forum.ircam.fr/collections/detail/technologies-ircam-free

Music Credits:
Répons by Pierre Boulez
L'esprit des dunes by Tristan Murail
Côte des Bars - by sound designer Roque Rivas for Krug Champagne

Paul Tingen Biog
Paul Tingen has been a contributor to Sound On Sound since 1990. In addition to his regular artist and producer interviews, he began writing the Inside Track: Secrets of the Mix Engineers series in January 2008.

He is also the author of a book, Miles Beyond: the Electric Explorations of Miles Davis, 1967-1991 (Billboard Books), a producer, and a guitarist with one album to his name.

Paul maintains two web sites: www.tingen.org and www.miles-beyond.com, and an Instagram page: @paultingenmusic

Catch more shows on our other podcast channels: https://www.soundonsound.com/sos-podcasts

Creators and Guests

Host
Paul Tingen
Paul Tingen has been a contributor to Sound On Sound since 1990. In addition to his regular artist and producer interviews, he began writing the Inside Track: Secrets of the Mix Engineers series in January 2008. He is also the author of a book, Miles Beyond: the Electric Explorations of Miles Davis, 1967-1991 (Billboard Books), a producer, and a guitarist with one album to his name.

What is People & Music Industry?

Welcome to the Sound On Sound People & Music Industry podcast channel. Listen to experts in the field, company founders, equipment designers, engineers, producers and educators.

More information and content can be found at https://www.soundonsound.com/podcasts | Facebook, Twitter and Instagram - @soundonsoundmag | YouTube - https://www.youtube.com/user/soundonsoundvideo

Welcome to the sound on sound people and music industry podcast channel in this episode SOS author Paul Ting and talks to Hugh Vinay director of innovations and means of research at the IRCAM Institute in Paris

The legendary French IRCAM Institute was founded in 1970 By composer Pierre Bole at the request of then President George Pompidou. The institute continues to be 100% funded by the French government. Here comes stands for Institute National Acoustic Music. which translates as Institute for Research and Coordination in Acoustics slash Music.

IRCAM is famous for developing a wealth of groundbreaking music technologies, and for playing a crucial role in the creation of avant garde electro acoustic classical music. Much of the software developed at IRCAM has also found its way into the mainstream, and some mainstream musicians have found their way to IRCAM.

For example, Frank Zappa. Sound On Sound first visited IRCAM in 1996, which resulted in an article that was published in the December issue of that year. Almost exactly 25 years later, we revisit IRCAM to see what is going on there today. We talk with Hugh Vinay, Here comes director of innovations and means of research.

Remarkably, Vinay also featured in our 1996 article. Hello Hugues, welcome to this Sound on Sound podcast. First of all, you've been working for IRCAM for a long time. How long exactly has it been? Well, yes, I arrived in October 94. And so, uh, I've stayed actually at this position until then. That's a long time.

So you must really enjoy your job. Yes, exactly. I didn't find better elsewhere. So do you have a musical background yourself? Yes. Uh, so, so I, I'm a pianist. Uh, I did the classical studies and then I played jazz with people. Uh, so we did concerts, we, we did a record, et cetera. But, uh, now I'm too busy to, to continue, but I hope I will be able to, to do it again later.

But your, your job at IRCCOM has a more technical focus. So what, what is your technical background? Yes, I'm, I'm an engineer in, in digital, uh, single processing. This is my training. Okay. So, let's get to IRCAM. So, um, so in 1996, I wrote an article and about IRCAM and I came to Paris and you were one of the people I interviewed.

Yes. I've been looking at the website of IRCAM and it's incredibly rich. There's many, many, many, many aspects and you could, you could spend two hours just on the website. But if you would have to encapsulate, uh, Today, how is IRCAM different from 25 years ago? What are the core shifts in focus of IRCAM?

Well, uh, of course, there are many technologies that did not exist at this time, I would say. Uh, for instance, the generalized use of AI, technologies like Web Audio. But there is also continuity, you know. Uh, so digital signal processing is still, uh, main focus. 3D audio specialization too. Uh, Maybe there has been also the rise of, uh, sound design, and I could explain in more detail what I mean by that, which is not exactly the usual meaning of sound design.

That is a functional role of sound in everyday's life. Also, when I arrived, there was almost nothing related to new instruments and gestural control of music, and so we developed also an axis of research So, yes, uh, and many other features, for instance, computer, uh, added orchestration, et cetera, so there are many things that we can speak about.

And even in the themes that we were already working on, like 3D audio, there has been, of course, many, many development, uh, since then. So would you say, because I mean, actually IRCAM was started, was an initiative by Georges Pompidou, who asked Pierre Boullée, uh, to start an institute that was dedicated to musical, acoustic, and computer research, and it was going to be very prestigious because it was right in the center of Paris, right next to the center of the Pompidou, um, um, So would you say that the, the, the main focus of, or the main purpose of IRCAM is still the same?

It's just a matter that you've, you've adapted to how technology and the world in general has evolved. Yes. And it's even more true today because, uh, Boulez, At the end of the 70s, had a vision of the needs of contemporary music production For advanced technology and it has become true and And so, so really, uh, the, this project is completely true.

And this is the main dynamics of, uh, activity at ACAM that is the specific, of course, we are, we are a big lab. Uh, we are more than, uh, I think 120 people are working on various, uh, research and technological aspects. But really, the main aspect is that this research is first applied to a contemporary music production, and so we host many different artists at IRCAN, and they use the technology, and so we have to develop it in a generic way so that it fits to very diverse aesthetic needs.

And so, yes, Boulez, I can say that Boulez project is, uh, still really the heart of IRCAM's dynamics today. But the people who are employed, I mean, does the French government still pay for IRCAM? Yes. Although you also have commercial partnerships, don't you? Like I noticed you had like with Renault and other big companies where maybe the sound design or, uh, there are certain connections you have with the business world as well.

Yes, and I should develop that aspect because we are very active in that. Um, So there was a first circle of, uh, dissemination and diffusion of our technology, uh, which is the IRCAM Forum, uh, which was founded about a little before I arrived in 94. And the goal here of the forum is to disseminate, to distribute all the software that is produced internally for our, for production to external users.

And it has grown a lot. I think now we are about 20, 000 users worldwide. So it's quite a big, uh, initiative. Yes, international community. But this is not commercial. Uh, you just, uh, register, uh, on a yearly basis and you get access to the software. Because I noticed on your website, you have a lot of free software to download.

Also, uh, but the main, uh, channel, uh, of diffusion is the forum, uh, because there are also, uh, a lot of documentation. There is a community. You can join a community and people, uh, so there are many discussion groups, if you like, on the various software. And then, of course, we have been very active in collaboration with companies.

Uh, first, it was focused on companies, uh, who develop tools for, uh, sound production and music production. And so we have had many licenses, for instance, of SuperVP, that you may know, which is, uh, an engine, uh, that does this. High quality time switching, transposition, et cetera. Um, and then we developed also other fields of expertise in relation to industry.

For instance, uh. At the beginning of music information retrieval, it was at the end of the 90s, we developed also some specific research. We coordinated the first large projects, European projects on that topic. And since then, we have developed a lot the technology that we have licensed to major actors like Universal Music Group, etc.

And now, uh, so, uh, two years ago, I know, yes, it was in 2019. We, we founded a new subsidiary of IRCAM, which is called now IRCAM Amplify. And so now this is really an independent company. So, so IRCAM is a majority shareholder. And the goal is to, to, uh, to bring a new step with investors, uh, on, on the, the application of IRCAM's research to, to, to the world, to, to the industry, et cetera.

The generalized power of sound, uh, so application in many, many aspects of, uh, our lives, not only professional, but also daily life. If you look at the musical side of things, I mean, are you still very much geared towards classical music or are you, uh, because of sound design, obviously, you know, we have all the plugin companies that, and the, and the door companies, you know, from avid and, and, and, and, and American Germany in particular, many companies that develop plugins and sound shaping.

Um, so there's a whole, when we spoke last time, all these boundaries were getting blurred. I mean, does that mean that you're also reaching out more to those kinds of companies? Yes. And that kind of technology, are you more geared towards the classical world still? No, and even, uh, a lot of our technology is also used, uh, in sound production, not only in music.

Uh, for instance, there is a collection, uh, with our partner Flux in France called Earcam Tools. Uh, so there is a series of plug ins, and for instance, there is a plug in called Trax, uh, which implements, uh, advanced sound, uh, audio. Yes, transformation of voice. So you can change, uh, voice in real time from male to female, et cetera, to elderly person.

You can change the, the, the, the pitch contour, et cetera. And this is used, uh, it can be used in, in music, but, but also in theater, in, in, uh, and a lot, of course, in, uh, in video and movie production. And also, in this collection, there is also the SPAT, our 3D Spatialization Engine, and it is used also in all sectors of sound production.

So, I would say that, yes, the basis of our musical production at IRCAM is, yes, focused on contemporary music production. This is true, but then our tools, the goal is that our tools can be applied to, to much broader, uh, musical, uh, aesthetics, to contemporary music. Because also I noticed that this software language Max, um, that went to Opcode and now it's owned by Ableton.

No, it's, yes, it was owned by Cycling74 after Opcode and Cycling74 was then bought by Ableton. Yes, you're right. So does that mean that your work in conjunction with Ableton? Yes. So what, what happens on that front? Is that developing Max or what is Ableton doing? Is Ableton using your technology and applying it or your research?

I mean, how does that work? Actually, they did something very clever with David Zicharelli, who was CEO of Cycling74. They, uh, they were very complementary because Max enabled to, uh, build some musical application, uh, using high level programming. And now this can run within live. And so, uh, IRCAM, uh, the, the, uh, ACAM's main tool for music production is still Max, uh, because Max is able to combine various, uh, music and sound processing modules.

And these modules are the basis of what our research produces. So the research produces, uh, advanced algorithms, and then these algorithms are packaged for our production as, uh, Max objects. And then this max objects can be combined, uh, for a given application, uh, uh, in, uh, in a plugin that will run on, uh, on live.

And so the workflow is, is quite a straightforward, uh, and it's only high level programming. You don't need to, to, to be a, an engineer to do that. And so, yes, we also have a collection called, uh, Ircamax. That is sold by Egleton, and that implements some, uh, specific sound processing modules, uh, and for that we, we developed them, uh, with Max.

Interesting. So anybody else in the world of electronic music can, can easily work with those tools? Yes. So, how has the world of Contemporary music developed because of course, the, the, the trajectory, you know, harmony became more and more abstract. And we got eight, we got first, we got, you know, 12 tone and we got a tonal music.

And then it was under particularly Pierre Boullée in many ways was like, uh, an exponent of, of going beyond tradition in music. And we got unusual sounds and strange things. Then of course, electronic music. I mean, we had music concrete and then. the whole electronic music thing after the second world war.

Um, all that is kind of known, but what, what have been the, the developments in classical music? Uh, would you say over the last 25 years stylistically has, has there been attempts made to integrate the electronic experiments more with melody and harmony? Have there been other developments? I, because actually I'm not, I don't think the main, the public at large is very aware of what's been happening in contemporary music.

So, yes, I think one, one important, uh, aspect has been what we call computer aided composition. Uh, so this, uh, is, uh, a line of research and application, and the main software, uh, that we developed for that is OpenMusic, but there are also other implementations in other environments, and the idea here is to, uh, to model the various, uh, elements, uh, uh, The various elements of language, of music, notes, chords, rhythms, in a structured way, so that this can be manipulated by the computer, and then can be generated algorithmically.

And many composers have worked on that. And this goes much beyond what you say that there is a scales with a quarter tone, eighth tone. You can basically design any scale you want. And you can also elaborate rhythms with no limit in complexity. And of course, a main concern has been the link with the sound world.

In the second half of the 20th century,

uh, spectral movement, uh, with several, uh, composers, uh, who, who, who developed links between this symbolic world and the sound world. And for instance, who analyzed real, uh, life, uh, sounds, uh, into combinations of notes and, and they would then orchestrate these sounds, uh, with all the complexity of, of the sound, for instance.

And vice versa. Uh, Computer aided composition, uh, also consists in controlling synthesis from, uh, discrete symbols. Uh, controlling synthesizers, if you like. So it's, you consider it as, as a very broad extension of MIDI, if you like. And so, then you can produce, and the goal, I think that is not attained at all, that is there is still no generalized language of sound.

Thank you. And so this is really a main concern today in, uh, in modern music. Uh, there are practices that enable to produce, uh, a musical idea in sound, I would say. There are many practices that consist mainly in combining a series of processes, of actions to the sound in order to reach a given goal. But there is no generalized language that you can apply in order to generate your sound idea.

It's still a question of practice, of tools, and, but it's a goal that we try to, to, to reach little by little. So you're talking about the equation because of course in music, you know, we have music notation, which is only rhythm and pitch, really. So what you're talking about language, basically to symbolize like I have a sample or I've sound designed some kind of sound.

And what you're looking for is a way of representing that so somebody else would be able to recreate that more or less somewhere else without actually having heard the sound. Yes, this is the idea. And precisely we have worked in sound design. That is how so how to describe. Is it possible to have a vocabulary that describes the sound world in a structured way?

And so we have also worked on that, and it combines various, I would say, viewpoints on sound. One is, of course, the sound source. When you hear a sound, to which sound source does it remind you? Uh, so is it a percussion? Is it a boat? An object? Et cetera. So, physical, uh, mode of production of sound. And then there are some more abstract levels of representation, for instance, timbre.

And then this can be, uh, this results also from joint studies between human perception and also signal processing. And we are in this articulation between the physical world and the signal world. The single world and the perceived world. Because, uh, we perceive some things, but other So, there are specificities in our perception, for instance, masking, etc.

Or the way we hear pitch as a superposition of harmonics, and we hear We listen to a single tone, for instance. And so these aspects of perception must be taken into account in the way we design algorithms for producing a given sound, uh, idea. But you mentioned earlier on sound design, because it sounds like a lot of what you're doing, uh, would overlap, you know, with sound design.

Yes. also happening in movies and in, in, in, in all sorts of fields, even in pop music. And then, you know, people doing all sorts of sound design and trying to create things that sound just a little bit different. So can you talk a little bit about that? How, what your overlap is between what you do and sound design as it happens more like people who working indoors and they have pro tools or logic or whatever.

Yes. What I understand from the usual meaning of sound design, maybe sound effects, for instance, you add. Uh, noise in a soundtrack, which is not voice nor, uh, music. It's all the other kinds of sound that are called environmental sounds. In a movie, for instance. But what we mean there is that these categories of sounds can be also used in a process of design of an object or a situation.

For instance, in a car, uh, there are, uh, alert sounds. And these sounds are the object of a specific study. That is, which kind of sound should we play for a given situation. So this is the functional aspect. And then there is the aesthetic aspect. For a given brand, a given company, and a given brand in this company, which kind of sound would fit or not.

And you see, it's a discussion both on the aesthetic aspect and on the functional aspect. And so, yes, we, we worked a lot on, uh, on that dimension, uh, with many companies. I mentioned cars, but, but, uh, there, there are other applications, uh, uh, for instance, with high fashion. They, now they produce also objects that generate sound and that have their own identity.

And, uh, each time you have a, an object that, that, uh, produces sound, uh, People at the beginning did not ask themselves the question of, of the aesthetics of sound or its role, uh, in terms of information. But now it, it, it begins to, to be more and more admitted by everybody that, that this requires a specific process of design.

So that also Requires technology also psychology like the kind of the kind of sounds that for example if you have a car You know what the sounds you make in the car what what makes people comfortable what will alert him? What will maybe kind of really kind of shock them a little bit if necessary? So you you working then with psychology with acoustics all sort of fields Yes, and also we work, so basically it's a process in three phases.

There is a process of definition, specification. Then there is a process of production of sounds. And then there is a process of evaluation. And it's a loop. There can be many iterations in such a process. And so, uh, The last phase, which is evaluation, is probably the one we must do the most because, uh, we use, uh, methodologies of experimental psychology to do tests with users, to test with people what they hear, what, how do they react, uh, with a given sound, et cetera.

Uh, I think the one, uh, in which we were less advanced in the first phase of specification because there is no universal language for speaking of sound and so for writing, uh, I would say, a definition of work, uh, of a sound designer. And how to specify what we want to, to attain. This actually requires, uh, a joint process, uh, with the client, with the, the company who wants to have these sounds, and the people who will design it.

And so we have developed this kind of ideation methodology that, uh, if you like, uh, matches, uh, um, uh, the given categories of sound, how we can speak of sound and the given categories of the brand, and functionality that we want to attain. So you want to go beyond, you know, in, in electronic music, two guys will sit in a studio and somebody plays a sound and says, Oh, that sounds good.

And you said, yeah, good, but add a bit of low end, or maybe make it a bit more scratchy. Or so you want to go beyond that kind of those kinds of discussions and be able to define that essentially. Yes, actually, we already do it. We have done that for a long time, for several decades. I can give another example, which is quite original.

Uh, it's, uh, people who produce champagne. Uh, so we work with the champagne, uh, the crude champagne. It's, uh, very, very I think great company for, for champagne and, uh, they consider that the best conditions must, uh, must be set for, uh, wine tasting. And so, uh, you know, a champagne is a mix of many different wines.

Every year the mix changes. But there is a given set, I don't remember how many, I think it's 200 basic wines that intervene in a mix. And we work with them, and each of the basic wines, we did actually a sound design related to it. That is, you hear some sound while you taste the wine. And the dynamics of the sound follows the dynamics of taste.

And so, for somebody who is less familiar with tasting, it helps a lot in, uh, in your attention, in the tasting experience, with the sound, for instance. And something very amazing, actually, uh, we, uh, we did, uh, a listening experiment to people who know very well these wines. And without testing, they recognize the wine only by hearing the sound.

You see, so there can be some synaesthetic relations between taste and sound, for instance. The goal is to really create, uh, um,

Situations, uh, with the best possible conditions of testing. And then the, the sound is a support of this testing. Well, I didn't say that also the sound production system is, uh, is very well designed. There are many loudspeakers, et cetera, for the sound diffusion. And so, so that the experience is really the best, uh, as best as, uh, as good as possible.

So what about, you talked about the, the three stages, second stage production. Um, what, what kind of things do you use? I mean, do you use modern synthesizers or soft synths or things, or do you still work with your own sound generation things? What, what, what is, what makes the sound? Yes, actually, any, any use for sample action can be used.

Uh, after it's the choice of the artist to, to, to choose the right, uh, modules. But we also develop now some specific tools for that. Okay. Uh, for instance, we just issued a tool that is called ScatArt. It's the result of a European project, uh, and the goal was to sketch, uh, sounds for sound design with voice and gesture.

Uh, intuitively, if you do, shoo, poof, you can do it. For instance, sound designers can understand what you want. So it could be a kind of specification. And now that we have algorithms that analyze various dimensions of sound, uh, we can map these dimensions to other sounds that we have in the library with similarity metrics.

And so we developed that, uh, that tool that is called, uh, ScatArt, uh, in which you can enter, uh, A sound from your voice, or you can enter a gesture, etc. So a dynamic, uh, uh, signal. And it will map to, to, uh, to sounds that you will have selected. Do you use commercial synthesizers? Like Moogs, or Oberheims, or whatever?

Or soft synths? I mean, Ableton has a whole bunch of soft synths inside. Yeah, so, so, so, so So, so it's up to the artist, there is no, no limitation on that. Maybe there are limitations when the sound is produced, uh, by, uh, an embedded hardware, for instance in a car, it's very much constrained, and then the artist must, uh, uh, work with that, uh, sound synthesis, uh, device.

So I'm just gonna go to your website for a moment because actually it's very, very extensive and comprehensive. And so you have at the top, you have research and creation and transmission innovations, and it just seems like five categories. Can you talk a little bit more because I'm not entirely sure. So what, what exactly is the research into artificial intelligence?

So you're building artificial intelligence type softwares that can create music or research music or, um, or you are, what, what exactly are you doing in that direction? Why is it such a hot topic? Well, to, to take again, the, the, The structuration that I already described in terms of, uh, symbolic representation or signal.

First, uh, the generation of, uh, the symbolic information can be now, uh, given to, to an AI. And it goes much further than before. And, uh, we developed, uh, for instance, uh, now, uh, computer improvisers. So you, you basically play, uh, the, the computer doesn't know anything about your musical style, whatever. There is only a mic, and, uh, he, the, the, the agent will play with you.

It will analyze what we do, and it will react in, in real time to what you do. And so we, we have worked a lot on that with jazz, uh, musicians, for instance. So the computer improvises. Yes. It improvises with you. Uh, and we have developed that such that either it's completely agnostic, that is, there is no.

knowledge on what you will play, or it can have also a scenario, what we call a scenario, which is a high level temporal structure that can be, for instance, a grid in jazz, and then it will constrain its responses to that scenario, for instance, to a grid. But a scenario can be also in terms of succession of different categories of sounds.

It's not only related to harmony. It can be what you want in terms of dimensions of description of sound. So this is an example of use of AI. Uh, using symbolic representation of sound because, uh, here we, actually, we analyze what the performer plays in terms of notes, succession of notes, chords, et cetera.

It would feed an algorithm and then it will analyze it and react also in a symbolic world and then take the sounds, for instance, of the performer to then generate that new synthesis. Thank you, Hugues, for taking the time to talk with us in this Sound on Sound podcast. Thank you for listening, and be sure to check out the show notes page for this episode, where you'll find further information along with web links and details of all the other episodes.

And just before you go, let me point you to the soundonsound. com forward slash podcasts website page, where you can explore what's playing on our other channels.