Ronald Bruce Smith Center for New Music and Technology An Interview with University of California, Berkeley 1750 Arch Street Tristan Murail Berkeley, California 94720-1210, USA [email protected]

Tristan Murail was born in , in of quasi-aleatoric experimentation. This was the 1947. Following university studies in economics, antidote to the structuralist trend in music. I was Arabic, and political science, he entered the com- not at ease with either of these approaches, nor position class of at the were many of my friends. This was why we had to Conservatoire National Supérieur de Musique de find something different. in 1967. Upon graduation in 1971, he was Smith: I believe that the term used in America was awarded the Prix de Rôme. On his return to Paris, “anti-establishment.” he founded the ensemble L’Itinéraire with com- Murail: In a way, serialism was the establishment. posers and former Conservatoire classmates It certainly became the establishment when Pierre and Michaël Lévinas. L’Itinéraire Boulez came back to Paris in the mid-1970s to es- soon became known as the starting point for an tablish IRCAM and the Ensemble aesthetic movement known as spectral composi- Intercontemporain. He was “establishment,” as he tion, its two main proponents being Tristan was heavily supported by the government. It Murail and Gérard Grisey. In a nutshell, much of seemed that there was little room for anything the material in a spectral composition is derived else. While there were other ensembles in Paris, from the frequencies of spectra and their behavior. such as L’Itinéraire, we had to fight just to survive, Tristan Murail has been involved with IRCAM to keep our meager subsidies. (Well, meager com- since 1980 as a composer, researcher, and profes- pared to what Boulez was receiving.) sor. In 1997 he moved to New York City, where he Smith: At the time of the inception of L’Itinéraire, is a professor of composition at Columbia Univer- what was its mission? sity. The following interview was conducted by Murail: We were trying to propose another ap- telephone on 7 February 1999. proach toward music. We tried to play everything that would not be considered establishment. At Smith: Can you briefly describe the contemporary times we played some music by Boulez, though we music scene in Paris in the early 1970s? were mainly looking for novelty, for things that Murail: At that time, I felt that serial composers would bring a new approach to music. and theorists were strongly predominate. With We programmed many premieres, such as the respect to theory and actual practice, many people French premieres of music by Giacinto Scelsi, who were still influenced by serial ideas and, of course, was virtually unknown at the time. We also gave by the theories of Pierre Boulez. However, 1968 the French premieres of pieces by George Crumb, being the year of the student uprising in Paris pro- because along with a certain freedom in his lan- vided a source for a contrasting viewpoint. In art, it guage, we believed that he was bringing something gave way to many experiments. This new aesthetic new to the field of timbre. We also played the mu- emphasized the art of destruction, although this sic of the Italian composer Salvatore Sciarrino, who model was already quite old, dating back to the was working on extended techniques. His music is Dada movement of the 1920s. John Cage was im- very, very special; it is not at all structuralist mu- portant at that time, as were Earle Brown and sic! These sorts of things we did quite often. Of André Boucourechliev, a French-Romanian com- course, we played the music of our smaller group. poser. Many concerts at that time involved a sort We played many pieces by Gérard Grisey, for ex- ample. We were also very interested in live elec- Computer Music Journal, 24:1, pp. 11–19, Spring 2000 tronics, so we built our own equipment. © 2000 Massachusetts Institute of Technology.

Smith 11

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 Smith: What sort of electronics did you use at that formation that we required was not as readily time? available as it is today. Gérard Grisey and I had Murail: Analog electronics: ring modulators, har- read books on acoustics that were designed more monizers, electric organs, electric guitars, reverb for engineers than for musicians. There we found units, and the Ondes Martenot. It was centered rare information on spectra, sonograms, and such around what was available at the time. A bit later that was very difficult to exploit. We also did our we bought synthesizers and even later we bought own experiments. For example, we knew how to computers. I don’t know whether people today real- calculate the output of ring modulators and, a little ize how limited the electronics were in the 1970s. later, frequency modulation. Those things were, Smith: And difficult to program, too, I would imagine. theoretically, quite easy to manipulate. Murail: Oh, yes! We had an old EMS synthesizer, When I began working at IRCAM, things be- which was a very temperamental machine! There came easier, even though it was still not very easy. is a Swiss composer, Thomas Kessler, who has In the early 1980s it was still quite complicated for done a lot with that machine. L’Itinéraire often us to perform a Fourier analysis on a computer. played his music, and it was always a nightmare. Scientists or computer technicians did this. Even- He was so meticulous. He would put tape all over tually, the data became more widely available. the synthesizer to immobilize all of the param- This was how I came across the Fourier analyses eters, which, as you said, were very temperamen- of several instruments that David Wessel created tal. We did quite interesting things with him; it at IRCAM. These were the basis for my first elec- was a tour de force every time! tronic piece, Désintégrations [1982–1983; for 17 We also used things like ring modulators and instruments and 4-channel tape; see Figure 1]. tape loops which were also temperamental. Actu- Smith: I would think that if one were to use mate- ally, just moving from one piece to another during rial derived from frequency modulation or ring a concert was complicated. modulation, it must have been an incredibly time- I did many experiments with the equipment consuming task to make all of the calculations. from that time. Even though I did not really use Murail: Yes. Time consuming, and boring, too. I the equipment in my pieces, I learned much about could show you pages and pages of calculations. electronics. I derived formal models from them Smith: How did you first become involved with and just learned about sound. I think that such IRCAM? electronic experiences were important for many of Murail: That was in 1980. They had organized a us at that time. summer session for people belonging to or related Smith: You have become known as one of the to L’Itinéraire. It was a way of signing an armistice! founding composers and one of the leading com- Smith: So things were that tense? posers of something which has been termed “spec- Murail: There was quite a lot of tension for the tral music.” What does that term mean to you? first few years of IRCAM. However, since we really Murail: I think that it is chiefly an attitude toward needed computers for the sorts of things that we musical and sonic phenomena, although it also were doing, it was quite natural (both aesthetically entails a few techniques, of course. We were trying and technically, if not politically) for us to be at to find a way out of the structuralist contradiction. IRCAM. At the same time, we did not want to be com- Smith: I would think that what you were inter- pletely intuitive like the aleatoric composers or ested in would have coincided very well with the like Giacinto Scelsi, or even György Ligeti (who research that was being done at IRCAM and with was and remains an experimentalist but an intui- the research that continued there through the tive composer). We were interested in those experi- 1980s, particularly toward synthesis and sound ences, but we wanted to build something more analysis. sound (pun intended). This was part of it, as was a Murail: Yes. I think, in fact, that there had been a curiosity about sounds. Also, at that time, the in- historic conjunction between an aesthetic move-

12 Computer Music Journal

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 Figure 1. Excerpt from the score of Désintégrations.

ment, the spectral movement, and the techniques, problems because, in terms of calculations, addi- research, and software developed at IRCAM. I think tive synthesis is very costly. we needed them and, in a way, they needed us. Computers at that time were big machines. At Smith: How were the electronics for IRCAM, there was one huge computer that every- Désintégrations realized? Did computer analysis of body shared (though it was probably less powerful sound play a role in the construction of the compo- than a Macintosh G3). When somebody was calcu- sition and in the creation of the electronic sounds? lating a sound, the whole system slowed down. Murail: The idea behind the piece was to integrate You also had to listen to what everybody else was the electronic sounds and the instrumental sounds working on. Each sound would circulate through as much as possible. I had worked on instrumental IRCAM. There was a line to listen to the sounds. analysis before this composition, although Grisey You had to wait your turn. had done more than I had. It was then quite natural Smith: It must have been a frustrating way to that I would use that as my starting point. I did not work, to wait for your sound, listen to it, and then do any new analysis for this piece. In fact, I simply after a short pause, hear somebody else’s sound. used the stock analysis that had been done by Murail: Oh, yes. At that time Xavier Rodet was David Wessel. They were just text files, and I used working on the program Chant. They were testing them as the basis for the synthesis as well as for the program with Mozart’s Queen of the Night aria the harmonic structures of the piece. In order to from The Magic Flute. I can’t imagine how many keep the richness of the information, I decided to times we heard that [sings a portion] coming from use additive synthesis, which is, of course, the in- the loudspeakers. verse of a Fourier transform. However, this gave us Anyway, my practical problem was that the ad-

Smith 13

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 ditive synthesis took too long. I think that it was with software like MSP, but it is not that easy. David Wessel’s idea to use the 4X, which was a Even trying to emulate a [Yamaha] DX7 on a com- very new machine at that time. There was no syn- puter is difficult. The DX7 is in fact a very special- thesis software available for it, so Andrew Gerzso ized, powerful computer. Perhaps a combination of wrote a program especially for this project. The 4X MSP with sampling is the only solution. Still, per- had been used to process sounds rather than for sonal computers today, while they might be ca- synthesis, so I think that this was one of the first pable of emulating a DX7, are not yet powerful applications for sound synthesis. enough to emulate a TX816. Also, I wrote a small program to create param- In January 1999, I transferred the programs that I eter files. Since some of the sounds required thou- used to make L’esprit des dunes [1993–1994; for sands of parameters, it was impossible to do it all eleven instruments and computer-generated manually. Moreover, as I could not produce more sounds; see Figure 2] at IRCAM to the Macintosh. than 30 oscillators at a time, I had to produce each It was originally done on the ISPW [IRCAM Sig- sound by layers and then mix everything on a 16- nal-Processing Workstation] with three cards (the track analog mixer. We had a number of assistants maximum configuration), and it can now run on to trigger the sounds at the right moments. A sort MSP. This shows how fast technology moves. It of real-time, real-life sequencer! In fact, we started means that the Macintosh G3 is more powerful by building the click track, and then everything than a NeXT with three specialized cards. was coordinated to the click track. Smith: Among many things, your music is remark- Smith: You have written for live performers and able for your ability to create unified environments tape, and live performers and live electronics. What between the electronics and live instruments. are the advantages of both, and do you have a pref- Could you talk about strategies for the integration erence for either? of electronics with live instruments? Also, could Murail: My composition Allegories [1989; for flute, you share any observations and experiences about clarinet, horn, percussion, violin, violoncello, and loudspeaker placement? live electronics] uses Max and a Yamaha TX816, Murail: In a piece like L’esprit des dunes, my pur- though I’ll have to find a replacement for the pose was not to have a perfect blend of electronics TX816 someday. The problem with tape, of course, and instruments. In a way, it is more difficult to is synchronization. In a piece like Désintégrations, have a convincing relationship between electronic there is a click track, but it gives the conductor no and acoustic sounds when they are not extremely liberty at all. He has to follow the click track. With related, because one can really feel the nature of Allegories, I wanted the exact opposite, for the the electronic sounds. I used a similar strategy for electronics to follow the conductor. The fact that L’esprit des dunes as I did for Desintégrations, al- you can adjust the time of the electronics is the though the techniques have changed very much. main advantage of live electronics. We have much more sophisticated techniques to- One might imagine other possibilities as well. day. I used partial tracking, and I could resynthe- For example, you could think of changing the tim- size the tracked partials with the ISPW and now bre of the piece according to the room acoustics. I with MSP. Of course, before resynthesizing one can had thought of this with Allegories, but decided it alter the parameters directly with Max or alter the was too dangerous. That would take too much time data with Patchwork. In this case, I introduced in the rehearsal and in the concert hall itself, so I some compositional elements in the resynthesis of didn’t think that it would be realistic. Nonetheless, the sound, although I kept some of the internal life it would be an interesting use of live electronics. of the sound. Acoustic sounds have a natural life in Smith: How are you dealing with the fact that which the frequencies can be understood as behav- some of your works employ equipment, such as the ing in ways that could be described as chaotic. It is Yamaha TX816, that is no longer easily available? that property, in fact, that is very hard to simulate Murail: The only real solution is to emulate them with synthesis. Even if one uses models and then

14 Computer Music Journal

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 Figure 2. Excerpt from the score of L’esprit des dunes.

alters the sounds drastically, by keeping the cha- be modified with automated movable panels. We otic behavior, the electronics and the instruments got much more interesting results by experiment- can communicate more easily (since the electronic ing with both the room acoustics and the loud- sounds act a little bit like acoustic sounds). I tried speakers there a year later. to accomplish this in L’esprit des dunes. Another question is, Should you amplify the in- You mentioned the placement of loudspeakers. struments or not? My experience has been that it The problem also involves the room acoustics, for is difficult to amplify instruments for a large en- which there is no a priori solution. You have to semble like Désintégrations, but with a smaller work with the room. The first performance of ensemble such as L’esprit des dunes, it is often L’esprit des dunes at IRCAM’s Espace de Projec- better to slightly amplify or add reverb to the in- tion was very disappointing. It took place in the struments to create a similar acoustic environ- middle of a long day of nonstop concerts, the ment between the electronics and the IRCAM Open House. At this event, nothing could instruments. Once again it really depends on the be changed to accommodate the room acoustics, room’s acoustics. My best experience with not even the placement of the loudspeakers. How- Désintégrations was in New York at the BAM ever, the acoustics of the Espace de Projection can Theater. It’s a small theater. The acoustics were

Smith 15

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 such that there was a perfect blend between the ine several programs that were being developed tape and the instruments, but also one could there, and among them was Patchwork. Of all of clearly hear the spatialization of the tape, which is them, I thought that Patchwork was the best. usually masked by the orchestra. Most of the work that was done on it at that time Smith: Are you interested in psychoacoustic re- was toward the interface. The idea was that a com- search and, if so, does this inform your composi- poser who had only a basic knowledge of comput- tions in any way? ers should be able to use the program. One just Murail: It interests me in a general way, but not in needed to be logical to get results. It is perhaps the a very specific way. Most of the research, while most widely used of the IRCAM software. important, has not directly applied to my concerns. Smith: When one compares your music from the At IRCAM, we did an experiment with Stephen 1980s to your music from the 1990s, one notices McAdams that dealt with the perception of com- some very distinct differences. One difference is that plex spectra (Pressnitzer, McAdams, Winsberg, and the formal structures in your more recent music are Fineberg 1996). I think that there are many things more complex and contain more self-reference to discover, but I’m limited to my own subjectivity through the recurrence of musical events or musical or my own intuition, or let’s say, that I am my own objects. The early works seemed to work through subject. However, I am not the best subject, as I linear processes, more or less. Furthermore, your would like to be a naive listener and I am not. recent works show an increase in harmonic activity Smith: Were you using any sort of computer-as- compared to the works of the 1980s. Are these differ- sisted composition software before IRCAM’s ences at all related to the use of Patchwork? Patchwork software first become available? Murail: The tendency toward more complex forms Murail: If we go back to the period of Désintégra- and more activity was an aesthetic necessity. It tions, I began to write a few programs for the com- was made possible, in one way, through the use of puters that were available at that time. I wrote my computers. It is clear that when you have to calcu- own programs on the machines at IRCAM to calcu- late harmonies and durations by hand, it takes an late frequency modulation and ring modulation. Af- enormous amount of time. Once one has generated ter that I used a small home computer which I also a certain amount of material, after a good amount programmed. I used that for many years, until the of work, one is simply happy to work with it. 1990s. I had quite a sophisticated program on my The linearity of the music of that period was home computer that worked very well. I had a MIDI partially an aesthetic decision. It was a way to cre- output, and I could bring up music staves on the ate a rupture, a strong break, with structuralist screen. I used a light pen instead of a mouse, which music. It was a conscious decision to write with enabled me to draw notation onto the screen. In linear processes. Still, we soon found the limits of terms of interactivity, it was superior to Patchwork, this way of composing. I say “we,” because at that but in general, it was much more limited than Patch- time, I shared many discussions about these mat- work. This is what I used for most of the 1980s. ters with Gérard Grisey. Either we were going to Then a few people at IRCAM began to work on a repeat what we had already done, or we were going program called Esquisse (Laurson and Duthen to need to find other solutions, especially toward 1990). This was not an official project at IRCAM. form. I think that I was aware of that fact fairly It was initiated by Jean-Baptiste Barriére and com- early. There are formal aspects of Gondwana posers and Marc-André Dalbavie. [1980; for large orchestra] that are nonlinear or Then, Patchwork was proposed by Mikael Laurson that break linear processes. The increasing ability around 1990. I remember that I wrote Allegories in to generate materials with a computer has facili- my old system in 1989. I switched to Patchwork tated such solutions. In a piece like Allegories, over the next two or three years. there are a good many spectral situations that I had worked as a consultant for Patchwork. would have been too time consuming to create Around that time, I was asked by IRCAM to exam- without the computer. This technology has con-

16 Computer Music Journal

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 tinued to develop since Allegories. In the early AudioSculpt and by eliminating weaker compo- pieces, the processes and the frequencies employed nents with Patchwork. However, I believe that in composition were identical to the form, Patchwork’s problem of manipulating larger struc- whereas the recent pieces employ a second layer tures will be mostly solved with OpenMusic. that uses processes and calculations that are not There is a dimension missing in Patchwork that is identical to the processes and calculations being the time dimension. In OpenMusic, there are fea- used on the primary layer. This allows for a form tures like the maquette function. There is a level that is more versatile with more references to al- of organization that is higher than in Patchwork. ready-heard objects and their transformation. Patches can be embedded in a higher order. One Smith: Do you use Patchwork to work out other can have a time dimension and maintain the aspects of a composition such as rhythm, duration, interactivity between the patches. dynamics, and musical gesture? Smith: Something that you share with the music of Murail: I use it mainly for devising forms and situ- Olivier Messiaen is that you have drawn on sounds ations and for planning durations, but I am very from nature as a source for compositional material. prudent with these. Psychoacoustic research dem- How does your use of sounds from nature differ onstrates that perception of musical time is not from how Messiaen used sounds from nature? Fur- identical to a precalculated formal structure. I use ther, how does your use of sounds from nature dif- Patchwork for these matters, but I am also very fer from how a composer like Jean-Claude Risset distrustful and cautious of it. Therefore, I often use uses sounds from nature in a composition like Sud? it as a guideline or a projection of the form, and Murail: There are several ways of using sounds then allow things to be changed by the music it- from nature. There is a metaphorical way and a self, the musical objects. Still, I use Patchwork to more direct way. In Messiaen’s music, one hears create complex musical objects and to transform both: the metaphor of nature occurs in many them and so on. Here, I’ve found limitations, be- pieces, as does the direct use through his use of cause Patchwork was not designed for such things. birdsong. However, even when he uses birdsong, it For my last orchestral piece, Le Partage des Eaux is quite transformed. [1996], I used very complex musical objects derived Risset has done a very interesting job in Sud by from the sound of waves in the ocean. Those struc- using the sounds as a source for filtering. It is a tures are very complex, as you can imagine. I tried little like what I said about L’esprit des dunes. to use them as such and then transform them in You keep the natural life of the sound, but you Patchwork. It was not easy. I had a similar experi- also sculpt the sound into composer-specified har- ence with L’esprit des dunes when I tried to pro- monies. This is a bit like what I do. It is a bit cess the analysis of real sounds. I found limits to metaphorical, but it is really the sounds them- the memory, the computational power, etc. selves that are used. I am not working with the Smith: Would that involve interfacing AudioSculpt sounds themselves, but with images of the sounds, with Patchwork? the structures of the sounds. In Patchwork, I ulti- Murail: Yes, but lately I’ve been using Additive, mately keep the energy, the amplitude shape of which is best for partial tracking. It is available on the sounds or their actual spectral content, and the latest IRCAM CD-ROM. It was part of analysis artifacts. The frequencies that I end up Diphone, but it is now a separate application con- with are partially artificial, but the grouping of the tained within the Diphone file. For L’esprit des formants is not. I keep the formants and I keep the dunes, I used Additive—it was only working on the movement of the sound. In the final orchestration SGI—then I used Patchwork, and last the ISPW. of Le Partage des Eaux, I kept the contours of the Now I can do the same thing on my Macintosh wave rather than the sound itself. I think that this using Additive, Patchwork, and Max MSP. can be quite interesting. For Le Partage des Eaux, I had to make a heavy In another piece, Bois Flotté [1996; for violin, data reduction by using the masking function in viola, violoncello, piano, trombone, and computer-

Smith 17

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 Figure 3. Sonogram show- ing a melodic fragment of a Mongolian voice that was used as one of the source materials for L’esprit des dunes.

generated sounds], I also use the sounds of the sea, the first time that I was able to do everything in but mostly to derive strange harmonies. It works my home studio. very well, and I don’t think that one could recog- Smith: In L’esprit des dunes, you use voices and nize the origins of the sound in this case. A bois instruments from nonwestern cultures. Can you flotté is a piece of driftwood that becomes a work speak about the sound sources of the composition, of art after a sculptor finds it and makes a few and how they are treated in the composition? changes to it. So, the idea is that one finds an ob- Murail: The main sources are Mongolian diaphonic ject in nature, an interesting object, and transforms singing, Tibetan trumps, and the chants of Tibetan it to make something artistic with it. The fact that monks. The one thing that they have in common is you do the selection yourself is in itself an artistic a very rich, strange spectra. The Tibetan trumps, attitude. I try to do this with sound, which ex- for example, have a spectrum that is distorted in a plains the title of the work. For this piece, it was manner opposite to the spectrum of a piano. Spe-

18 Computer Music Journal

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021 cifically, it is compressed. There is distortion of L’esprit des dunes, has the music of other about –3 percent. I countered this distortion to nonwestern cultures been an interest to you or make it more harmonic. The voices are also been an influence on your music? slightly distorted. The diaphonic singing is a kind Murail: It has always been an interest. An influ- of spectral singing where the overtones are used to ence? Perhaps. There is always the formal aspect make melody. The melody is produced inside the and the sense of time that is implied in the music spectrum through a sort of filtering. This is exactly of India or Indonesia. However, I think that it is what I did at times in the 1980s. I made melodies difficult to translate that into our own culture. from the spectral content, which is what these Smith: What are your current projects? singers have been doing for centuries! I thought Murail: I have two pieces for instruments and elec- that this would be an obvious reference. All the tronics to write this year. I am now working on a sources were analyzed and resynthesized using 40 piece for large ensemble with a few electronic partials, which is usually enough. We made cue sounds. The electronics will not be as involved as lists for the ISPW to make the resynthesis. The in L’esprit des dunes. It will involve only a synthe- idea was not to recreate the original sounds, but to sizer using synthetic sounds that I will produce transform them to create electronic sounds. The with my home system. The second piece is for trumps were used to create special musical objects three instruments and electronics. I intend to use that were percussive but with resonances. The MSP for this piece and, I hope, to use it in real Mongolian voices were used as a resource of me- time. Technically, I will be working more on par- lodic contours. I would filter all of the harmonics, tial tracking, MSP, and OpenMusic. I also hope to and the various harmonics that were accentuated transfer my patches and libraries from Patchwork over time formed the basis of the melodies. It was a to OpenMusic. very long process, but one that worked well. I re- tained all of the microinflections of the sounds. Furthermore, I tried to “tune” the sounds to the References harmonies of my music instead of using the sounds to create harmony. In certain parts of the Assayag, G. 1993. “CAO: Vers la partition potentielle.” piece, I superimposed the trumps and the two Les Cahiers de l’IRCAM. Paris: IRCAM. styles of singing. I altered the sounds by shifting Laurson, M., and J. Duthen. 1990. “A compositional en- vironment based on Preform II, Patchwork, and the partials so that their spectra matched. This is Esquisse.” Proceedings of the International Com- what I mean by tuning the sounds. It was fascinat- puter Music Conference. San Francisco: ICMA. ing, because I could go beyond processing and Pressnitzer, D., S. McAdams, S. Winsberg, and J. make interpolations in the sound in ways that Fineberg. 1996. “Tension et rugosité dans une were compositional. musique non tonale.” Journées d’Informatique Musi- Smith: Besides the music of the cultures used in cale 1996. Caen, France.

Smith 19

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/014892600559146 by guest on 25 September 2021