Vous êtes sur la page 1sur 7

Interview with Iannis Xenakis

Author(s): Henning Lohner and Iannis Xenakis


Source: Computer Music Journal, Vol. 10, No. 4 (Winter, 1986), pp. 50-55
Published by: The MIT Press
Stable URL: http://www.jstor.org/stable/3680096
Accessed: 10/01/2009 10:20

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=mitpress.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Computer Music
Journal.

http://www.jstor.org
HenningLohner Interview with lannis
Aha Tonstudio
P.O. Box 10 01 27
5000 Cologne 1
Xenakis
FederalRepublic of Germany

Introduction programmingand wrote programsmyself, of course


I cannot immerse myself in the details of elec-
lannis Xenakis is one of the pioneers of computer tronics. If I had to spend my time doing that, I
music, having developed the ST (stochastic) music would never compose anymore! But, I know what
composition programin 1962. He has been involved direction to go, and if you have accurate and good
in numerous computer-basedprojects since then. collaborators,then it works.
His book Formalized Music (1971) summarizes his I have a team that works with me who are infor-
work of the 1960s. A more recent survey is avail- matics [computerscience-Ed.] people and elec-
able in (Xenakis 1985). tronic or software engineers. Of course they are
This interview, held in English, took place 17 awareof what's going on, but nevertheless it would
August 1985 in Delphi, Greece. be a very difficult task to produce new equipment
Lohner: You started the Centre d'Etudesde Mathe- in addition to our other projects because it's not our
matique et Automatique Musicales (CEMAMu)in goal here. This is why it is so important to collabo-
the 1960s as a cooperative. What have been some of rate with a manufacturerwho is obliged to go ahead
the important developments at CEMAMusince with new discoveries and make them cheaper and
then? of better quality.
Xenakis: I founded it twenty years ago when we Lohner: As a composer who likes to work in both
didn't have any money, with some friends who were instrumental and electronic fields, how do you see
mathematicians, statisticians, psychologists, or this electronic music work?
professors at the University in Paris. Gradually,we Xenakis: Well, the differencefrom writing for the
evolved and finally we received some grant money, orchestra is that in writing electronic music you
to start with from the Gulbenkian Foundation. also have to direct the invention of new tools. With
That was more than 16 years ago. In the early the UPIC at the level of the sound this means sound
1970s, we producedthe first converters in the world synthesis, and at the level of musical architectures:
with a resolution of 52,000 samples per second and how to manipulate architectural structures. Then
16 bits per sample. We had to produce sound by pro- you must determine which theories are most appro-
gramming computers and then storing the samples priate and simple to be used in music and sound
on digital tape. Then the digital tape was translated transformation.They must be usable by everybody
by our converter into sound. Laterwe received more in a standardway because that is very important.
grant money in orderto create the UPIC system, After all, if you can introduce standardways, then
with graphic input via a drawingboard.The UPIC that doesn't mean you cannot also be specific on
is linked to a minicomputer, or to microprocessors your own. A specific approachhas to be developed
in the present case. I hope that this development by the composers who are using the systems, but
will continue also with the help of manufacturers; based on something more general, more universal.
their task is to forge ahead in implementation tech- Lohner: Here, alongside your writing music, the
nology and to make it better, faster, and cheaper! CEMAMuproject comes in....
Lohner: To what extent must you as a composer be Xenakis: Yes, the UPIC is a tool that enables one
concerned with the technical details? to work in a systematic way on various levels at the
Xenakis: In spite of the fact that I did studies of same time-which is not the case with instrumen-
tal music-and especially not with most computer
music so far, when you have to filter your ideas
? Copyright 1986 by Henning Lohner. through some theories that might not be the best

50 Computer Music Journal


versus time, which is much easier-and you can see
it and hear the results immediately. But you needn't
analyze it on a theoretical basis. Instead of correct-
ing a sound by modifying the program,which is
much more complicated and time-consuming (com-
puter waiting-times, etc.), it's much more accessible
to just hear the lines and correct them in a much
easier way: by hand.
Lohner: Workingwith graphicshas always been
close to you, especially through the architectural
work you do. Did CEMAMudevelop out of the ne-
cessity of finding appropriatemeans of translation?
Xenakis: Yes, except that the UPIC did not come
from architecture, but it came from music itself.
Because when I wrote for orchestras, some of the
things were too complicated to be specified in stave
notation. So I had to introduce a graphic notation
which, by the way, is also more universal.
Everybodycan understanda line, whereas you
have to do specific studies in order to understand
what the symbols of traditional writing in music
mean. If you have training, it's OK, but then there
are also complicated things that cannot be dis-
cussed or studied with that kind of symbolic and
graphicnotation; as you know, traditional notation
is actually a combination of those two. Beyond that,
I think this more universal notation is possible for
ones, such as Fourieranalysis. Although it's a very everybody because it is the end of the hand that cre-
powerful tool in other domains, in music its use is ates the drawings.The hand is the organof the body
limited. So you have to invent different theories in that is closest to the brain.
order to produce, explore, and create new musical Lohner: But automatic functions for rhythmic pat-
worlds. terns, for arrangingmathematical curves, or for the
With the UPIC you have the potential to enter creation of an envelope transformationare things of
into the problem of composition in a much more mechanical nature that are not possible on UPIC-
simple and direct way-and by this I mean direct to and this has been criticized by some.
the mind. That is, you explore sounds on a much Xenakis: Yes, there are no machines or functions
larger scale, from the micro or almost micro ele- that can do this in UPIC; you need another device
ments of a sound up to more general architectures. that is completely different. But these systems can
The system is based on elementary acoustics, only do simple things; you can't compose some-
that is, acoustics based on a sound signal-a varia- thing that is complex and would last a long time.
tion of atmospheric pressure versus time.-You con- For example, keyboardsas a control device are not
struct these variations by drawing them. So you very sophisticated; continuity doesn't exist with
have them directly. Instead of superimposing or them. UPIC, considered as an instrument, extends
piling up sine waves, which is the traditional way to the limits of human perception, utilizing sys-
of synthesizing sound or music by computer, you tematic pitch, dynamics, page control, and so on.
delve into the other part of that equation. You spe- Decide if you really want to use a tool that is very
cify directly the waveform or the shape of the signal sophisticated, so that the music itself will be rich,

Lohner 51
or, whether you are happy with something that is
unsophisticated. As far as pedagogyis concerned,
these unsophisticated instruments would be harm-
ful: everyone would get used to something that is
simple, like pop music. And there's nothing more
unsophisticated than pop music. But to add special-
purpose machines like this to the UPIC: that's an-
other realm.
Lohner: And what about constantly varying a wave-
form? With acoustical instruments the waveforms
are subject to constant change, which is especially
noticeable in different registers. In UPIC these
changes don't exist, and many people insist that this
should be innate also to electronic instruments for
musical reasons. Is there a way of foreseeing (ran-
dom) variations of intensities, pitches, dynamics,
and other sound parameters?And will it be possible
on this machine? elasticity of the materials used to make the instru-
Xenakis: This is very important aesthetically. Two ments. Given the large waves, the rate of air flow,
things were brought up in this observation: on the and the dimension of the tube as air passes by the
one hand, different drawingsof waves sometimes little tongue of the pipe, things are simplified with
produce the same effect. This is well known. The increasing speed due to elastic inertia. On the other
ear doesn't always hear the differencesbetween hand, there's also an effect on the ear. In listening,
graphisms (the contour of the waveforms);it only the inner ear simplifies the complexity of the per-
distinguishes different classes of waveforms. It ceived effects. Forthis phenomenon we don't know
would be interesting to examine this. As for the very well how the ear functions. We don't have
variation of a waveform or a dynamic envelope, enough technology to test these types of things to
you'd need to have special transformationfunc- see if there is a continuity. Does the ear respond
tions, maybe algebraic or other types with special equally to the high notes as it does to the low notes?
keys on the drawing tablet itself. This could happen And does it converge if there's a variation in timbre?
by varying the same function, or it could vary sto- Anyway, since the ear is not linear, there is cer-
chastically. Forexample, presently we can do this tainly variation in the perceived spectrum or in the
by following the same curve that you draw to de- perceived waveform. Both are equivalent inasmuch
scribe different envelopes. So graphicallywe have as the waveform is a function of time and the spec-
certain ways of achieving variation. trum is the series of Fouriercomponents.
On the other hand, if you want to simulate musi- If you wanted to simulate traditional instruments
cal instruments, you have to see what the wave- you would have to obtain instruments, study them,
forms correspondto. Forinstance, if you take the record them, class them by sufficiently differenti-
low G tone on an organ, the waveform has a certain ated registers, and then use them as if they were
complexity. As you go towards the higher pitches, specific waveforms in various frequencies and in-
the complexity diminishes until it becomes almost tensities. UPIC certainly wasn't made to simulate
a sine wave, and this you hear. So when you start existing instruments.
with a given timbre of an organ pipe, the more you Lohner: You have written two pieces for the UPIC
gravitate toward the higher notes, it converges to- machine so far?
ward the sound of a little flute. Xenakis: Yes, I do too much administration for the
On the one hand, this is valid for all of the instru- UPIC system, so I don't have that much time to
mental sounds and is due also-probably-to the work on it. I hope to compose more with UPIC in

52 Computer Music Journal


the near future. One of my UPIC pieces is music in ard. Fororchestral music, this might be enough,
itself. It was the first piece to be written on UPIC, but I think if we want to go fartherin acoustical
made directly with it. experimentation and also improve digital sound
Lohner: Mycenae Alpha. quality, then we have to go on to a higher sampling
Xenakis: Mycenae Alpha, yes. And the other one rate for the kind of music that is much more com-
is called Pour la Paix (ForPeace), based on texts by plex than instrumental music and less recognizable.
my wife, who is a writer, with UPIC sounds and Other extensions to the UPIC could be simultane-
also a chorus. ous with this development.
Lohner: How do you feel writing for a machine like Lohner: Computers in music are very supportive of
this, knowing that it has technical limitations to this line of abstract thought, that is, expanding dif-
which solutions are foreseeable in the future? At ferent dimensions at the same time.
the time of Mycenae Alpha, the mixing function Xenakis: Exactly.This means that you have to con-
was not yet possible. Did you conceive the piece trol the tiniest events as much as possible, or-
rather as an etude, an exercise or practice piece that when you talk about informatics which deals in
you might change later on? numbers-the range and the level of the sample.
Xenakis: Sure, there are things tht you can do as an This is quite a new aspect opened up by computer
exercise, but if they are good enough, then they are science.
not exercises anymore. They are results. Mycenae Lohner: Continuity is a word that has often been
Alpha I did as a technical exercise, but also as a used during this conversation. Our views on musi-
musical exercise, and I knew the limitations of the cal objects have changed in this respect and we are
machine. I could not do any mixing at all, so I did now aware of the fact that we are in a continuous
pages of the score in a way that they could be linked medium.
to each other just by timewise juxtaposition. Xenakis: Yes, we are in a kind of continuum from,
Lohner: Have you planned any compositions now? say, usual objects that we use in music down to
Are you going to wait for the next generation? the aspects of music that are inaudible, but which
Xenakis: I am waiting for the next generation of produce these events on a higher level. Computers
UPIC synthesis, which will be in real time. The have brought thought down to that "lower"level.
sample calculation will be done immediately. That Before it was not possible.
is, as you draw on the page, you would press on the Lohner: Is this awareness reflected in your instru-
"listen" button and immediately you have the result mental music as well as in your projects with a
of the page or any other function or part of the ma- computer?
chine that is used for this effect. Xenakis: Yes, because the things that you use, the
Lohner: What sampling frequency will that have tools or the ways of thinking which are ascertain-
then? able can be applied or used on a much lower level
Xenakis: It will go to about 45,000 or 50,000 sam- or on a much higher level.
ples per second with 16 bits per sample-for the Lohner: How do you decide where the lowest or
time being. I hope that it will be extended further the highest level is? The other day we were talking
within the next year or so to a higher sampling rate about the possibility of having universes beyond the
and maybe also a better sample resolution than 16 one we know of. When we explore microstructures
bits. of matter, and dissect the atom, soon we get to the
Lohner: For the moment, the present sampling rate limits of perception. We encounter a boundarybe-
implies alias distortion which is audible because of yond which lies an unknown micro universe.
the extremely large frequency and amplitude range Xenakis: That's right, but I can't help that. I live in
of the instrument. this epoch. These are the tools and possibilities to
Xenakis: Correct. In the manufacture of the com- investigate lower ranges and higher ranges and that's
pact discs, for instance, they use up to 44,000 sam- it! I am bound by the technology or the techniques
ples per second, with 16-bit resolution as a stand- or the thoughts of my time.

Lohner 53
Lohner: Nevertheless you are looking for ways to also depends on organizations. Now it's too late
supersede that condition. anyway because you have to do those things in a
Xenakis: Yes, because this is the way one lives. continuous way during a whole lifetime.
That is, not only to do what one does, but also to As I see it, music is a domain where the most
try and understandwhy and where and what the profoundquestions of philosophy, thought, behav-
targets or the aims are, if there are any. ior, and the theory of the universe ought to pose
Lohner: For a composer who has consequently themselves to the composer. The role of the musi-
gone his own way, what is your motivation to write cian must be this fundamental research: to find
music? answers to phenomena we don't understand,and to
Xenakis: Well, I write music because it's the way enlarge our powers of conception and action. So it is
for me to feel less miserable. I would have liked to a perpetual exploration.
do other things at the same time, like scientific re- Lohner: "Perpetual"is the word I'd like to pick
search or architectural research and so on, but it's up; isn't that quite in line with Parmenides,whom
very difficult to put them together in 24 hours. It you've often cited?

54 Computer Music Journal


Xenakis: No. Let's just say that Parmenidesis in completely different type of music in correspond-
the pure philosophical ontological domain, whereas ence between image and sound.
when you write music you have to make decisions But this is a future stage which could happen ac-
and compromises. There is no doubt about that. cording to the amount of subsidies grantedby the
Otherwise, you don't do anything. If you admit that government.
the main problem is creativity or originality, that is, Lohner: One would have pens of certain colors, for
doing something new as much as possible-differ- instance, and composers would be able to write
ent from what you have received from your envir- pieces at home, send them in, and they'll be tran-
onment, if you are not a philosopher, but just a scribed on an UPIC system.
composer, then you have to decide. This in spite of Xenakis: Yes, yes. It would open up completely
the fact that it might not be absolutely original or new paths in thinking the music all together. This
new. So it's a perpetual compromise, which is very is predictablein a way, precisely because it has
painful most of the time. never been done before. But nevertheless, it is a
Lohner: Speaking of decisions, what majordevelop- new task based on the most advancedcontempo-
ments have you planned for this next and following rary technology that uses image and picture cells.
generations of the UPIC? Lohner: So the idea behind this is to make it more
Xenakis: Forinstance, we plan to introduce trans- accessible to all people?
formation functions for the graphicinput with prob- Xenakis: Exactly.
ability calculations and other algorithms. These Lohner: And another general goal for UPIC is the
should be standardand usable by anybody.And in pedagogical side.
the fourth phase of development, we will introduce Xenakis: Oh, yes!
image processing. With colors and lines you have a Lohner: What are the main points of interest for
camera that stores the color image in the computer. you and CEMAMuto have the machine developed?
Analysis by the computer will be a very interesting Xenakis: Forone thing, we are not trying to make
decoding of this color painting. money from it-that's not the point at all. I want to
In Stockholm 1984 I said that an image processor have a tool for myself and for other people that will
for a future stage of UPIC is where you could draw be general enough to be used in pedagogy,so as to
or paint a page on UPIC and explore it with a cam- bridge the gap that exists between "normal"people
era. Through a coding system you could transform and contemporarymusic developments. If anybody
this exploration into sound, into music. It's very, is able to use such a machine, it will heighten the
very difficult to do. For example, take the Mona awareness of the averageperson who will then be
Lisa-almost everyone knows that if you explored involved in composition also. This makes a much
it with a camera, the camera would need to read more homogenous environment for music.
each color type and intensity of color and see what Lohner: Everyperson being essentially creative.
it correspondsto. Each point-each digital point- Xenakis: I think so, yes. This is a conclusion I
would have more or less 250 different values: 8 bits. came up with after many years of personal experi-
This is an enormous quantity of information in one ences in how people react to music, and what they
little point of colored surface. In transformingthis, feel. It also reflects how I feel about them.
you could put in about 250 values; each one could
be structures, pitches, intensities, but also more
complicated structures. Each point therefore would References
open up a fantastic number of possibilities, differ-
ent ways of coding, depending on how you travel
Xenakis, I. 1971. Formalized Music. Bloomington:
inside the picture. You could explore a type of coding IndianaUniversityPress.
-by lines-and receive music in time. Speed, Xenakis, I. 1985. "Music Composition Treks."In C.
by the way, is also translated in a simple way, in an Roads, ed. 1985. Composersand the Computer.Los
automatic way in music. Each line could give you a Altos: William Kaufmann,pp. 171-192.

Lohner 55

Vous aimerez peut-être aussi