Jean-Baptiste Barrière - Chréode I (1983)

Jean-Baptiste Barrière was born in Paris, France in 1958. He studied music, history of art, philosophy and mathematical logic. Parallel to composing, he made a career at Ircam/Centre Georges Pompidou in Paris, beginning in January 1981 as researcher with the projects Chant (synthesis of vocal singing with computer), and Formes (control of synthesis and composition with computer). From 1984 to 1987, he directed Ircam’s Department of Musical Research, from October 1989 that of Education, and from 1993 to 1997 he headed the production department. 1997-98, he taught computer music composition at the Sibelius Academy in Helsinki. In September 1998, he left Ircam to concentrate on composing. 1

For tape produced at Ircam
World Premiere: Festival International des musiques expérimentales de Bourges, 1 June 1983
Length: 10 minutes
Publisher: Unpublished
Discography: Wergo 2024-50
First Prize for digital music, International Electroacoustic Music Contest, Bourges, 1983.

Programmnote:

Chréode  is a term borrowed to Morphogenesis and Biology (from the greek 'cre' we need, and 'odos' path : necessary path). it serves here of metaphor for a cross systematic investigation of sonic materials and organizations.
Though sonic materials have been worked carefully, attention in this piece is more specially on organisation. Chréode is the first step towards a grammar of processes I want to try to elaborate.
This research on musical processes, their fields of action and their limits, has been thought of a strategy of approach of the musical territory, as it is renewed by the possibilities made accessible through the use of computer.
A very general destination of this project consisted to experiment with different types of organizations, and at a higher level to structure them in time and formally.
Compositional structures as well as scores were elaborated in the FORMES programming environment with the help of Pierre Cointe and Yves Potard. Sonic materials were chosen to allow a certain type of control over timbre, concentrated of a small number of compositionally relevant parameters. Therefore the work on timbre is essentially based on the compression and expansion of formants or spectral enveloppes, by departing from vocal types material and derivating from this models towards others, with or without references to musical families of instrumental timbres.
Sound synthesis was made with the CHANT program at IRCAM.
The piece is dedicated to the CHANT/FORMES project and to Kaija Saariaho. 2

In Chréode I, the composer exploits the resources of the program Chant to explore the limits of simulated voice.
Chréode is a term borrowed from morphogenesis and biology (from the Greek cre, there must be, and odos, path, 'necessary path'). Here it's used as a metaphor of the systematic investigation of material and organisation.
Both the compositional structures and the scores were elaborated in the programming environment FORMES, with the assistance of Pierre Cointe and Yves Potard.
The sound materials were selected with an eye to certain kinds of timbral control, focussed on a small number of compositional parameters. So, the timbral work is essentially based on compressing and expanding the formants or spectral envelopes, starting from vocal-type materials, and thereafter displacing these models toward other paradigms, with or without reference to the families of instrumental timbres. Synthesis was effected using the program Chant.
In this example, a decelerando ends the first part, describing a predetermined curve.
chreode424.mov


Illustration of frequency variation.


Illustration of amplitude variation

This second excerpt from Chréode I, after a movement of synthetic speech very rich on the level of timbral articulation, allows us to hear clearly a process of interpolation from the voice to the sound of a bowed string; what emerges is a melodic curve preceding the coda:
chreode528820.mov 3


The choice of vocal material as a reference point in this piece produces an abstract presence which is never grasped, always assumed, endlessly sought and ever non-existent. With the "disembodied" vocal synthesis, there is a "voice," but no singer. There is only a corpus of material which is terribly abstract, but which retains a physical link with every listener. The reference to vocal material has real mimetic value: on one hand it provides schemes for organization of material which we may draw upon, transform and anamorphize; on the other hand, it is a carrier of meaning, furnishing a real learning experience for the imagination and perceptions.
The work on timbre is essentially based on the compression and expansion of formants or spectral envelopes, by departing from vocal types of material and deviating from these models towards others, with or without references to musical families of instrumental timbres. These models and their deviations provide a great homogeneity of material despite the fact that there is constantly a wide variety of transformations in the texture. The sound synthesis was made with the CHANT program at IRCAM, which is a program for synthesis through rules, based on the model of vocal production. Developed by Xavier Rodet and Yves Potard, it uses a library of "instruments," which are materials centered around the voice, but also departing from it. These include all families of instruments like strings, winds and percussion, and "instruments" more distant from existing instrumental references.
Though the sonic materials have been carefully worked out, special attention in this piece is given to organization. The research on musical processes, their fields of action and their limits, has been thought of as a strategy for approaching the musical territory as it is renewed by the possibilities made accessible through the use of computers. The types of processes in the piece are concerned with the phenomena of change, of transition, or of trajectories defining lines between points, structured formally and in time. These trajectories are strategies for the exploration of a terrain, designed to invade the sound space in all of its dimensions. Based on these processes, the piece navigates through libraries or structured lists of values for parameters such as frequency, amplitude, duration, vocal phonemes (which basically describe spectral envelopes) and the compression/expansion of formant frequencies.
The beginning of the piece employs processes that reveal the development of the timbre at different harmonic and melodic levels according to the compression/expansion of spectral envelopes based on vocal phonemes. This leads to an ostinato section also controlled by processes at various levels, and in the latter portion of the piece these processes produce a slow ascent in a synthesized vocal choir, which becomes increasingly rapid and high, rounded off by a gong-like stroke that dissolves into a distant choir. Compositional structures as well as scores were elaborated in the FORMES programming environment, with the help of Pierre Cointe and Yves Potard. Developed by Xavier Rodet, Yves Potard and Pierre Cointe, FORMES is essentially devoted to the development and manipulation of structures in time. A number of elaborate rules describing the behavior of the different parameters was defined to generate evolutions for the internal, or microscopic, structure of sounds. The macroscopic structure was managed through various processes interacting at different levels of organization, controlling pitch, rhythmic, metric and timbral structures. For instance, libraries of selected values for a given parameter were first built in FORMES; complex patterns were written that were used to "read" these libraries, and then calculate the values for a given sequence. The resulting data was then sent to CHANT for sound synthesis.

barriere.mov

Le programme Chant

Design: Xavier Rodet, Yves Potard
Development: Analysis/Synthesis group for Unix, Ircam Signal
Processing Workstation (ISPW)ß and Macintosh versions: Gerhard Eckel, Francisco Iovino, Adrien Lefèvre and Dominique Virolle
Resonance Models: Pierre-François Baisnée, Yves Pochard, Jean-Baptiste Barrière and Francesco Marti.

At Ircam, research on synthesizing sung voice had already begun in 1977, thanks to the work of Xavier Rodet. With Yves Potard and Gérard Bennett, Rodet developed the program Chant.
Originally conceived for singing voice synthesis, Chant turned out to be well-suited to simulating other instruments, as well as rich for synthesis in general. The program has branched out in numerous ways: artificial voice, sound filters, resonance models.

Vocalising - from a man's voice to a woman's voice:
rodet16a.mov

Decomposition - progressive movement from isolated glottal stops to a man's voice:
rodet17.mov 4


1 http://www.petals.org/Barriere/bio-english.html
2 http://www.petals.org/Barriere/Chreode-english.html
3 http://musicweb.koncon.nl/
4 http://www.music.columbia.edu/masterpieces/notes/barriere/index.html