The fundamental principle of IRCAM is to encourage productive interaction among scientific research, technological developments, and contemporary music production. Since its establishment in 1977, this initiative has provided the foundation for the institute’s activities. One of the major issues is the importance of contributing to the renewal of musical expression through science and technology. Conversely, sp…
IRCAM is an internationally recognized research center dedicated to creating new technologies for music. The institute offers a unique experimental environment where composers strive to enlarge their musical experience through the concepts expressed in new technologies.
In support of IRCAM's research and creation missions, the educational program seeks to shed light on the current and future meaning of the interactions among the arts, sciences, and technology as well as sharing its models of knowledge, know-how, and innovations with the widest possible audience.
Engaged with societal and economic issues at the intersection of culture and IT, research at Ircam has forged a reputation for itself in the world of international research as an interdisciplinary benchmark in the science and technology of sound and music, constantly attentive to the new needs and uses in society.
The fundamental principle of IRCAM is to encourage productive interaction among scientific research, technological developments, and contemporary music production. Since its establishment in 1977, this initiative has provided the foundation for the institute’s activities. One of the major issues is the importance of contributing to the renewal of musical expression through science and technology. Conversely, sp…
IRCAM is an internationally recognized research center dedicated to creating new technologies for music. The institute offers a unique experimental environment where composers strive to enlarge their musical experience through the concepts expressed in new technologies.
In support of IRCAM's research and creation missions, the educational program seeks to shed light on the current and future meaning of the interactions among the arts, sciences, and technology as well as sharing its models of knowledge, know-how, and innovations with the widest possible audience.
Engaged with societal and economic issues at the intersection of culture and IT, research at Ircam has forged a reputation for itself in the world of international research as an interdisciplinary benchmark in the science and technology of sound and music, constantly attentive to the new needs and uses in society.
Next season, Philippe Leroux will be working on his first opera, commissioned by Angers-Nantes Opéra and in association with IRCAM. The composer has been working on this music for almost forty years, without ever having had the opportunity to produce it for lack of a suitable text. The text will be Paul Claudel's L'Annonce faite à Marie, from which Raphaèle Fleury drew the libretto, to be directed by Célie Pauthe.
From left to right, the composer Philippe Leroux © Pierre Raimbault and the and the director Célie Pauthe © Ishak Ali Anis
When it comes to composing, the composer has focused primarily on two topics. The first area of focus was the generation of the melodic and harmonic movements, which compose the musical framework of the opera, drawn from the graphics of Claudel's writing. "By analyzing the gestural data (shape of the letters, thickness of the strokes...) and by transposing these data in the musical field, I generated most of the melodic and harmonic movements which compose the musical framework of the opera," says Philippe Leroux.
Phrase from the Paul Claudel’s manuscript La jeune fille Violaine (first version of L'Annonce faite à Marie). This was used to calculate the opera’s scales and pitch.
This work echoes that of a previous project, Quid Sit Musicus, a sort of musical dialogue between the composer and one of his elders: Guillaume de Machaut. Using augmented paper, Philippe Leroux extracted several parameters from Machaut's manuscripts corresponding to the gestural data of calligraphy (position of the graphism on the page itself, spatial arrangement of the staves, speed and pressure of the pencil, time required to draw the strokes, thickness, perimeter and surface of the forms... ) to determine melodic gestures and pitch scales, sound treatments, synthesis, spatialization, or even to generate distortion processes and the use of the sound, modulation, transposition or acoustic effects simulation processes based symbolic data (pitches, durations, etc.).
The second subject the composer focused on, with the help of Christophe Veaux and Carlo Laurenzi, was the recreation of Paul Claudel’s voice.
"My idea is to stage Claudel musically, sometimes having him intervene in his opera, as if he were dreaming, or in the process of writing his text, reciting it to himself, or as if he were guiding us in our listening by underlining this or that expression. It goes without saying that I only have Claudel recite fragments of his text. I wanted, in fact, to accentuate the deeply autobiographical and human nature of this drama, in which Claudel crudely depicts the various passions and love rivalries of the two sisters, as well as the reactions of their mother and sometimes the cowardice of the men who love them. One finds many connections with the Claudel family in these characters, including Camille."
Original voice of Paul Claudel
Paul Claudel's voice, synthesized
To recreate the playwright's voice, Philippe Leroux, Christophe Veaux, and Carlo Laurenzi used a neural synthesizer: "This is composed of two neural networks," explains the composer. "The first one predicts a spectrogram based on the vocabulary of the libretto, and a second network (called neural vocoder) converts the spectrogram into audio samples. The first network (which works from the graphemes [letters]) is thus responsible for intelligibility and prosody, while the second one is responsible for acoustic quality. As we have few recordings of Claudel's voice—and they are of poor quality—we had to carry out several tests and use different types of neural vocoders. In the very beginning of the opera, you first hear real recordings of Claudel's voice and then, little by little, the synthesized voice takes over."
Data extracted from words
In a way, L'Annonce faite à Marie is an extension of Philippe Leroux's long-term research on the voice. Many of the sound files used in the work come from vocal transformations. "Claudel's voice is transformed by concatenative synthesis, various filters and all sorts of effects, but also some pre-recorded singers. I also introduced musical moments elaborated from Gregorian chants (mentioned by Claudel in the didascalies), which are filtered as if they came from far back in time."
"Musical elements are born, like the author’s voice and calligraphy, » concludes Leroux.
A word (moi), from the first phrase
An example of the creation of a scale using the script of the word 'moi’ to calculate the pitch profiles in OpenMusic
1997 - M for 2 pianos, 2 percussions and electronics, 15 mn
2002 - Voi(Rex) for voice, 6 instruments and electronics, 23 mn
2005-2006 - Apocalypsisfor voice, 16 instruments and electronics, 25 mn
2014 - Quid sit musicus ? for 7 voices, 2 instruments and electronics, 18 mn