Three Questions for Marco Stroppa on the Electronics in Poésie pour pouvoir
Poésie pour pouvoir, for tape and three orchestral groups, was premiered on October 19, 1958, in Donaueschingen by the Südwestrundfunk orchestra conducted by Hans Rosbaud and Pierre Boulez. The work was inspired by a poem by Henri Michaux, Je rame (I Row), which appeared in a book-object whose general title the composer adopted. Ambitious and complex, with its spatialized electronic part (the loudspeakers being distributed behind and around the audience), Poésie pour pouvoir did not satisfy the composer, who withdrew it from his catalogue but continued to explore the questions and materials that had emerged from it in other projects. His response was, precisely, Répons for ensemble and electronics—conceived a quarter of a century later at IRCAM, where Boulez was then director.
Premiere of Poésie pour pouvoir in 1958
With the agreement of the rights holders, IRCAM and the Lucerne Festival joined forces to bring Poésie pour pouvoir back to life for a concert. The redesign of the electronic part, made possible by 21st-century technological advances, was entrusted to Marco Stroppa and Carlo Laurenzi.
You have agreed to collaborate with computer music designer Carlo Laurenzi on a redesign of the electronic part of Poésie pour pouvoir. What interests you in this adventure, more than half a century after the first performance and Boulez’s decision to withdraw the work from his catalogue?
First of all, the nature of this project is exciting and unique. It's as if we had found a painting in a cellar that had been stored there for 62 years, damaged by humidity and mould, and partially covered with graffiti. Since restoration is out of the question, our task is to reconstruct it without having precise information about the techniques used. As our knowledge of the tools of the time is incomplete, we will not be able to achieve this without a substantial dose of imagination.
This “we” refers to a particular mode of collaboration that has nothing to do with the “classic” duo of a new work being produced in a studio by a guest composer and an IRCAM computer music designer. With the composer absent, Carlo and I work a little like the members of Renaissance artists’ workshops (the botteghe). We have complementary expertise and mastery of the techniques we use.
The picture, in this case, is the only existing recording: that of the work’s creation, made with a single microphone placed in the hall, which simultaneously captured the three orchestras, the electronics diffused around the audience by some twenty loudspeakers, the acoustics of the hall, and the ambient noise. Because of the spatialization, this recording is particularly unsuitable for reproducing the details of the electronics.
Another of my motivations stems from the technological approach adopted: cleaning up the existing recording as best as possible (with iZotope RX) and then, using AudioSculpt, selecting the most important spectral components by hand with the “Pencil” filter. Sometimes this is enough, but more often than not, a very detailed analysis is required using a technique known as partial tracking, which allows them to be manipulated symbolically. This is in line with one of the main research activities I have been pursuing at IRCAM since the 1980s, “sound synthesis writing,” which led me to develop a system for symbolic processing of synthesis data, now integrated into a library in the OpenMusic environment called OMChroma. In embarking on this project, I did not pretend to be a musicologist; rather, I approached an aesthetic problem (creating the best-sounding electronics in absolute accordance with the composer’s wishes) with tools that I know I can put at his service. But I see them in an unusual light—which is particularly stimulating—since this time I am not “the” composer, and my goal is not to invent a different kind of electronics.
In composing Poésie pour pouvoir, Boulez tackled several challenges at once: working on the spatial distribution of the orchestra; creating an organic, or at least continuous, relationship between the instruments and the electroacoustic part; in the latter, bringing together synthetic sounds and a human voice subject to transformation; and finally, giving the electroacoustic diffusion a sense of motion by spatializing it through loudspeakers. These issues were burning ones for the avant-garde at the time, as evidenced by Stockhausen’s works from the same period, such as Gesang der Jünglinge (coherent and spatialized use of synthetic and concrete sounds) and Gruppen (three orchestras creating a complex spatio-temporal polyphony). Many other musical proposals extended these early explorations, which quickly became outdated—even to their authors, as was the case here. This is even more true on a technical level, as contemporary computer music has revolutionized the studio practices established in the 1950s. Knowing that neither the original sounds nor the technical tools are available, what choices did you have to make between what could be preserved as is and what had to be completely or partially redone?
Two families of electronic sounds are easily discernible when listening: synthesized sounds (sometimes created from filtered and stretched instrumental samples) and a processed voice reciting Henri Michaux’s poem Je rame (I Row). These families are sometimes isolated, sometimes mixed with the orchestra. We reconstruct the synthesized parts using partial tracking: a large number of frequency and amplitude curves are produced (between tens and hundreds of thousands per minute), which are then manipulated in a symbolic environment that we developed for this project and which is much more manageable, powerful, and expressive than anything else currently available. This environment is integrated into OMChroma and coupled with the Csound synthesizer, which ensures the best possible resynthesis quality without limitations in duration or density. I have been using this kind of approach for a long time in my work with electronics. Whenever we find imperfections, we try to improve them. For example, sometimes certain attacks are not as clear as in the original. So, we programmed an attack detector coupled with an exciter: when a weak attack is detected, the system automatically makes it more incisive.
However, this process no longer works when the electronics are superimposed on the orchestra (which plays the role of the graffiti in my initial metaphor). With the help of Juanjo Burred, a former researcher with the Analysis–Synthesis team at IRCAM, we are currently testing methods for separating electronic and instrumental sources using the latest deep-learning algorithms, in order to achieve better results than we could obtain using manual techniques. Once the extraction is complete, we will reconstruct the missing parts using rules or by ear, supplementing the symbolic data with new data.
Unfortunately, this strategy cannot be used with the voice, due to the semantic role of the text and the poor quality of the original recording. It had to be reconstructed by ear, asking a contemporary actor (Yann Boudaud) to reread the text while respecting the diction of the original narrator, Michel Bouquet.
Finally, as the monophonic recording made it difficult to perceive the original spatialization correctly, we consulted Boulez’s sketches at the Paul Sacher Foundation in Basel with the help of Angela Ida De Benedictis and took into account Ai Higashikawa’s doctoral thesis on the electronics of Poésie, as well as the scores used by the conductors. These give us clues, especially regarding the vocal part, which was broadcast by a rotating loudspeaker placed high up in the middle of the audience. Compared to 1958, spatialization and loudspeaker technology have evolved so much that we believe we can offer a contemporary interpretation of the original space that is closer to the composer’s wishes, knowing how he later realized them—starting with Répons.
Poésie pour pouvoir is the work in which Boulez took his work on sound synthesis furthest, the only one in which a voice is treated, and the one that seems to me to come closest to the “limits of the fertile land” that he pointed out in a prophetic text in 1956. Admittedly, the electronic part does not reach the level of perfection of Stockhausen’s electronic works written at the same time, but it is one of the very first examples of mixed music, and the first work in which Boulez seeks a fusion that is both musical (through the materials used), spatial (through the placement of the orchestras and loudspeakers), and formal (through the succession of parts for electronics alone and with orchestra) between the world of synthesis, voice processing, and an already highly personal orchestral style. A sign of this desire to treat electroacoustics on an equal footing with the instrumental universe that Boulez mastered so well, Poésie pour pouvoir begins and ends with a section of electronics alone.
Originally, Poésie pour pouvoir was a book-object published by an art gallery. The two poems printed on linocuts in the book deal with rage and exorcism—a theme concretely reflected in the teak cover studded with nails that was created for part of the print run. Even if it comes across as attenuated in the historical record, this force fascinated Boulez, who still referred to it in 1999 in a testimony about Michaux, “Despite the incompleteness, not to say failure, of this essay, I still have a special regard for this poem, one of the strongest and least restrained he ever wrote. For the strange, even the precious, disappear when confronted with anger, threat, and pain. The personal universe explodes under the pressure of the unexpected and catastrophic event. Prudence has disappeared.” Does your restoration–reconstruction work revive your perception of the colours and emotions of the painting?
Cover and page from Henri Michaux's book-object Poésie pour pouvoir
After I arrived at IRCAM in June 1982, initially for a six-week course in computer music, I attended all of Boulez’s masterclasses and concert-lectures. I was fascinated by his dazzling intellect, his coherence and musicality, the depth of his references, and the effectiveness of the combinatorial processes he mastered to perfection. I also noticed the anger he could express in public interviews against what he perceived as the inertia of the musical system he wanted to improve or reform. The fact that IRCAM has remained a unique institution in the world for almost half a century is a testament to his fabulous abilities as a builder.
As I immersed myself in the sketches, I rediscovered this multidimensional spirit in its purest form: the initial path of thought that I had discovered at IRCAM appeared before my eyes with remarkable force and assurance. Indeed, the book-object produced by Michel Tapié, which “spaces out” two poems by Michaux and which researcher Lorraine Dumenil defines as a hapax, contains a text that is terribly violent and animated by a truly destructive rage linked to his wife’s illness. This text has great incantatory force, with abundant anaphora. “Power” is understood here in the operational sense of acting (Michaux likened the poet to the doctor). We find a form of violent energy in the sharp, accent-filled orchestration of Boulez’s work. And the incantatory aspect that appears here would return with force in 1974 in Rituel – In memoriam Bruno Maderna.
Because the recited voice is not always intelligible due to the electronics applied to it, Poésie pour pouvoir develops the continual interplay between presence and absence of Michaux’s text. At times the processing anticipates the voice; at others, the voice is reduced to an expressive contour, with only the narrator’s emotion perceptible. This fleeting presence, situated at the intersection of electronics, transformed voice, and orchestra, seems to me one of the essential interpretive keys to the work.
By Nicolas Donin, musicologist and professor at the University of Geneva
Photo 1: Premiere of Poésie pour pouvoir in 1958
Photo 2: Partition of Poésie pour pouvoir © SWR
Photo 3: Pierre Boulez in Baden Baden in the Südwestfunk studio in 1958.
Photo 4: Cover and page from Henri Michaux's book-object Poésie pour pouvoir
