Interview with Brigitte d’Andréa Novel
Engineer and research director in automation, Brigitte d’Andréa-Novel has been a professor at MINES ParisTech since 1992 and the head of the Systèmes de Contrôle Avancés (Advanced Control Systems) team in the school’s robotics center.
In January 2018, she was named director of the “Sciences and Technologies of Music and Sound” research lab (IRCAM-CNRS-Sorbonne Université-Ministère de la Cuture / UMR 9912) as well as professor at Sorbonne Université, teaching classes in control and robotics, a discipline with applications in the field of active control for musical instruments. Brigitte d’Andréa-Novel is also an accomplished pianist and enthusiastic teacher.
Brigitte d’Andréa-Novel, you are a scientist and director of research in automation, internationally renowned for your work on non-linear commands for which you received the bronze medal from the CNRS in 1989. Can you tell us a little more about your research subject?
My research subjects focus on the theory of non-linear control and applications for underactive mechanical systems, like robotized rolling bridges, as well as automotive control and applications for intelligent transport systems. I am also interested in the stabilization of hybrid systems coupling ordinary differential equations and partial differential equations with applications for the internal control of flexible mechanics, irrigation channels, modelling, estimating, and controlling wind instruments.
As a musician, you are dedicated to bringing together science and music. You have worked with the teams at IRCAM throughout your career.
I started to study piano, harmony, and analysis when I was very young and still enjoy playing chamber music today. I have to admit, it was really hard to choose—as we often must in our country—between mathematics and music. I really felt that the two paths were intertwined and part of the same intellectual practice. Fortunately, attitudes change, and we should be overjoyed by the existence of the “Sciences and Musicology” program at Sorbonne Université that enables undergraduates with a high level of musical proficiency to continue a double major through their bachelor’s studies.
At the École des Mines de Paris, where several students are musicians and I organize the festival “Musiques aux Mines”, I created the multidisciplinary course “Acoustique-Informatique-MusiquE” (Acoustic-Computer Science-Music) (AIME). The idea was to present a large palette of scientific tools for music: the connections between mathematics, harmony, and composition with software tools and materials associated with the MIDI standard, examining also sound processing methods used for the analysis and synthesis of musical sounds, for automation of real and virtual instruments formed using the laws of acoustics. A book* was written based on this course with Benoît Fabre (acoustician at LAM, SU) and Pierre Jouvelot (computer scientist at MINES ParisTech) that presents the major works on synthesis via physical models, estimation, and controlling wind instruments. I carried out a part of this research from 2005 to 2008 in the ANR project CONSONNES (Contrôle des Sons instrumentaux Naturels et Synthétiques), coordinated by Jean Kergomard (LMA, Marseille). This research was carried out in collaboration with Thomas Hélie (STMS, CNRS-IRCAM), Benoît Fabre (LAM, SU), and Jean-Michel Coron (Laboratoire Jacques-Louis Lions, SU) and extended the results obtained on mechanical and hydraulic hyperbolic partial differential equations as well as on acoustic partial differential equations.
But my first collaborations with IRCAM was in 1992. I worked with the researchers from the Analysis/Synthesis team on the control of the equation of sound waves with derivatives fractioned over time, modelling the effects of viscothermic losses in the guiding waves. In terms of applications, the objectives of this research is to develop mathematical models from physics to represent, with precision, the dynamic of wind instruments for a better analysis of their acoustic behavior as well as for automatic control in real-time.
The STMS lab at IRCAM © Philippe Barbosa
What are your projects for the “Sciences and Technologies of Music and Sound” lab?
Just like in several other national labs, we must diversify our funding due to the decrease in support from the ANR (national French research agency). The researchers at IRCAM have a good rate of success with European calls for projects and I would like to encourage them in this endeavor. Submissions to the European Research Council and participation in projects for industrial research chair programs or sponsoring would also be fitting given the quality of the lab’s research.
I also want to reinforce the events opening the laboratory to a non-specialized audience taking into account the evolution of musical practices—notably in composition—with the increase in the democratization knowledge and technologies. Today, with the development of standards such as MIDI and MP3 and software for synthesis, sequencing, and composition, anybody can build their “home studio”. IRCAM needs to anticipate this change in creative artistic forms, and, more generally, reinforce the connections between scientists and composers.
Aside from the software aspect, we must also take the democratization of technologies developed from motion capture that enable users to develop new SmartInstruments, sometimes at a lower cost. It is therefore essential to continue our strategy of innovation by encouraging researchers to file patents, with the support of the CNRS and the DIRE (Direction de l’Innovation et des Relations avec les Entreprises) as well as the creation of start-ups based on this valorization such as Antescofo founded by Arshia Cont on score-following software and synchronous programming for musical composition or SYOS founded by Pauline Eveno on the fabrication of mouthpieces for wind instruments using 3D printing technologies.
I think it is also important to mix music with other live performance medias: theater, images, film… This incites investigation into issues of artificial vision and image processing using 3D cameras, for example. And, more generally, investigation into virtual and augmented reality. It will be interesting for the lab to connect with labs specialized in the field, like the robotics center at MINES ParisTech or ISIR at SU. This theme could, in effect, further the analysis of scenes in real-time with the fusion of data, therefore enabling a better interpretation of musicians’ movements that should set off electronic sounds and also a better music/image alignment.
And what about your projects as a researcher? Because you are also a part of the Sound Systems and Signals team at IRCAM…
Given my training as an automation expert and the research I have developed over the years at the robotics center at MINES ParisTech, I am more and more convinced of the importance of the discipline of “automation” for research on music and instruments. This feeling was made even stronger during my participation in the CONSONNES project—the word control is even the first word of the project’s acronym! A musical instrument is basically a mechanical system we can equip with captors and actuators to which theories of automation apply in order to respond to different objectives: modelling the system, estimating parameters, inversing the system, developing observers, and controlling the instrument. This can all be done without the musician: we use a “robot musician” (see the mecatronic projects developed at MINES ParisTech) or in relationship with one, with the objective of a “human-machine” collaboration using technologies for motion capture, recognition, learning, and following. It is in the context of “active control of an instrument” that I intend to carry out my research, on an academic level, on a practical level, but also on an applicative level using conclusions drawn by the augmented instruments team and the COALA sensor/actuator system developed by the start-up HyVibe, founded by Adrien Mamou-Mani and based on research carried out at IRCAM.
* éditions Presses des Mines, 2005
Photo: The COALA sensor/actuator system © Philippe Barbosa