The laboratory’s scientific project organized around 3 fields of interaction:
Here, sound is understood as a physical phenomenon (mechanical, acoustic) as well as digital data involving mathematical and computer science (signal processing) techniques to process data audio streams and musical or multimedia scenes.
Digital audio signals are addressed in as sources and fields. We characterize or synthesize digital audio sources with the accuracy, subtlety, and quality required to meet the demands of contemporary creation. Our research focuses on the specificities of the musical signal and has led us to develop more precise models employing physics, biomechanics, perception and cognition, and even symbolic musical data. The extraction of more a symbolic, semantically pertinent description and the control of models via high-level qualitative data (e.g. prosody in the case of the synthesized singing voice) are two of the laboratory’s objectives in this domain. As for sound fields, our studies aim at understanding their spatial representations and couple them with time/frequency representations, at capturing them in high resolution, and at reconstructing them realistically.
The Musical Body
The domain focused on here is that of music and sound, perceived and produced by humans who interact with an environment, a machine, or other human beings. Sound is not reduced to a simple vibration, but becomes a vehicle enabling interaction produced by a biological body and operating on individuals and behaviors. This domain is addressed through four issues—gesture, voice, the spatial perception of sound and its multi-sensorial integration—while examining the connections among sonic or musical contents and emotional valency.
While music is a phenomenon produced and perceived by a body that capitalizes on physical objects and organizes a sound world, it is also thought, imagined—or dreamt— and occasionally formalized before being realized. The question of creativity emerges today as a scientific1 issue and a useful working paradigm for the conception of intelligent systems. Contrary to what a superficial analysis of the new directions taken by AI2 could suggest, the challenge seems to be more about mentoring and companionship than imitation: what ideas and tools are necessary to inspire, guide, and assist an artist in their creative process and make the machine a partner? We are therefore opposed to an approach that aims at replacing humans, making autonomous machines and that which aims at augmenting human capacities. We can therefore talk about augmented intelligence.
In the artistic field, the first approach could possibly lead to understanding mechanisms that are replaced, but the second offers novel creative dimensions that could radically renew the science-music relationship. Using one of our major assets—our intimacy with musical creation—the laboratory’s work endeavors to apprehend and document composers’ creative process; to formalize, analyze, calculate, and produce symbolic musical structures in interaction with the creators’ goals; to develop new tools for writing that address contemporary issues of notation; and to capture systems that assist collective creation.
These three fields ensure the legibility of our research and cover all our activities, they cross the physical and digital dimensions of the sound and musical world as well as its perception and production by humans in connection with issues surrounding creation and creativity. Among these research subjects, we can cite a few examples of subjects we intend to investigate more thoroughly in the upcoming years:
- Multiphysical modelling
- Expert expressive gesture
- Cyber physical & human systems
- The contribution of neurosciences, cognitive sciences, and connections with health
- AI and musical creativity
- New dimensions of musical notation
1- See the new international conferences in which the joint research lab has participated since their creation “Digital Intelligence (#DI)”, “Musical MetaCreation (MUME)”, “International Conference on Computational Creativity, (ICCC)”, or the European initiative FET “Creative ICT” that funded 5 projects in 2016.
2- Like the Magenta project developed by Google: “(how to develop) algorithms that can learn how to generate art and music, potentially creating compelling and artistic content on their own”.