My current research is concerned with the representation and automated reasoning on time in music, from two point of views: notation (software for music edition and computer-aided composition) and performance (interactive music systems).
From the notation point of view, I am working on the development of techniques and tools for processing digital documents in common western music notation (music scores), with a focus on the proportional notation of rhythm. Our approach is based on symbolic computation, with formalisms for tree-structured data processing, such as Term Rewriting Systems and Tree Automata. Trees are indeed a natural way to represents rhythms, which are defined hierarchically in music notation. In this setting, we are addressing Music Information Retrieval problems such as rhythm transcription and efficient querying in music scores.
From the point of view of music performance, our objective is to improve the reliability and predictability of realtime interactive music systems involved in live music performances with human musicians (a popular example being automatic accompaniment systems). For this purpose, we are using formal methods for verifying statically that a system will behave as expected during a concert. This involves in particular symbolic model checking procedures and static analysis of the property of schedulability.
You can look at some selected publications to get more details.
- Partenariat Hubert Curien (PHC) Amadeus franco-autrichien LETITBE (Ircam, U. Salzburg).
- EFFICACe. ANR-13-JS02-0004.