From natural gestures to AI signal (SigNature)
Person(s) in charge
Team
Hajer Ben Boubaker
Partners
Ensemble Flashbackk, Association Flashback 66
Start date: 2026
Presentation of the project
Following on from the prototype work SINE (Gilbert Nouno, 2023, revived in 2025), the SigNature project explores gesture as a compositional signature: a situated inscription that transforms physical action into sound and visual writing. The device converts gestures captured by a camera and analysed by an AI algorithm into audiovisual decisions. The performer becomes a virtual percussionist in a choreographic-musical score where gestural mapping takes the place of grammar. It is a contactless organology where the perceptual and computational chain structures the very form and tempo of the work, and where decision-making is distributed between performer and machine.
In concrete terms, an interactive environment links gesture, sound and image: a movement calls up a colour, a texture, a shift in the sound space; a trajectory responds with a visual motif or an acoustic relief. SigNature shapes and shares this chain of gestures interpreted by AI through to the audiovisual rendering, using open and reusable tools, workshops and a pilot creation.
The research focuses on: the reliability of the device on stage (despite changing lights and different viewing angles), a vocabulary of correspondences that allows for both control and the unexpected, and a replicable transmission with clear resources that anyone can use (kits, examples and bilingual documentation).
Supported by the HEM and an artistic, cultural and technical partner, SigNature offers a philosophy of instrumentality adapted to contemporary practices: from the studio to the stage, from teaching to installation, it opens up an ecology of uses (augmented instrument, scenography), while re-articulating human-machine mediation at the heart of the musical experience.
