Detalhes do Documento

This paper investigates the potential of real time movement analysis as a driver for generative audiovisual performance. Premised on a generative logic of movement fueling transformation, we propose an experimental pipeline — AVLMA — that fuses Laban Movement Analysis (LMA) with real time motion capture using MediaPipe. The system extracts spatial and qualitative features of bodily movement to modulate generative sound and visuals in a live performance context. Key components include: particle systems that trace bodily landmarks with expressive variation; a swarm based visual model employing boid dynamics to reflect motion complexity; cellular automata that evolve in response to movement flow characteristics; and a FAUST based audio engine responsive to LMA derived parameters. These layers interconnect through a unified interface developed in Unity3D, forming a reactive, self organizing audiovisual ecosystem. Rather than functioning as background embellishment, the audiovisual system becomes an active agent in the experimental performance—interpreting and responding to the dancer’s evolving gestures. This study expands on the notion of choreography as computation and demonstrates how movement analysis can serve not only as input but as an aesthetic grammar for generative media