Exploring Musical Agents with Embodied Perspectives

This chapter presents a retrospective of five interactive systems I have developed focusing on how machines can respond to body movement in music performance. In particular, I have been interested in understanding more about how humans and non-human entities can share musical control and agency. First, I give an overview of my musical and aesthetic background in experimental music practice and a less conventional approach to sound and music control. Then follows a presentation of embodiment and music cognition theories that informed the techniques and methods I employed while developing these systems. Then comes the retrospective section structured around five projects. Biostomp explores the unintentionality of body signals when used for music interaction. Vrengt demonstrates musical possibilities of sonic microinteraction and shared control. RAW seeks unconventional control through chaos and automation. Playing in the “air” employs deep learning to map muscle exertions to the sound of an “air” instrument. The audiovisual instrument CAVI uses generative modeling to automate live sound processing and investigates the varying sense of agency. These projects show how an artistic–scientific approach can diversify artistic repertoires of musical artificial intelligence through embodied cognition.