[radical] signals from life: from muscle sensing to embodied machine listening/learning within a large-scale performance piece

This paper describes an approach to designing responsive, intelligent sonic interaction in a choreographed dance/media context in which biophysical signals from five dancers are mapped across multiple sense modalities in an immersive black box context. The sound of muscle activity was used as the sole driving signal to generate a musical composition, whose structure was determined by an intelligent agent, designed through an embodied cognitive view on machine listening and learning. This work was undertaken in the context of the [radical] signs of life piece, a collaborative project that involved the author as composer and interactive sound designer, in collaboration with an international team of artists and technologists.