Live coding with Adapt

Adapt is an ongoing project that aims to explore the use of machine listening and adaptive signal processing to achieve distributed agency in a live coding environment. Adapt’s behaviour emerges from the relationships between smaller components that can perceive changes in their environment and react accordingly. This is achieved by combining the analysis of audio signals and the tracking of system behaviours, with an adaptive control of certain sound parameters and the ability of the software to reconfigure itself.

Rather than focusing on the creation of virtual agents modelled on human behaviour, Adapt takes a different approach distributing the agency between various smaller components that can perceive changes in their environment and react accordingly. The sonic outcome is the result of all these components playing an active casual role and jointly governing the behaviour of the system influencing each other, distributing agency working as a cognitive assemblage. In Adapt, the cognition of the machine emerges through a series of action-perception relationships that connect each component with its perceived environment and keep track of analysed data.

Adapt was first presented at the International Conference on Functional Programming (ICFP 2022), Ljubljana, Slovenia, 11-16 September 2022
A full video of the performance is available at this link
To cite this work please refer to this source