Soundscapes of an Earthly Community

This project is inspired by Bernie Krause’s soundscape ecological approach, which sees sonic environments as equally constituted by biological sounds, such as plants (biophony), geophysical sounds, such as wind (geophony), technological sounds, such as machines, (technophony) and human-made sounds (anthrophony). These categories are not fixed but rather intersecting and relational. While some sounds might be distinctively human, others might only emerge when different spheres meet. For instance, the sound of the geophonic wind emerges when it brushes the leaves of the biological tree. It is these intersections then, where we intend to find the in-common that does not deny the uniqueness of its parts.

The project employs the Realtime Audio Variational autoEncoder (RAVE) AI model developed at IRCAM by Antoine Caillon and Philippe Esling. To be presented as a surround sound installation each speaker is fed by a distinct, generative RAVE model trained either on biophonic, geophonic, or anthrophonic sounds mediated by the technopohnic sounds of AI to jointly represent the speculative soundscape of an Earthly Community. The installation is interactive, as a series of microphones pick up any sound in the space including that of the AI-driven speakers and the audience and feed these as inputs into the RAVE model, thus creating interactive feedback loops.