First Iteration

Current work-in-progress using RAVE + Claude + SuperCollider: github.com/IAndy-10/liveCoding-rave

Sound Example

Concept

NeoChucao takes the Chucao — a secretive Chilean forest bird — as its sonic seed. A neural model is trained on bird recordings, compressing ecological sound into a latent space. The performer doesn't play notes. They navigate a landscape of learned sonic behavior.

The system responds to the environment it runs in: wind speed, humidity, and air quality modulate the sound in real time. Gesture and movement shape the trajectory through the sonic space. The goal is an instrument that feels alive — not because it is programmed to, but because it is listening to the same world the performer is.

Technical Detail

• —Neural audio model (RAVE) trained on Chilean bird recordings — the latent space is a compressed map of ecological sound.
• —Environmental data (weather, air quality) feeds into the instrument in real time, shifting its timbral behavior with the conditions outside.
• —Performer gesture and movement navigate the latent space, replacing traditional note-based performance with spatial exploration.