Multimodal Recursion is a live audiovisual performance by Joel Schäfer and Anna Ferro that treats sound and image as a single, entangled instrument. At its core runs a bidirectional modulation system in which visual parameters shape sound and sound reshapes visuals, a continuous loop in which control never fully rests with either medium. This automated entanglement is deliberately interwoven with the performers' own hands: audio-reactive elements and live MIDI-driven modulation sit alongside the feedback system, so the piece moves fluidly between moments the machine generates on its own and moments the performers actively steer. What emerges is neither a fixed composition nor a purely reactive visualization, but a real-time negotiation between system and performer, automation and gesture.

The piece unfolds in three stages: an atmospheric opening of clean geometric forms synchronized to harmonic sound, a fracturing into glitch-driven rhythm and live-performed particle systems, and a final dissolve into noise-inspired soundscapes via fluid simulation and pixel-sorting.

Sound is built from a modular synthesizer, an electronic drum machine, a tongue drum with a pickup mic, and granular processing in Ableton Live. Visuals are generated live in TouchDesigner and controlled via a MIDI controller. Bidirectional OSC links the two systems in real time, so sound shapes the image and performed visual gestures modulate the synthesizer in return.