2026 Webcam, contour software (red traces), generative sound system, local LLM (not connected to the network), screen Performers: dancers (to be defined) First presented at: to be defined
A webcam captures the dancers’ movement and translates it into thin red contours that appear on the screen. The contours remain visible for a few seconds, then fade. The sound arises from the same logic: when the bodies move, the sound field is excited and generates frequencies; when they stop, the sound collapses towards silence. Every spatial relation between the bodies — distance, proximity, repetition — is converted into an acoustic event. A local language model, installed on the machine and not connected to the network, generates short aphorisms that appear on the screen and vanish after a few moments. The visual, sonic and textual traces are erased immediately after they appear. Nothing is recorded, nothing is accumulated. The system continuously produces and continuously eliminates. The dancers do not see the contours, do not read the aphorisms, do not control the sound. The device runs in parallel with the bodies, without feedback.
Parallel regime. The LLM generates aphorisms, the visual system traces red contours, the sound system generates frequencies from movement. All traces are erased immediately. The dancers do not see the contours, do not read the aphorisms, do not control the sound. Here the device runs in parallel with the bodies, without feedback.
Photo: Cristina Crippa Produced by artQ13.