light box instrument · gesture recognition · ongoing

Built in Max/MSP with MFCC-based sound analysis and machine-learning gesture recognition, Cubis explores the space between architecture, body, and sound.

Research question — How can touch alone generate an organic interactive instrument, one that responds to the hand the way a living thing would?

A wooden light box that is played by hand. Microphones and sensors read touch and gesture across its surfaces; a machine-learning classifier trained on my own playing distinguishes one kind of contact from another; sound and DMX lighting respond in the same instant, so the object lights from within as it is touched.

Cubis comes out of a simple curiosity: an interface that can react to different touches. 

The small cube (18-inch) is an object to hold, to hug. The large one (36-inch) is a shelter — something that can be broken from the inside out, and perhaps a rare kind of instrument to play from within.

Wrapped in tracing paper, the cube resonates acoustically like a drum. Through electronics, gesture recognition, and machine learning tools, it develops a richer sonic vocabulary, responding to touch and movement like a living instrument — or perhaps a living being.

Versions

Still Warm (small cube, 18-inch) — for one performer

Rebirth (large cube, 36-inch) — for two performers