New plans for Junction in 30’ x 30’ (9m x 9m) space. 10 tower multichannel array with “kiosk” of video grid. 

More testing with Junction. Better hue calibration depending on time of day, improved tracking, larger sound palette. Thanks Rachel for the stills.

Junction is a sound sculpture powered in real-time by thousands of New York City taxi cabs controlling Cycling74’s Max/Jitter and Symbolic Sound’s Kyma. Through the use of publicly available live footage from New York’s Traffic Management Center, Junction tracks the movements of taxis in some of the city’s busiest intersections. It uses their position, velocity and overall density to synthesize sounds. The onlooker gains a feeling of omniscience, as if “watching from above”, while every tracked taxi’s indeterministic movements become part of a larger symphony.


A taxi’s position might control only one partial of a complex frequency spectrum, or a single frame of a granular cloud. Up to 40 taxis are tracked on the grid simultaneously. They are indexed and vectored every 50 milliseconds. Once a taxi leaves the scene, its index dies and a new taxi is tracked. Junction is powered by Cycling 74’s Max and Jitter with the cv.jit externals by Jean-Marc Pelletier. Data from Max are sent to a Kyma system for realtime synthesis based on dynamic event variables.


The included video documentation is but one iteration of Junction, edited to provide a sampling of the sound palette that is created. Just as traffic morphs over many hours so does a sonification of that pattern. Thus, for brevity’s sake, I have consolidated many of the generated sequences into the short documentation.

(Source: vimeo.com)