Neon Music – turning cities into sound


Presentation Title:

  • Neon Music - turning cities into sound




  • “Neon Music” is a 3 channel interactive video installation, designed to analyze video for motion and color information, and turn it into music.

    It was initially based on footage of neon signs that I shot in Tokyo, since these have a strong rhythmic and repetitive aspect. Over time, I have collected similar footage from other cities, and started to identify characteristic “sounds” for each city, based on the activity of their night-time signage.

    My interest in this area came from my background in providing live visuals for concerts and events. My challenge was always to try to match the mood of the music with my imagery. In this piece, I have reversed the dynamic: I start with imagery, and generate music to match.

    The piece is implemented in the Max/MSP programming environment, and is meant to be played live. A performer chooses from a bank of between 48 and 96 video clips, which are then analyzed and interpreted in real time. When presented in its three channel version, three clips at a time are projected onto separate screens, each controlling a different instrument sound.

    In order to achieve musical results, the program embodies some musical rules. Horizontal motion maps on to harmonic chords, vertical motion to changes in octave, overall saturation to chord type, and dominant hue to key.

    The process of collecting appropriate footage for this piece is interesting, and has become a complex project in its own right. It turns out that night-time signage in cities around the world has rhythms and colors that are specific to a place. Filming these in a way that retains rhythm and color information, yet remains abstract can be a challenge.

    Here are a couple of recordings of the piece:

    As performed in New York, using footage from Tokyo:

    As performed in Tokyo, using footage from New York:

    (This version was an experiment, using non-neon sign footage!)

    This piece can be presented live, as a three channel video performance, as a single channel performance (all three clips on one screen side-by-side) or as a high definition video recording.