About the Project

Eavesdropping is an internet-based, interactive audio system that explores network mediated, musical performance in shared public spaces. In public environments, individuals interact by means of a variety of bodily and auditory cues and gestures. These ambient communication techniques can be directed at specific individuals or may be general expressions of mood meant for anyone who happens to notice. Visitors to public spaces, such as a café, seek the passive awareness of others to achieve a sense of connectedness born of shared experience, like the audience in a music venue. This project highlights the exhibitionism and voyeurism in the public sphere by amplifying participants' moods via music and increasing shared experiences to encourage deeper interaction. It aims to develop an environment which increases audience interaction and connectedness in a localized, computer-controlled performance.

The system is a client-server architecture made of three components: (1) an audio preparation interface, (2) an interactive performance interface, and (3) a machine learning-based conductor.  Musician's have contributed files to represent participants' moods.  Participants input their mood during the performance.  An artificial conductor mixes an acoustic ecology based on mood data.  Participants are encouraged to respond to whether the audio represents the mood they've input.  This allows the system to learn from audience response to more accurately represent participants' moods.

The link above begins the most recent version of the performance.

Eavesdropping was designed as a composition environment to allow musicians to compose using moods.  The performance was intended for a group of networked computers in the same room such as an internet cafe.  Mood-tagged music files play at each participant's computer exposing individual moods publicly from private computers.  Initial performances with this system drew various responses from participants but most expected more interaction and the ability to control or shape the music in the room.

• Other Stories: An evening of interactive arts. March, 2007. Surrey, BC.
• 404 International Festival of Electronic Art. May 2007. Triest, Italy.

• Stockholm, Jack. Eavesdropping: Network Mediated Performance in Social Space. 2007.

The second version of Eavesdropping was developed to address issues of audience agency raised in the first version as well as to extend the artificial intelligence of the system in validating mood tagging of audio files.  This version removes the composition environment and instead requests the participants to enter their own moods.  The system then selects audio files which match those moods to perform in the room while finding files that mix well with each other.  Users respond whether they feel the music matches their mood.  This is recorded by a machine learning system to improve the mapping of music files to moods.

• ACM Multimedia Presentation. October 2008. Vancouver, BC. <FUTURE>

• Stockholm, J., Pasquier, P. Eavesdropping: Audience Interaction in Network Audio Performance, 2008.