Date of Submission

Spring 2013

Academic Program

Computer Science

Project Advisor 1

Sven Anderson

Abstract/Artist's Statement

Is it just a statistics problem?

How can a listener tell which instruments are playing together without reading the musical score? Can a computer algorithm separate different instrument voices automatically? Blind signal separation tries to separate instrumental voices based on statistical regularities in the signal. This project focuses on recordings of two different instruments simultaneously playing the same notes. Independent components analysis, one form of blind signal separation, was used to separate signals collected by two independent microphones. The experimental results show that voices combined by adding signals mathematically are separated better than voices combined acoustically and collected by two independent microphones Instrument timbres also affect the ability to separate voices. Statistics can separate some artificially constructed signals, but separating recorded signals is much more difficult.

Distribution Options

Access restricted to On-Campus only

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.

Download

This work is protected by a Creative Commons license. Any use not permitted under that license is prohibited.

Bard Off-campus Download

Bard College faculty, staff, and students can login from off-campus by clicking on the Off-campus Download button and entering their Bard username and password.

Share

COinS