Simon Fraser University

Naming that tune just got easier

February 23, 2006 , vol. 35, no. 4

By Marianne Meadahl

The simple, rhythmic bopping of a finger has led a trio of SFU computing science students to solve a musical dilemma - how to name that unknown tune.

They call it Song Tapper, and it can be found at The website has been designed to enable users to identify songs by tapping the melody on their spacebars.

The site has grown from a few hundred songs to well over 11,000 tunes and attracts as many as 10,000 hits a day.

Geoff Peters, a jazz pianist who graduated from SFU in the fall, says the group came up with the idea last spring while brainstorming for a project for an artificial intelligence course.

"We were originally thinking of creating an automated song lyric generator," says Peters, creator of Google Duel, another internet tool designed to gauge name popularity on the internet. "But instead of matching rhythms in music to rhythms in lyrics, we decided to try matching rhythms in music to other rhythms in music."

Students came up with an algorithm, which takes a rhythm and determines a rhythmic contour, or profile of the way the rhythm changes, like a first derivative from calculus. They generated a string of characters that represent the rhythmic contour, then used approximate string matching to do the search.

The students ran a test with 30 children's songs. After initial success, they turned the lab project into a website.

Peters and colleagues Caroline Anthony and Michael Schwartz did the coding and implementation of Song Tapper's algorithms. They also worked on the documentation, added new ideas and raised funds to present the concept at a major conference in Pittsburgh.

"A problem with some music search programs is that they require the user to hum or whistle into a microphone, and many users don't have that," notes Peters, "or they might be tone deaf. The space bar makes it more accessible."

The site also uses an automatic song-learning technique allowing users to teach the system new songs by tapping and then entering the song name.

Students developed the project in a computing course taught by professor Diana Cukierman, who encouraged them to submit a paper to an international conference on artificial intelligence and give a number of presentations, which have all led to international interest in the project.

The students have submitted a chapter about Song Tapper to a book on artificial intelligence called Ambient Intelligence for Scientific Discovery. "We have lots of ideas on where the tapping project could go, including commercial and educational applications," says Anthony, who is majoring in cognitive science.

"It could be used on internet music sites for music search and discovery, on mobile devices for ring tone search, or in schools to teach rhythm. It could also possibly be used as a way for a computer to identify a person - like a password."

Tapping could also eventually be incorporated into children's toys, adds Schwartz. "Our goal is to help the site grow, through enlarging the number of songs it can recognize, and perhaps down the road, hooking it up to an online music store. There are many possibilities."