University of California, San Diego artificial intelligence researcher Luke Barrington is developing software that can analyze a piece of music and compile information about it that could be useful in making a playlist. The software can assign the music a genre or give it subjective descriptions such as whether or not a track is "funky."
Barrington wants to create a system that can distinguish between different styles of music within a single song. For example, if a user chooses a song with a mellow verse and a loud chorus, the system would be able to recommend songs that fit that pattern.
However, before software can analyze a piece of music, it must understand what distinguishes one genre of music from another. Early approaches to this problem used speech recognition technology such as the mel-frequency cepstral coefficients approach, which is useful for determining which instruments are being used in a piece.
However, the University of Sao Paolo's Luciano da F. Costa is using rhythm to assign a genre to music, which he says is simple to extract and is independent of instruments or vocals. After analyzing a collection of MIDI files, da F. Costa's team was able to establish models of the note transitions characteristic of rock, blues, reggae, and bossa nova songs.
From New Scientist
View Full Article
Abstracts Copyright © 2010 Information Inc., Bethesda, Maryland, USA