Skip to content | Change text size
 

Music to our ears

29 July 2009

Person listening to iPod
 

Researchers from the Faculty of Information Technology are developing a new music classification system to make digital music downloading easier.

The team from the Gippsland School of Information Technology are developing software that will improve the way music is classified and annotated so users can find what they want faster and those selling music can add more products to their range.

The three-year Australian Research Council-funded project will develop techniques to automatically annotate music pieces based on automatically-extracted music, perceptual features and manually-annotated pieces.

Music classification is currently a manual process where music is divided into broad classes such as genre. Music annotation provides a more detailed description of music by using terms such as "fast and exciting" or "happy."

Professor Guojun Lu said it would mean a computer could automatically find a piece of music, not based on a text description or title, but rather on the music itself.

"Music is a huge industry currently undergoing a major revolution. The industry is shifting from music making to music retrieval, with music being incorporated more and more into products from TV and film, to music streaming into locations and events, as well as MP3 players and all kinds of electronic devices," Professor Lu said.

"If someone listens to music, they can tell if it is upbeat or sad and if it is country and or pop.

"We are developing a program that will automatically determine this, supporting the immediate retrieval of music, based not just on titles, composers and/or performers, but on the actual properties of the music itself."

Professor Lu said it could open up opportunities for emerging song writers because it would help, for example, filmmakers find music without knowing anything about the song or the composer.