Hugo Geiger Prize 2012

Automatically classifying world music

In her thesis at the Fraunhofer Institute for Digital Media Technology IDMT in Ilmenau, Anna Marie Kruspe developed an automated classification process for non-Western music genres. Using this process, she achieves an accuracy of 70 percent. This is equivalent to the existing systems for classifying Western music.

The music market is in a state of flux: Since the era of digital formats, music is being increasingly offered online to a globalized market. Automated processes help genres to be reliably classified and music archives to be managed affordably and efficiently. For that to happen, music databases must be pre-processed. Commercial solutions are already in place for standard music genres such as rock and pop, but not for world music. In her thesis at the Fraunhofer Institute for Digital Media Technology IDMT in Ilmenau, Anna Marie Kruspe developed an automated classification process for non-Western music genres. Using this process, she achieves an accuracy of 70 percent. This is equivalent to the existing systems for classifying Western music. So in future, “world music” can also be integrated into the international music market - benefitting both composers and consumers alike. The thesis emerged as part of »GlobalMusic2one«, a project supported by the Federal Ministry of Education and Research.