The dominant approach in the research area of Music Information Retrieval
(MIR) for audio signals has been to characterize musical content information
using overall statistics over time-frequency representations such as the
"classic" Mel-Frequency Cepstral Coefficients (MFCC). Although
this approach has yielded impressive results in various applications including
genre classification, similarity retrieval and music browsing it seems
to have reached a plateau. In this talk I will describe the efforts of
my research group at the University of Victoria to go beyond this "classic"
approach in two main directions: 1) extracting and analyzing specific "types"
of sound sources from audio music mixtures 2) connecting MIR techniques
with the process of music creation and performance. In addition I will
describe in more detail two specific projects: 1) formulating dominant
melody separation from polyphonic audio using perceptually-informed cues
as a graph partitioning problem 2) a system for capturing and analyzing
indian sitar performance for interaction with a robotic drummer.
George Tzanetakis is an assistant Professor of Computer Science (also cross-listed
in Music and Electrical and Computer Engineering) at the University of
Victoria, Canada. He received his PhD degree in Computer Science from Princeton
Uiversity in May 2002 was a PostDoctoral Fellow at Carnegie Mellon University
working onquery-by-humming systems with Prof. Dannenberg and on video retrieval
with the Informedia group.
His research deals with all stages of audio content analysis such as feature
extraction, segmentation, classification with specific focus on Music Information
Retrieval (MIR). His pioneering work on musical genre classification is
frequently cited and received an IEEE Signal Processing Society Young Author
Award in 2004. He is the principal designer and developer of the Marsyas
open source audio processing framework (http://marsyas.sourceforge.net
) which has been used for MIR projects both in academia and industry.
He has presented tutorials on MIR and audio feature extraction at several
international conferences. He is also an active musician and has studied
saxophone performance, music theory and composition. More information can
be found at: http://www.cs.uvic.ca/~gtzan