Polyphonic instrument identification using independent subspace analysis

A system which tries to identify the musical instruments playing concurrently in a mixture is investigated in this paper. The features used in classification are derived from independent subspace analysis (ISA) which somewhat decomposes each source, and the mixture, into its statistically "independent" components. Without re-grouping or actually separating the sources, they offer physiologically-motivated classification of instruments, assuming the decomposition is robust to the mixing process. The system is evaluated on two-tonal instrument mixtures from a set of five instruments and a phrase of a real song from CD.