BLUES from Music: BLind Underdetermined Extraction of Sources from ...
Source: www2.imm.dtu.dk
Topic: Blues
Sort Desciption: (BLUES) to separate the instruments in a real music stereo recording. We combine two strong separation techniques to segregate instruments ...
Content Inside: BLUES from Music: BLind Underdetermined Extraction of Sources from Music Michael Syskind Pedersen, Tue Lehn-Schiøler, and Jan Larsen Intelligent Signal Processing, IMM, Technical University of Denmark ⋆ {msp,tls,jl}@imm.dtu.dk Abstract. In this paper we propose to use an instantaneous ICA method (BLUES) to separate the instruments in a real music stereo recording. We combine two strong separation techniques to segregate instruments from a mixture: ICA and binary time-frequency masking. By combining the methods, we are able to make use of the fact that the sources are differently distributed in both space, time and frequency. Our method is able to segregate an arbitrary number of instruments and the segregated sources are maintained as stereo signals. We have evaluated our method on real stereo recordings, and we can segregate instruments which are spatially different from other instruments. 1 Introduction Finding and separating the individual instruments from a song is of interest to the music community. Among the possible applications is a system where e.g. the guitar is removed from a song. The guitar can then be heard by a person trying to learn how to play. At a later stage the student can play the guitar track with the original recording. Also when transcribing music to get the written note sheets it is a great benefit to have the individual instruments in separate channels. Transcription can be of value both for musicians and for people wishing to compare (search in) music. On a less ambitious level identifying the instruments and finding the identity of the vocalist may aid in classifying the music and again make search in music possible. For all these applications, separation of music into its basic components is interesting. We find that the most important application of music separation is as a preprocessing step. Examples can be found where music consists of a single instrument only, and much of the literature on signal processing ...
blues from music blind underdetermined extraction of sources from music