Beta 1

Title Instrument Detection in Music
Author Hjorth-Jensen, Frederik
Supervisor Larsen, Jan (Intelligent Signal Processing, Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
Institution Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
Thesis level Master's thesis
Year 2008
Abstract In this thesis the field of Musical Information retrieval was introduced in order to provide the setting for the main subject of the thesis: ´´Instrument Detection in Music”. A spectrogram alike quantity was developed using a Constant Q Transform, to facilitate the investigation of harmonic content of music signals, and to form the basis components used in one of two methods being developed. A classification framework consisting of two methods, namely Sparse Nonnegative Matrix Factor 2 dimensional Deconvolution and reduced Kernel Partial Least Squares, were implemented and applied to Instruments detection. Samples of a philharmonic orchestra were used to construct several test sets, produced in a software sequencer. Experiments showed that the proposed methods are suitable for the task of Instrument Detection in Music.
Original PDF ep08_25NEWweb.pdf (2.33 MB)
Admin Creation date: 2008-03-10    Update date: 2008-07-15    Source: dtu    ID: 211621    Original MXD