Source-Filter Model - Sound Software .ac.uk

Overview

MATLAB code implementing a source-filter model based on the beta-divergence. Details about the method can be found a supplementary technical report (see 'Publications' tab).

The repository contains a MATLAB class 'CSourceFilter' implementing the model. An example script 'sourceFilterExample' illustrates the use of the class and some supplementary files for the example script are provided.

There are no prerequisites for the class other than a basic MATLAB installation. The additional files 'dceps' and 'idceps' implement the 'discrete cepstrum spectrum envelope' from D. Schwarz' thesis (see reference in ICASSP paper). These are only required for the estimation of the smoothed version of the estimated filter curve.

Related publications

H. Kirchhoff, S. Dixon, and A. Klapuri, Missing template estimation for user-assisted music transcription. Vancouver, Canada, 2013.: [More Details] [BIBT_EX]; @proceedings{kirchhoff2013a, author = {Kirchhoff, Holger and Dixon, Simon and Klapuri, Anssi}, title = {Missing template estimation for user-assisted music transcription}, address = {Vancouver, Canada}, booktitle = {International Conference on Acoustics, Speech, and Signal Processing (ICASSP)}, month = {May}, note = {submitted}, year = {2013} }
H. Kirchhoff, S. Dixon, and A. Klapuri, “Derivation of update equations for a source-filter model based on beta-divergence,” Queen Mary University of London, C4DM-TR-10-12, Nov. 2012.: [More Details] [BIBT_EX] [URL (ext.)]; @techreport{kirchhoff2012a, author = {Kirchhoff, Holger and Dixon, Simon and Klapuri, Anssi}, title = {Derivation of update equations for a source-filter model based on beta-divergence}, address = {Queen Mary University of London}, month = {Nov}, number = {C4DM-TR-10-12}, year = {2012} }

Members

Manager: Holger Kirchhoff