README for artist20/baseline This is a set of Matlab routines that implement a baseline artist identification system, designed to work with the artist20 twenty-artist pop music data set distributed by LabROSA. For more details, see: http://labrosa.ee.columbia.edu/projects/artistid/ The main routine is trntest_folds.m, which takes one file containing definitions of set of train/test folds. The default example is 6fold.list, in this directory. This then calls the routine trntest_1fold.m, passing it each of the train/test set definition files read from the master file, and accumulating the confusion matrix results. trntest_1fold.m reads the training and test set definition files, trains models for each class using train_model.m, then reads each of the test tracks and evaluates it against all the models using test_track.m. train_models.m calls subfunction model_train_data.m, and test_track.m calls model_match.m, both of which actually employ Kevin Murphy's GMM code (which itself calls Ian Nabney's netlab). If you don't have it, get the whole lot from: http://www.cs.ubc.ca/~murphyk/Software/HMM/hmm.html We also include readhtk.m from Mike Brookes's VOICEBOX toolkit. The full toolkit is available at: http://www.ee.ic.ac.uk/hp/staff/dmb/voicebox/voicebox.html .. but you don't need to download it to run this code. You can reference the use of this data with the following paper: "CLASSIFYING MUSIC AUDIO WITH TIMBRAL AND CHROMA FEATURES", Daniel P. W. Ellis, Proc. Int. Conf. on Music Information Retrieval ISMIR-07, Vienna, Austria, Sep. 2007. % Copyright (c) 2007 Columbia University. % % This file is part of LabROSA-artist20-baseline % % artist20-baseline is free software; you can redistribute it and/or modify % it under the terms of the GNU General Public License version 2 as % published by the Free Software Foundation. % % artist20-baseline is distributed in the hope that it will be useful, but % WITHOUT ANY WARRANTY; without even the implied warranty of % MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU % General Public License for more details. % % You should have received a copy of the GNU General Public License % along with artist20-baseline; if not, write to the Free Software % Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA % 02110-1301 USA % % See the file "COPYING" for the text of the license. -- 2007-07-05 Dan Ellis dpwe@ee.columbia.edu