view examples/Automatic Music Transcription/SMALL_AMT_SPAMS_test.m @ 61:42fcbcfca132

(none)

%%  DICTIONARY LEARNING FOR AUTOMATIC MUSIC TRANSCRIPTION EXAMPLE 1
%
%   Centre for Digital Music, Queen Mary, University of London.
%   This file copyright 2010 Ivan Damnjanovic.
%
%   This program is free software; you can redistribute it and/or
%   modify it under the terms of the GNU General Public License as
%   published by the Free Software Foundation; either version 2 of the
%   License, or (at your option) any later version.  See the file
%   COPYING included with this distribution for more information.
%   
%   This file contains an example of how SMALLbox can be used to test diferent
%   dictionary learning techniques in Automatic Music Transcription problem.
%   It calls generateAMT_Learning_Problem that will let you to choose midi,
%   wave or mat file to be transcribe. If file is midi it will be first
%   converted to wave and original midi file will be used for comparison with
%   results of dictionary learning and reconstruction.
%   The function will generarte the Problem structure that is used to learn
%   Problem.p notes spectrograms from training set Problem.b using
%   dictionary learning technique defined in DL structure.
%
%%

clear;


% Defining Automatic Transcription of Piano tune as Dictionary Learning
% Problem

SMALL.Problem = generateAMT_Learning_Problem();
TPmax=0;
%%
for i=1:10
    %%
    %   Solving AMT problem using non-negative sparse coding with
    %   SPAMS online dictionary learning (Julien Mairal 2009)
    %
    
    %   Initialising Dictionary structure
    %   Setting Dictionary structure fields (toolbox, name, param, D and time)
    %   to zero values
    
    SMALL.DL(i)=SMALL_init_DL();
    
    %   Defining fields needed for dictionary learning
    
    SMALL.DL(i).toolbox = 'SPAMS';
    SMALL.DL(i).name = 'mexTrainDL';
    
    %   We test SPAMS for ten different values of parameter lambda
    %   Type 'help mexTrainDL in MATLAB prompt for explanation of parameters.
    
    lambda(i)=1.4+0.2*i;
    
    SMALL.DL(i).param=struct(...
        'K', SMALL.Problem.p,...
        'lambda', lambda(i),...
        'iter', 300,...
        'posAlpha', 1,...
        'posD', 1,...
        'whiten', 0,...
        'mode', 2);
    
    %   Learn the dictionary
    
    SMALL.DL(i) = SMALL_learn(SMALL.Problem, SMALL.DL(i));
    
    %   Set SMALL.Problem.A dictionary and reconstruction function
    %   (backward compatiblity with SPARCO: solver structure communicate
    %   only with Problem structure, ie no direct communication between DL and
    %   solver structures)
    
    SMALL.Problem.A = SMALL.DL(i).D;
    SMALL.Problem.reconstruct=@(x) SMALL_midiGenerate(x, SMALL.Problem);
    
    
    %%
    %   Initialising solver structure
    %   Setting solver structure fields (toolbox, name, param, solution,
    %   reconstructed and time) to zero values
    %   As an example, SPAMS (Julien Mairal 2009) implementation of LARS
    %   algorithm is used for representation of training set in the learned
    %   dictionary.
    
    SMALL.solver(1)=SMALL_init_solver;
    
    %   Defining the parameters needed for sparse representation
    
    SMALL.solver(1).toolbox='SPAMS';
    SMALL.solver(1).name='mexLasso';
    
    %   Here we use mexLasso mode=2, with lambda=3, lambda2=0 and positivity
    %   constrain (type 'help mexLasso' for more information about modes):
    %
    %   min_{alpha_i} (1/2)||x_i-Dalpha_i||_2^2 + lambda||alpha_i||_1 + (1/2)lambda2||alpha_i||_2^2
    
    SMALL.solver(1).param=struct(...
        'lambda', 3,...
        'pos', 1,...
        'mode', 2);
    
    %   Call SMALL_soolve to represent the signal in the given dictionary.
    %   As a final command SMALL_solve will call above defined reconstruction
    %   function to reconstruct the training set (Problem.b) in the learned
    %   dictionary (Problem.A)
    
    SMALL.solver(1)=SMALL_solve(SMALL.Problem, SMALL.solver(1));
    
    %%
    %   Analysis of the result of automatic music transcription. If groundtruth
    %   exists, we can compare transcribed notes and original and get usual
    %   True Positives, False Positives and False Negatives measures.
    
    AMT_res(i) = AMT_analysis(SMALL.Problem, SMALL.solver(1));
    if AMT_res(i).TP>TPmax
        TPmax=AMT_res(i).TP;
        BLmidi=SMALL.solver(1).reconstructed.midi;
        writemidi(SMALL.solver(1).reconstructed.midi, ['testL',i,'.mid']);
        max=i;
    end
end %end of for loop
%%
% Plot results and save midi files

figAMTbest=SMALL_AMT_plot(SMALL, AMT_res(max));

resFig=figure('Name', 'Automatic Music Transcription SPAMS lambda TEST');

subplot (3,1,1); plot(lambda(:), [AMT_res(:).TP], 'ro-');
title('True Positives vs lambda');

subplot (3,1,2); plot(lambda(:), [AMT_res(:).FN], 'ro-');
title('False Negatives vs lambda');

subplot (3,1,3); plot(lambda(:), [AMT_res(:).FP], 'ro-');
title('False Positives vs lambda');

FS=filesep;
[pathstr1, name, ext, versn] = fileparts(which('SMALLboxSetup.m'));
cd([pathstr1,FS,'results']);
[filename,pathname] = uiputfile({' *.mid;' },'Save midi');
if filename~=0 writemidi(BLmidi, [pathname,FS,filename]);end
[filename,pathname] = uiputfile({' *.fig;' },'Save figure TP/FN/FP vs lambda');
if filename~=0 saveas(resFig, [pathname,FS,filename]);end

[filename,pathname] = uiputfile({' *.fig;' },'Save BEST AMT figure');
if filename~=0 saveas(figAMTbest, [pathname,FS,filename]);end