wolffd@0: function varargout = mirbeatspectrum(orig,varargin)
wolffd@0: %   n = mirbeatspectrum(m) evaluates the beat spectrum.
wolffd@0: %   [n,m] = mirbeatspectrum(m) also return the similarity matrix on which
wolffd@0: %       the estimation is made.
wolffd@0: %   Optional argument: 
wolffd@0: %       mirbeatspectrum(...,s) specifies the estimation method.
wolffd@0: %           Possible values:
wolffd@0: %           s = 'Diag', summing simply along the diagonals of the matrix.
wolffd@0: %           s = 'Autocor', based on the autocorrelation of the matrix.
wolffd@0: %       mirbeatspectrum(...,'Distance',f) specifies the name of a dissimilarity
wolffd@0: %           distance function, from those proposed in the Statistics Toolbox
wolffd@0: %               (help pdist).
wolffd@0: %           default value: f = 'cosine'
wolffd@0: %   J. Foote, M. Cooper, U. Nam, "Audio Retrieval by Rhythmic Similarity",
wolffd@0: %   ISMIR 2002.
wolffd@0: 
wolffd@0: 
wolffd@0:         dist.key = 'Distance';
wolffd@0:         dist.type = 'String';
wolffd@0:         dist.default = 'cosine';
wolffd@0:     option.dist = dist;
wolffd@0: 
wolffd@0:         meth.type = 'String';
wolffd@0:         meth.choice = {'Diag','Autocor'};
wolffd@0:         meth.default = 'Autocor';
wolffd@0:     option.meth = meth;
wolffd@0: 
wolffd@0: specif.option = option;
wolffd@0: varargout = mirfunction(@mirbeatspectrum,orig,varargin,nargout,specif,@init,@main);
wolffd@0:     
wolffd@0: 
wolffd@0: function [x type] = init(x,option)
wolffd@0: if not(isamir(x,'mirscalar'))
wolffd@0:     if isamir(x,'miraudio')
wolffd@0:         x = mirmfcc(x,'frame',.025,'s',.01,'s','Rank',8:30);
wolffd@0:     end
wolffd@0:     x = mirsimatrix(x,'Distance',option.dist,'Similarity');
wolffd@0: end
wolffd@0: type = 'mirscalar';
wolffd@0: 
wolffd@0: 
wolffd@0: function y = main(orig,option,postoption)
wolffd@0: if iscell(orig)
wolffd@0:     orig = orig{1};
wolffd@0: end
wolffd@0: fp = get(orig,'FramePos');
wolffd@0: if not(isa(orig,'mirscalar'))
wolffd@0:     s = get(orig,'Data');
wolffd@0:     total = cell(1,length(s));
wolffd@0:     for k = 1:length(s)
wolffd@0:         for h = 1:length(s{k})
wolffd@0:             maxfp = find(fp{k}{h}(2,:)>4,1);
wolffd@0:             if isempty(maxfp)
wolffd@0:                 maxfp = Inf;
wolffd@0:             else
wolffd@0:                 fp{k}{h}(:,maxfp+1:end) = [];
wolffd@0:             end
wolffd@0:             l = min(length(s{k}{h}),maxfp);
wolffd@0:             total{k}{h} = zeros(1,l);
wolffd@0:             if strcmpi(option.meth,'Diag')
wolffd@0:                 for i = 1:l
wolffd@0:                     total{k}{h}(i) = mean(diag(s{k}{h},i-1));
wolffd@0:                 end
wolffd@0:             else
wolffd@0:                 for i = 1:l
wolffd@0:                     total{k}{h}(i) = mean(mean(s{k}{h}(:,1:l-i+1).*s{k}{h}(:,i:l)));
wolffd@0:                 end
wolffd@0:             end
wolffd@0:         end
wolffd@0:     end
wolffd@0: else
wolffd@0:     total = get(orig,'Data');
wolffd@0: end
wolffd@0: n = mirscalar(orig,'Data',total,'FramePos',fp,'Title','Beat Spectrum'); 
wolffd@0: y = {n orig};