-
Notifications
You must be signed in to change notification settings - Fork 0
/
time_stretch.m
183 lines (151 loc) · 5.44 KB
/
time_stretch.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
function [result, locs, costs, min_in_step] = time_stretch(sample, suspect, sample_no, suspect_no)
[data_orig,fs] = audioread(sample);
[data_copy,fs] = audioread(suspect);
% data_orig = bsxfun(@rdivide, data_orig, rms(data_orig,1));
% data_copy = bsxfun(@rdivide, data_copy, rms(data_copy,1));
data_orig = downsample(data_orig,2);
data_copy = downsample(data_copy,2);
fs = fs/2;
%% Computing the STFT spectrograms
data_orig = mean(data_orig,2);
data_copy = mean(data_copy,2);
window = 4096;
hop = 1024;
[Xo, freq, time] = spectrogram(data_orig, window, window-hop);
Xs = spectrogram(data_copy, window, window-hop);
clear data_orig data_copy
f = 0:(length(freq)-1);
f = f*((fs/2)/length(freq));
midi = 69 + 12*log2(f/440);
%% Performing Non-negative Matrix Factorization on the original sample spectrogram
n = 0; % no pitch shifts
k = 10;
[Bo, Ho, r] = nnmf(abs(Xo), k);
% Check for low-rank < k
rank_check = sum(Ho,2);
if ~isempty(find(rank_check == 0))
Ho(find(rank_check == 0),:) = [];
Bo(:,find(rank_check == 0)) = [];
k = k - numel(find(rank_check == 0));
end
%% Computing pitch shifted templates for detecting pitch shifted samples
% N = numel(Bo(:,1));
% Bo_concat = Bo;
%
% % n : number of semi-tone shifts.
% n = 12;
% for i=1:12
% t1 = 1:N;
% t2 = (t1)/(2^(-i/12));
% B_shift = interp1(t1, Bo, t2,'spline');
% % if t1(end)<t2(end)
% % B_shift(floor(t1(end)/t2(end)*t1(end)):end,:) = 0;
% % end
% Bo_concat = [Bo_concat B_shift];
% end
%
% clear B_shift Bo
%% Performing partially fixed NMF using the precomputed template matrix
[Bo1, Ho_hypo, ~, ~, err] = PfNmf(abs(Xs), Bo, [], [], [], 20, 0);
%% Performing partially fixed NMF using the pitch-shift templates concatenated as well.
% use either one of the PFNMF sections
% [~, Ho_hypo, ~, ~, err] = PfNmf(abs(Xs), Bo_concat, [], [], [], 0, 0);
%% Normalize the two activation matrices before computing correlation
% DO NOT RUN
for i = 1:k
Ho(i,:) = Ho(i,:)/norm(Ho(i,:),1);
end
for i = 1:(n+1)*k
Ho_hypo(i,:) = Ho_hypo(i,:)/norm(Ho_hypo(i,:),1);
end
%% time stretch detection
% tic;
D = pdist2(Ho', Ho_hypo','correlation');
costs = zeros(1,size(Ho_hypo,2)- floor(size(Ho,2)/4));
locs = zeros(1,size(Ho_hypo,2) - floor(size(Ho,2)/4));
% new_costs = zeros(1,size(Ho_hypo,2) - floor(size(Ho,2)/4));
costs_m = zeros(1,size(Ho_hypo,2)- floor(size(Ho,2)/4));
locs_m = zeros(1,size(Ho_hypo,2) - floor(size(Ho,2)/4));
% p = cell(size(Ho_hypo,2),1);
%MATLAB DTW call
tic
% for i=1:(size(Ho_hypo,2) - floor(size(Ho,2)/4));
% [a,c] = DTW(D(:,i:end));
% % p{i} = a;
% [costs_m(i), locs_m(i)] = min(c(end,:));
% locs_m(i) = locs_m(i) + i;
% costs_m(i) = costs_m(i)/size(a,1);
% end
% toc;
%C++ DTW call
distfile = 'C:/Users/SiddGururani/Desktop/MUSIC-8903-2016-exercise3_dtw/bin/release/distmat.bin';
outfile = 'C:/Users/SiddGururani/Desktop/MUSIC-8903-2016-exercise3_dtw/bin/release/output.bin';
executable = 'C:/Users/SiddGururani/Desktop/MUSIC-8903-2016-exercise3_dtw/bin/release/MUSI8903Exec.exe';
tic
fid = fopen(distfile, 'w');
% Row-wise writing instead of column-wise writing
fwrite(fid,D','float');
fclose(fid);
for i=1:(size(Ho_hypo,2) - floor(size(Ho,2)/4))
system([executable ' ' num2str(size(D,1)) ' ' num2str(size(D,2)) ' ' num2str(i-1) ' ' distfile ' ' outfile ' 1>NUL 2>NUL']);
fid = fopen(outfile,'r');
dtw_out = fread(fid, [3 1], 'float');
fclose(fid);
locs(i) = dtw_out(3) + i + 1;
costs(i) = dtw_out(2)/dtw_out(1);
[a,c] = DTW(D(:,i:end));
% p{i} = a;
[costs_m(i), locs_m(i)] = min(c(end,:));
locs_m(i) = locs_m(i) + i;
costs_m(i) = costs_m(i)/size(a,1);
end
toc;
% For testing purposes
% filename = ['test_', num2str(sample_no), '_', num2str(suspect_no),'.mat'];
% save(filename,'locs', 'costs');
% result = 1;
% locs = 1;
% load(filename);
% code for including the step length for cost computation
locs_diff = diff(locs);
locs_diff = [1 locs_diff];
steps = find(locs_diff~=0);
steps = [steps, numel(locs)];
length_steps = diff(steps);
length_steps(end) = length_steps(end)+1;%to compensate for last step
for i = 1:numel(costs)
for j = 2:numel(steps)
if(i < steps(j) && i >= steps(j-1))
new_costs(i) = costs(i)/length_steps(j-1);
end
end
end
new_costs(numel(costs)) = costs(end)/length_steps(end);
% Get minimum cost within each step
min_in_step = zeros(1, numel(steps));
locs_in_step = zeros(1, numel(steps));
for i = 1:numel(steps)-1
[min_in_step(i), locs_in_step(i)] = min(costs(steps(i):steps(i+1)));
locs_in_step(i) = locs_in_step(i) + steps(i);
end
min_in_step(end) = min(costs(steps(end):end));
locs_in_step(end) = locs_in_step(end) + steps(end);
mean_cost = mean(costs);
stdev = std(costs);
% Removing candidates whose value is greater than the mean cost
locs_in_step((min_in_step-mean_cost)>0) = [];
min_in_step((min_in_step-mean_cost)>0) = [];
% Removing candidates whose value isn't lower than 1.5 times the standard
% deviation of the costs.
locs_in_step((abs(min_in_step-mean_cost)-3*stdev)<0) = [];
min_in_step((abs(min_in_step-mean_cost)-3*stdev)<0) = [];
if(isempty(locs_in_step))
result = 0;
else
result = 1;
end
% figure; plot(costs);title(['costs for sample# ',num2str(sample_no),' and suspect# ',num2str(suspect_no)]);
% figure; plot(locs);title(['locations for sample# ',num2str(sample_no),' and suspect# ',num2str(suspect_no)]);
% figure; plot(new_costs);title(['normalized costs/length for sample# ',num2str(sample_no),' and suspect# ',num2str(suspect_no)]);
locs = locs_in_step;
end