Mercurial > hg > smallbox
comparison Problems/generateAudioDeclippingProblem.m @ 161:f42aa8bcb82f ivand_dev
debug and clean the SMALLbox Problems code
author | Ivan Damnjanovic lnx <ivan.damnjanovic@eecs.qmul.ac.uk> |
---|---|
date | Wed, 31 Aug 2011 12:02:19 +0100 |
parents | b14209313ba4 |
children | 9c418bea7f6a |
comparison
equal
deleted
inserted
replaced
155:b14209313ba4 | 161:f42aa8bcb82f |
---|---|
2 %% Generate Audio Declipping Problem | 2 %% Generate Audio Declipping Problem |
3 % | 3 % |
4 % generateAudioDeclippingProblem is part of the SMALLbox [1] and generates | 4 % generateAudioDeclippingProblem is part of the SMALLbox [1] and generates |
5 % Audio declipping is a problem proposed in Audio Inpaining Toolbox and | 5 % Audio declipping is a problem proposed in Audio Inpaining Toolbox and |
6 % in [2]. | 6 % in [2]. |
7 % | |
8 % The function takes as an optional input | |
9 % soundfile - name of the file | |
10 % clippingLevel - (default 0.6) | |
11 % windowSize - 1D frame size (eg 512) | |
12 % overlap - ammount of overlaping frames between 0 and 1 | |
13 % wa,ws,wd - analisys, synthesis and dictionary window functions | |
14 % | |
15 % Dict_fun - function to be used to generate dictionary | |
16 % redundancyFactor - overcompletness of dictionary (default 2) | |
17 % | |
18 % The function outputs the structure with following fields: | |
19 % original - original signal | |
20 % clipped - clipped signal | |
21 % clipMask - mask indicating clipped samples | |
22 % clippingLevel - (default 0.6) | |
23 % Upper_Limit - maximum value of original data | |
24 % fs - sample rate of the original signal in Hertz | |
25 % nbits - the number of bits per sample | |
26 % sigma - added noise level | |
27 % B - dictionary to be used for sparse representation | |
28 % M - measurement matrix (non-clipped data in b) | |
29 % b - matrix of clipped frames | |
30 % m - size od dictionary atom | |
31 % n - number of frames to be represented | |
32 % p - number of atoms in dictionary | |
33 % windowSize - 1D frame size (eg 512) | |
34 % overlap - ammount of overlaping frames between 0 and 1 | |
35 % wa,ws, wd - analisys, synthesis and dictionary window functions | |
7 % | 36 % |
8 % [1] I. Damnjanovic, M. E. P. Davies, and M. P. Plumbley "SMALLbox - an | 37 % [1] I. Damnjanovic, M. E. P. Davies, and M. P. Plumbley "SMALLbox - an |
9 % evaluation framework for sparse representations and dictionary | 38 % evaluation framework for sparse representations and dictionary |
10 % learning algorithms," V. Vigneron et al. (Eds.): LVA/ICA 2010, | 39 % learning algorithms," V. Vigneron et al. (Eds.): LVA/ICA 2010, |
11 % Springer-Verlag, Berlin, Germany, LNCS 6365, pp. 418-425 | 40 % Springer-Verlag, Berlin, Germany, LNCS 6365, pp. 418-425 |
101 data.wa = wa; | 130 data.wa = wa; |
102 data.wd = wd; | 131 data.wd = wd; |
103 | 132 |
104 data.fs = x.fs; | 133 data.fs = x.fs; |
105 data.nbits = x.nbits; | 134 data.nbits = x.nbits; |
106 data.Upper_Limit = max(solutiondata.XClean); | 135 data.Upper_Limit = max(solutionData.xClean); |
107 [data.m, data.n] = size(x_clip); | 136 [data.m, data.n] = size(x_clip); |
108 data.p = windowSize*redundancyFactor; %number of dictionary elements | 137 data.p = windowSize*redundancyFactor; %number of dictionary elements |
109 | 138 |
110 cd(TMPpath); | 139 cd(TMPpath); |
111 | 140 |