Mercurial > hg > ltpda

comparison m-toolbox/classes/@ao/fngen.m @ 0:f0afece42f48

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Import.
author Daniele Nicolodi <nicolodi@science.unitn.it>
date Wed, 23 Nov 2011 19:22:13 +0100
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:f0afece42f48
1 % FNGEN creates an arbitrarily long time-series based on the input PSD.
2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3 %
4 % DESCRIPTION: FNGEN creates an arbitrarily long time-series based on the input PSD.
5 %
6 % CALL: b = fngen(axx, pl)
7 %
8 % PARAMETERS:
9 % 'Nsecs' - The number of seconds to produce
10 % [default: inverse of PSD length]
11 % 'Win' - The spectral window to use for blending segments
12 % [default: Kaiser -150dB]
13 %
14 %
15 % NOTE: this function requires the Statistics Toolbox in order to create
16 % a chi^2 distributed random variable.
17 %
18 % <a href="matlab:utils.helper.displayMethodInfo('ao', 'fngen')">Parameters Description</a>
19 %
20 % VERSION: $Id: fngen.m,v 1.33 2011/04/08 08:56:16 hewitson Exp $
21 %
22 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23
24 function varargout = fngen(varargin)
25
26 % Check if this is a call for parameters
27 if utils.helper.isinfocall(varargin{:})
28 varargout{1} = getInfo(varargin{3});
29 return
30 end
31
32 import utils.const.*
33 utils.helper.msg(msg.PROC3, 'running %s/%s', mfilename('class'), mfilename);
34
35 % Collect input variable names
36 in_names = cell(size(varargin));
37 for ii = 1:nargin,in_names{ii} = inputname(ii);end
38
39 % Collect all AOs and plists
40 [as, ao_invars] = utils.helper.collect_objects(varargin(:), 'ao', in_names);
41 pl = utils.helper.collect_objects(varargin(:), 'plist', in_names);
42
43 if nargout == 0
44 error('### fngen cannot be used as a modifier. Please give an output variable.');
45 end
46
47 % combine plists
48 pl = parse(pl, getDefaultPlist());
49
50 % Extract necessary parameters
51 Nsecs = find(pl, 'Nsecs');
52 swin = find(pl, 'win');
53
54 % Loop over input AOs
55 bs = [];
56 for j=1:numel(as)
57 if ~isa(as(j).data, 'fsdata')
58 warning('!!! %s expects ao/fsdata objects. Skipping AO %s', mfilename, as(j).name);
59 else
60 % Properties of the input PSD
61 N = 2*(length(as(j).data.y)-1);
62 fs = as(j).data.x(end)*2;
63 % Extract Fourier components
64 Ak = sqrt(N*as(j).data.getY*fs);
65 Ak = [Ak; Ak(end-1:-1:2)]; % make two-sided
66 % redesign input window for this length
67 switch lower(swin.type)
68 case 'kaiser'
69 swin = specwin('Kaiser', N, swin.psll);
70 otherwise
71 swin = specwin(swin.type, N);
72 end
73 % Compute time-series segments
74 Olap = 1-swin.rov/100;
75 win = [swin.win].';
76 segLen = N/fs;
77 if segLen > Nsecs
78 cNsecs = 2*segLen;
79 else
80 cNsecs = Nsecs;
81 end
82 Nsegs = 1+floor(cNsecs/segLen/Olap);
83
84 % Prepare for generation
85 rphi = zeros(N,1); % Empty vector for random phases
86 xs = zeros(fs*(cNsecs+segLen), 1); % Large empty vector for new time-series
87 e1 = 1; e2 = segLen*fs; % Indices into large vector
88 step = round(segLen*fs*Olap); % step size between each new segment
89 lxs = length(xs);
90
91 % Loop over segments
92 for s=1:Nsegs
93 % Generate random phase vector
94 rphi(2:N/2) = pi*rand(1,N/2-1); % First half
95 rphi(N/2+1) = pi*round(rand); % mid point
96 rphi(N/2+2:N) = -rphi(N/2:-1:2); % reflected half
97 %---- Compute Fourier amplitudes
98 % Use chi^2 distribution to randomize amplitudes.
99 % - from Percival and Walden: S_est = S.*chi2rnd(2)/2
100 % so A_est = A.*sqrt(chi2rnd(2)/2)
101 % Here we take the measured input data to be a good estimate of
102 % the underlying power spectrum
103 X = (Ak.*sqrt(chi2rnd(2)/2)) .*exp(1i.*rphi);
104 % Inverse FFT
105 x = ifft(X, 'symmetric');
106 % overlap the segments
107 xs(e1:e2) = xs(e1:e2) + win.*x;
108 % increase step
109 e1 = e1 + step;
110 e2 = e2 + step;
111 if e2>lxs
112 break
113 end
114 end
115 % Make ao from the segment of data we want
116 e1 = fs*segLen/2;
117 e2 = fs*(Nsecs+segLen/2)-1;
118 b = ao(tsdata(xs(e1:e2).', fs));
119 b.name = sprintf('fngen(%s)', ao_invars{j});
120 b.data.setXunits('s');
121 % Add history
122 b.addHistory(getInfo('None'), pl, ao_invars(j), as(j).hist);
123 % Add to outputs
124 bs = [bs b];
125 end
126 end
127
128 % Set output
129 if nargout == numel(bs)
130 % List of outputs
131 for ii = 1:numel(bs)
132 varargout{ii} = bs(ii);
133 end
134 else
135 % Single output
136 varargout{1} = bs;
137 end
138
139 end
140
141 %--------------------------------------------------------------------------
142 % Get Info Object
143 %--------------------------------------------------------------------------
144 function ii = getInfo(varargin)
145 if nargin == 1 && strcmpi(varargin{1}, 'None')
146 sets = {};
147 pl = [];
148 else
149 sets = {'Default'};
150 pl = getDefaultPlist;
151 end
152 % Build info object
153 ii = minfo(mfilename, 'ao', 'ltpda', utils.const.categories.sigproc, '$Id: fngen.m,v 1.33 2011/04/08 08:56:16 hewitson Exp $', sets, pl);
154 ii.setModifier(false);
155 end
156
157 %--------------------------------------------------------------------------
158 % Get Default Plist
159 %--------------------------------------------------------------------------
160 function plout = getDefaultPlist()
161 persistent pl;
162 if exist('pl', 'var')==0 || isempty(pl)
163 pl = buildplist();
164 end
165 plout = pl;
166 end
167
168 function pl = buildplist()
169
170 pl = plist();
171
172 % Win
173 p = param({'Win', 'The spectral window to use for blending data segments.'}, paramValue.WINDOW);
174 pl.append(p);
175
176 % Nsecs
177 p = param({'Nsecs', 'The number of seconds of data to produce.'}, paramValue.EMPTY_DOUBLE);
178 pl.append(p);
179
180 end
181