


Helper function to calculate tapers and, if precalculated tapers are supplied,
to check that they (the precalculated tapers) the same length in time as
the time series being studied. The length of the time series is specified
as the second input argument N. Thus if precalculated tapers have
dimensions [N1 K], we require that N1=N.
Usage: tapers=dpsschk(tapers,N,Fs)
Inputs:
tapers (tapers in the form of:
(i) precalculated tapers or,
(ii) [NW K] - time-bandwidth product, number of tapers)
N (number of samples)
Fs (sampling frequency - this is required for nomalization of
tapers: we need tapers to be such
that integral of the square of each taper equals 1
dpss computes tapers such that the
SUM of squares equals 1 - so we need
to multiply the dpss computed tapers
by sqrt(Fs) to get the right
normalization)
Outputs:
tapers (calculated or precalculated tapers)
eigs (eigenvalues)

0001 function [tapers,eigs]=dpsschk(tapers,N,Fs) 0002 % Helper function to calculate tapers and, if precalculated tapers are supplied, 0003 % to check that they (the precalculated tapers) the same length in time as 0004 % the time series being studied. The length of the time series is specified 0005 % as the second input argument N. Thus if precalculated tapers have 0006 % dimensions [N1 K], we require that N1=N. 0007 % Usage: tapers=dpsschk(tapers,N,Fs) 0008 % Inputs: 0009 % tapers (tapers in the form of: 0010 % (i) precalculated tapers or, 0011 % (ii) [NW K] - time-bandwidth product, number of tapers) 0012 % 0013 % N (number of samples) 0014 % Fs (sampling frequency - this is required for nomalization of 0015 % tapers: we need tapers to be such 0016 % that integral of the square of each taper equals 1 0017 % dpss computes tapers such that the 0018 % SUM of squares equals 1 - so we need 0019 % to multiply the dpss computed tapers 0020 % by sqrt(Fs) to get the right 0021 % normalization) 0022 % Outputs: 0023 % tapers (calculated or precalculated tapers) 0024 % eigs (eigenvalues) 0025 if nargin < 3; error('Need all arguments'); end 0026 sz=size(tapers); 0027 if sz(1)==1 && sz(2)==2; 0028 [tapers,eigs]=dpss(N,tapers(1),tapers(2)); 0029 tapers = tapers*sqrt(Fs); 0030 elseif N~=sz(1); 0031 error('seems to be an error in your dpss calculation; the number of time points is different from the length of the tapers'); 0032 end;