LTPDA Toolbox™ contents Using spectral windows   Power spectral density estimates

Spectral Estimation Methods

Linear and Log-scale Methods

The LTPDA Toolbox offers two kind of spectral estimators. The first ones are based on pwelch from MATLAB, which is an implementation of Welch's averaged, modified periodogram method [1]. More details about spectral estimation techniques can be found here.

The following pages describe the different Welch-based spectral estimation ao methods available in the LTPDA toolbox:

As an alternative, the LTPDA toolbox makes available the same set of estimators, based on an implementation of the LPSD algorithm [2]).

The following pages describe the different LPSD-based spectral estimation ao methods available in the LTPDA toolbox:

More detailed help on spectral estimation can also be found in the help associated with the Signal Processing Toolbox.

Computing the sample variance

The spectral estimators previously described usually return the average of the spectral estimator applied to different segments. This is a standard technique used in spectral analysis to reduce the variance of the estimator.

When using one of the previous methods in the LTPDA Toolbox, the value of this average over different segments is stored in the ao.y field of the output analysis object, but the user obtains also information about the spectral estimator variance in the ao.dy field.

The methods listed above store in the ao.dy field the standard deviation of the mean, defined as


For more details on how the variance of the mean is computed, please refer to the the help page of each method.

Note that when we only have one segment we can not evaluate the variance. This will happen in
  • linear estimators: when the number of averages is equal to one.
  • log-scale estimators: in the lowest frequency bins.


The following example compares the sample variance computed by ao/psd with two different segment length. 



% create white noise AO 
pl = plist('nsecs', 500, 'fs', 5, 'tsfcn', 'randn(size(t))');
a = ao(pl);

% compute psd with different Nfft
b1 = psd(a, plist('Nfft', 500));
b1.setName('Nfft = 500');
b2 = psd(a, plist('Nfft', 200));
b2.setName('Nfft = 200');

% plot with errorbars
iplot(b1,b2,plist('YErrU',{b1.dy,b2.dy}))


References

  1. P.D. Welch, The Use of Fast Fourier Transform for the Estimation of Power Spectra: A Method Based on Time Averaging Over Short, Modified Periodograms, IEEE Trans. on Audio and Electroacoustics, Vol. 15, No. 2 (1967), pp. 70 - 73
  2. M. Troebs, G. Heinzel, Improved spectrum estimation from digitized time series on a logarithmic frequency axis, Measurement, Vol. 39 (2006), pp. 120 - 129. See also the Corrigendum.


Using spectral windows  Using spectral windows   Power spectral density estimates Power spectral density estimates

©LTP Team