ECG Database Applications Guide Table of Contents


lomb - estimate power spectrum using the Lomb periodogram method


lomb [ options ... ] input-file


lomb transforms a real-valued time series (from the specified input-file, or from the standard input if input-file is specified as ``-''; input-file must be in text form) into a power spectrum (on the standard output), using a technique known as the Lomb periodogram [Lomb, N.R., ``Least-squares frequency analysis of unequally spaced data'', Astrophysics and Space Science 39:447-462 (1976)].

The input is a text file containing a sampled time series, presented as two columns of numbers (the sample times and the sample values). The intervals between consecutive samples need not be uniform (in fact, this is the most significant advantage of the Lomb periodogram over other methods of power spectral density estimation). lomb writes the Lomb periodogram (the power spectral density estimate derived from the input time series) on the standard output, in two columns (frequency and power). If the units of the sample times in the input file are seconds, the units of the frequencies in the output are Hz.

Options are:

Print a usage summary.
Smooth the output.

The original version of this program was based on the algorithm described in Press, W.H, and Rybicki, G.B., ``Fast algorithm for spectral analysis of unevenly sampled data'', Astrophysical J. 338:277-280 (1989). It has been rewritten using the version of this algorithm presented in Press, W.H., Teukolsky, S.A., Vetterling, W.T., and Flannery, B.P., Numerical Recipes in C: the Art of Scientific Computing, pp. 575-584 (Cambridge Univ. Press, 1992).

Among many other applications, this program can be used to estimate heart rate power spectra, in combination with ihr (see Moody, G.B., ``Spectral analysis of heart rate without resampling'', Computers in Cardiology 1993, pp. 715-718 (IEEE Computer Society Press, 1993)). The Lomb method is ideal for analysis of any time series with missing or noisy data (the noisy data may be removed from the time series and need not be replaced, as would be necessary if conventional PSD estimation algorithms were employed).


fft(1) , hrfft(1) , memse(1)


This program is part of the psd package, distributed with the DB Software Package, version 9.3 and later.

Table of Contents