dspperiododgram for power spectral density

Share icon

Fri, 07/31/2015 - 01:38 pm

Hi,

So far, I used the no-longer-updated igor procedures available through #include <power spectral density> to estimate power spectral densities. As some of my phase data has random walk noise, I modified the PSD code to remove both DC offset and slope of each segment before multiplying with the window function and calculating the FFT.
Trying to use newer code, I looked at dspperiodogram. While I found the /NODC option to remove the DC part, I couldn't find a way to remove the slope of each segment; this procedure is recommended e.g. in [1].
Is there a way to remove the slope of individual segments with dspperiodogram?

Thanks,
Esther

[1] Heinzel, Gerhard, Albrecht Rüdiger, and Roland Schilling. “Spectrum and Spectral Density Estimation by the Discrete Fourier Transform (DFT), Including a Comprehensive List of Window Functions and Some New at-Top Windows,” 2002. http://pubman.mpdl.mpg.de/pubman/faces/viewItemOverviewPage.jsp?itemId=….
Hello Esther,

There is no built-in option to remove the individual slopes of time-series segments. If you really want to implement that you might want to take a look at DSPDetrend operation that allows you to remove any polynomial (not just a line) from the data. You would have to apply this to individual segments which suggests that you may be better off running the old DSP procedures and adding your choice of de-trending for the segments.

A.G.
WaveMetrics, Inc.

July 31, 2015 at 02:07 pm - Permalink