Getting the same output as crosspsd without MathScript?

[jeanlucpicard]

I have posted this question in the MathScript forum, but have not received any answer, hopefully someone here can help me. 

I have a program which works fine by using MathScript to calculate auto- and cross-spectra by using psd_welch and crosspsd. However, I would like to ditch MathScript, so I could make the installer smaller by not having to include the MathScript runtime installer as well.

I heavily use the overlap parameter, and I have managed to replicate psd_welch by using TSA Welch.vi as suggested here (image taken from linked topic).  (And by using a scale factor of the FFT-length.)

Is there a way to obtain same results as an overlapping crosspsd by using regular vi's? I am also looking for exact representations of coherence_ms function, if possible.

Thanks

[jeanlucpicard]

does anybody have a solution?

[Bob_Schor]

If you carefully read the earlier posts you reference, you'll notice that it seems that the "Gold Standard" (a.k.a. the "correct answer") is not clear -- should one "trust" MatLab/Mathscript or LabVIEW's function?

I suspect that not too many Forum readers are experts on the topic of Welch's method for getting Power Spectral Density, which is why, after two days, noone has ventured an opinion.  Are you an expert?  Do you know any?  Be patient, and if no one speaks up, you may have to "hit the books" (or "search the Web") ...

Bob Schor

P.S. -- I just did such a search.  This does not seem to be such a difficult thing to do yourself.  I could well believe that a MatLab and LabVIEW algorithm might give slightly different results, as there are numerous "free parameters" and Advanced Handwaving in these algorithms.  Recall that you are trying to estimate the Power Spectral Density, and the Welch method is nice because it can be a little more immune to noise than just taking an FFT ...