[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Finding PSD Mean Frequency

Mark McGillion <mm@fs1.co.umist.ac.uk> writes:
> I have a PSD and would like to find the mean frequency. The goal of this
> is to fit a gaussian curve to the PSD to represent the mean and variance
> of the power in the given frequency range. I am making the assumption
> that the frequency power is normally distributed and can be modelled by
> 1 or more gaussians. The frequency range is 0-1000Hz.

Procedures you will want to look at are

 GAUSSFIT - fits a gaussian to a series
 CURVEFIT - fits an arbitrary IDL function to a series
 MPFIT, MPFITFUN - fits an arbitrary IDL function to a series

GAUSSFIT is the least general, but may be enough for your purposes.
If you decide that your function needs to be more complicated, for
example gaussian plus a constant, then you should try one of the
latter solutions.  CURVEFIT is built into IDL and has on-line

MPFIT, MPFITFUN and MPFITEXPR are available from my web page,
http://astrog.physics.wisc.edu/~craigm/idl/idl.html.  They are IDL
versions of the MINPACK robust least squares fitting routines.  For
noisy data I have found them to be more reliable than CURVEFIT.  The
headers of each have complete documentation, and a tutorial example of
fitting a guassian is on the web page.

Good luck,
Craig B. Markwardt, Ph.D.         EMAIL: craigmnet@astrog.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response