expfit vs harminv - which is faster/smaller ?

[Peter L. Søndergaard]

ons, 17 12 2008 kl. 07:04 -0800, skrev Sergei Steshenko:
> 
> 
> --- On Wed, 12/17/08, Peter L. Søndergaard <peter at sonderport.dk> wrote:
> 
> > From: Peter L. Søndergaard <peter at sonderport.dk>
> > Subject: Re: expfit vs harminv - which is faster/smaller ?
> > To: sergstesh at yahoo.com
> > Cc: "Francesco Potortì" <Potorti at isti.cnr.it>, help-octave at octave.org
> > Date: Wednesday, December 17, 2008, 12:57 AM
> > Another option that might be of interest is the NFFT
> > (non-equispaced
> > FFT) library from
> > 
> > http://www-user.tu-chemnitz.de/~potts/nfft/
> > 
> > Cheers,
> > 
> > Peter.
> > 
> 
> I think I've already visited this page, and I would use the algorithms for
> non-equispaced sampling.
> 
> However, I'm not sure it decomposes input sequence onto a sum of 
> exponentially decaying sinusoids.
> 
You are right, it decomposes the input into non-equidistantly spaced
complex exponentials. They do not decay.

Cheers,

Peter.