> > The 'fit lines' tool in Splat (from the 'Analysis' menu) is able to
> > fit various profiles to line features in spectra. It returns
> > parameters such as line center, line width and line strength. Sadly,
> > the Splat line fitting tool does not return uncertainties for any of
> > these parameters.
I would expect that this surprising omission is on Peter Draper's
SPLAT list.
> The ELF... commands in DIPSO return errors on the fitted parameters. See
>
> http://star-www.rl.ac.uk/star/dvi/sun50.htx/sun50.html
Gosh that old URL still works. The modern address is
http://www.starlink.ac.uk/star/docs/sun50.htx/sun50.html
> > Are there other packages/routines within Starlink which do give line
> > parameters _and_ their uncertainties? Would Figaro or Dipso be useful?
The SPECDRE FIT* commands buried in Figaro do calculate errors and in
fact allow for covariance, but they don't include the likes of
Lorentzian fits. They're more geared to spectral-cube analysis, where
the line centre and dispersions are the most important data to extract,
rather than equivalent widths.
Malcolm Currie
|