[Ifeffit] mxan and gnxas
mamarcus at lbl.gov
Sat Jul 29 15:54:15 CDT 2006
Thanks. Do you mean that it's available as source-code only? Is there any
tutorial material available? If it does background subtraction, does that
you feed it raw signal-vs-energy files?
----- Original Message -----
From: "Francesco Giannici" <zenarcade at libero.it>
To: "ifeffit" <ifeffit at millenia.cars.aps.anl.gov>
Sent: Saturday, July 29, 2006 11:19 AM
Subject: [Ifeffit] mxan and gnxas
> 1. how to get mxan: as far as i know, mxan is available free of charge,
> although not ready for download. you should directly contact maurizio dot
> benfatto @ lnf dot infn dot it. caveat: the documentation is *very* short,
> and there is no interface, so here's another reason to keep in touch with
> the developer.
> mxan means "minuit xanes", so it uses the minuit minimization procedures
> to fit either structural (i.e. atomic positions) or physical (i.e.
> broadening, plasmon, etc.) parameters. it seems the best choice if you
> deal with a few atoms (biomolecules, coordination compounds, and such).
> 2. my two cents on gnxas: this package is either very powerful, and very
> scary. you have to be always aware of what you're doing, because it's no
> black box. but once you manage to learn the basics, you are awarded with
> wonderful quality and stability of the fit (minuit is used to avoid false
> minima during minimization), and total control of the background
> subtraction (spline and multi-electron excitation refined simultaneously
> during the fit). other useful features are asymmetric distributions, and
> the automatic inclusion of all ms paths in the n-body signal generation.
> the cons: there's no fourier filtering (the fit is done on chi(E)), there
> are a few bugs in the potential calculation that prevent a smooth
> generation of signals above 7 angstrom, and you have to spend a lot of
> time learning the various modules of the package.
> hope this can help,
> francesco giannici
> dept. of inorganic and analytical chemistry
> università di palermo
> giannici at pa.ismn.cnr.it
> Ifeffit mailing list
> Ifeffit at millenia.cars.aps.anl.gov
More information about the Ifeffit