Pseudo-Voigt FWHM of Gaussian and Lorentian

[Kale Letterman]

Hi!

Thanks for the package it's working very well!

I was wondering if there is a method of, once I've fitted a PV and gotten the fraction (alpha) and FWHM to calculate the respective FWHM for the Gaussian portion and the Lorentian portion of the PV fit?

Thanks again!

[Matt Newville]

For a pseudo-Voigt (but also: why use that instead of a real Voigt?), the `sigma` is used for the Lorentzian component, so its FWHM will be 2*sigma.    The Gaussian component uses a sigma_gauss = sigma/sqrt(2*log(2)) so that it will also have the same FWHM.    That is, the two components will have the same FWHM.  This would not be true, in general for a real Voigt function, which better expresses the convolution of a Gaussian and Lorentzian function.