[Ifeffit] [HOWTO] Ifeffit f1f2() function

Hi to all, Is it correct to use as energy convolution width (in the f1f2 Ifeffit function) the natural widths of atomic levels tabulated by Krause & Oliver (J. Phys. Chem. Ref. Data, Vol. 8, No. 2, 1979)? My problem is to smooth (arctg-like) Cromer-Liberman f"(E) at the edge region [NOTE: I don't use feff8 because is too slow]. Thanks, Mauro -- Mauro Rovezzi PhD student - GILDA BM08 c/o ESRF (Grenoble, FR)