Anisotropic Lorentzian broadening

[stestoll]

pepper and chili should support an anisotropic Lorentzian broadening, akin to HStrain. It should be defined in the frequency domain, like HStrain. This new broadening and the Lorentzian part of lw and lwpp should be mutually exclusive. Naming unclear.

See #57 for chili.

Possible input syntax:

Sys.lwpp = [xG yG zG; xL yL zL];  % both Gaussian and Lorentzian anisotropic broadening

With this, the term "convolutional" broadening somewhat changes its meaning - it's not the powder spectrum that is convoluted by a lineshape, but each single-orientation spectrum separately.

Alternative: introduce two new fields:

Sys.lwppG = [xG yG zG];
Sys.lwppL = [xL yL zL];

This is a little less cryptic. Sys.lwpp could still be supported as the abbreviated input for isotropic broadening.