Replace fourier filter in POLARIS total scattering reduction #40647
Description
Is your feature request related to a problem? Please describe.
Even after background subtraction in issue #40646 - there remains artefacts in low-r region of pdf.
To deal with this (in the same spirit as Gudrun - though implementation differs as Gudrun uses a top hat convolution to subtract Q-space background) it is necessary to find a background in Q that reproduces the low-r region of the pdf and subtract this from S(Q) - 1.
To construct the ‘error’ term in g(r) I use a raised cosine window which goes to 0 at rmin, with optional flat portion until r0 (can set r0=0). This should not produce any ringing/artefacts (as long as rmin - r0 >~ 3*pi/Qmax).
Which produces a very flat g(r) at low r - regardless of background chosen.
