Rick,
Thanks! So, there is no published reference yet, I guess?
Btw, as an aside, does not the use of the mixture of the Gaussian and mono-exponential create issues with using the 3d FFT? (This mixture can not be decomposed for each dimension.) Is the result very different (in terms of fitting) if a mixture of the Gaussian and monoexponential (with the same parameters) were fit in each dimension? This would also permit the use of the 3dFFT and decomposition of the correlation function.
Many thanks,
Ranjan