In comp.dsp Martin Brown <|||email@example.com> wrote:
(snip) > Anyone who says deconvolution isn't complicated demonstrably does not > have the first clue about the subject of signal processing. Convolution > is easy but deconvolution is usually a very difficult inverse problem.
Well, linear deconvolution isn't complicated, but it also often gives less than useful results. As usual, the non-linear case is more complicated than the linear case.
I have previously recommended "Deconvolution of Images and Spectra" by Jansson. Rumors are that it is now available in an affordable paperback reprint. Otherwise it is pretty expensive.