| yeah, Stroker's right, pure magic...
of course, you can make spectral analysis of the enlarged pic and push/boost/copy higher frequencies for sharpening; or, if the shot is coherent enough, you can try to extrapolate more data from subsequent pix [i think there are few papers on this on citeseer], but that's about it...
also, as tv signal is noisy as hell, the sharpening phase doesn't give any good results...
and exactly as Stroker says, you simply can't get data that are not there [don't forget that tv signal is in itself digital, thus loosing lots of information]... |