>>2565661The example shown is probably one of those 'perfect conditions', they probably knew which pixels shifted. In real life applications you won't know which pixels shift and the results won't come close to those of that image, though the results can be pretty good as seen in
>>2564934.
>>2565799If you don't see a clear difference you must be doing something wrong. PhotoAcute example, upper right shows the normal 9000F flatbed result, upper right shows 2 scans in PhotoAcute, left 4, and right 6. Note that you might have to experiment a bit with PA to get the right profiles to avoid excessive noise and oversharpening.
