[Klaus]

[quote name='Class A' timestamp='1290803667' post='4481']

Thanks. It appears they are aiming at sub-pixel precision by exploiting blur distribution.



Do you know to what extent their approach depends on the assumption of adequate anti-aliasing, i.e., the presence of an AA filter?



Is it possible that feeding the Imatest software with images from a sensor with no AA filter might produce questionable results and/or what precautions have to be taken to make sure that this doesn't happen?

[/quote]





They offer test data for phone cams so I reckon there's no need for an AA filter (but phone cam optics serve as a AA filter for sure).



However, we've fed Leica M files into Imatest and the results are reasonable (unless you apply "normal" sharpening - the absolute values skyrocket in this case).