philip_pj Offline Upload & Sell: On
|
Every kind of factor comes into the answer to this simple request, Svenning, and testing is a fine idea. People making huge prints and looking for highest IQ will tell you f8 is as small as you can go. But, effectively, it depends on your aims, your repro size, and your need for depth of field.
Many more images are lost to inadequate dof than to diffraction. If you stick with f8 and end up with slightly sharper data in the zone of reasonable dof but lose this quality from not using f11, clearly that is not optimal.
Lenses vary widely. Taking 50mm recent releases, two high end 55mm lenses, Sony's FE55 and Zeiss's Otus 55/1.4; both these decline in IQ from f8 to f11 right across the frame. By contrast, Nikon's 58/1.4 loses a little centre IQ but picks up significantly from f8 to f11. As this lens is very deficient in corner definition, the user may want to take this trade-off. And at f11 the Sony lens is still far better in the corners at f11 than the Nikon despite its f11 IQ being worse than its f8.
The general 'rule' is that best overall IQ occurs at two stops down from wide open, but this is very often very wrong. That Sony FE55 is best at f5.6, 3.3 stops down from wide open. Zeiss's 55mm Otus is best at f4, 3 full stops down.
Then preferences we all have come into it. I shoot a lot of images with focus at the middle to edge of the frame, so I use lenses with strong outer frames at relatively wide apertures, for the low light shooting I do demands that from me. Huge numbers of lenses are really only good in the center of the image, typically described as 'portrait lenses'. Myself I find this problematic, dull and boring. Leica have made an art form of producing lenses that are stunning in the 'magic donut' centre but don't even think of using them for landscapes! So you see - it all depends.
|