Markus M writes:
I have read just that moment your test with the D4 as regards the optimal aperture and diffraction. Again a highly useful and true life test, thank you very much for that!
To be sure, the D4 is not for me (cost, bulk), but according to your tests I would not have hesitations using apertures from f/8 to f/ 11 in landscape photography to get reasonably sharp pictures (foreground and background, I know there will be some limitations but it seems possible). With my D700, I do also use f/16 if necessary, though I normally stick to f/ 8 - f/ 11.
But now there is "my problem": I have also read your optimal aperture and diffraction test for the new D800. Also a great, eye-opening test!
From your mosaic pictures, f/ 4 is optimal, f/ 5,6 is still very good, but f/ 8 is already noticeably worse (to be sure no more optimal), and f/ 11 up is not worth talking about.
Well, this is a high resolution camera for landscapes. From my own experience, I can hardly shoot at apertures from f/2,8 - f/ 5,6 if I want depth of field. My Nikkor 24 f/ 1,4 will be fine at f/4, my 24-70 will "just" be fine at f/ 8. But f/ 4 - f/ 5,6 is a joke for landscape shooting.
Does that mean, if you want depth of field like "before the 36 mp era", that you have to use either focus stacking or tilt-shift-lenses? A "normal" wide angle prime lens like the Zeiss ZF 25 f/ 2 or the Nikon 24 f/ 1,4, as good as they are, won`t cut it. To be honest, I am not sure any more if I want this camera, and how you should shoot landscapes or cityscapes if you still want to have fun when doing that (and not just carrying lots of equipment, because with focus-stacking or tilt-shift-lenses you can`t shoot handheld reasonably). What is your impression about that?
DIGLLOYD: depth of field is an amorphous concept that takes into account the print size, the “acceptable” blur, etc. In reality, there is a narrow zone of sharp focus at all apertures, with increasing blur away from that zone. Depth of field is in a sense, an illusion created by the reproduction size of the image; make the image (print) small enough, and f/2 with a 100mm lens looks sharp on a 4 X 6" print.
The lens projects exactly the same image on a 36-megapixel D800 as it does on a 12-megapixel D700. Each camera’s image sensor samples that projected image at some resolution— imagine if a 100-megapixel sensor were used— same image, just higher sampling density. Some of those samples (photosites/pixels) are crisply rendered, and others are blurred. The higher the sampling resolution, the more individual details look blurred, on a per-pixel basis. But the total image is never less sharp because it is sampled at 36MP instead of 12MP. Looked at another way, it is the photosite size which influences the relative blur, as the D4 example shows relative to the D800 example.
Depth of field by stopping down makes the spot size that can be resolved smaller (more resolution), but diffraction steadily blurs the spot. At some point, there is a happy medium for any sensor resolution. For the D800, that point is arguably f/5.6, but really anything in the f/4 - f/11 range, depending on whether one wants peak sharpness and contrast in a narrower zone, or higher overall sharpness over a deeper zone. For a planar (flat) subject, that optimal aperture with a world-class lens is f/4, as shown in the example. For other lenses, it might be f/5.6 or f/8, since they aren’t very good to begin with.
Even on film this was an issue: using a 10X loupe on a fine grain slide film like Velvia, it was clear that f/32 lost resolution and contrast compared to f/22. Yet I often shot at f/32 to gain the extra depth of field (on my 4X5 view camera and 6X7 panoramic camera).
The question that arises is “how much sharp detail can be captured/sampled” for any given camera’s potential on a particular scene. We should not confuse “best possible pixel quality for every pixel” with “best quality for the image to be made”. For landscape photography, the D800 will make a better image at any aperture than the D3x or the D700, end of story.
As for tilt/shift lenses, they do allow the diffraction issue to be evaded with some suitable subjects (e.g. stopping down less), because the zone of focus can be aligned more closely to the shape of the 3D subject, thus making it less necessary to stop down so far.
Side note— your eyes do matter: there is one blogger out there telling people that the D800 vs the D800E doesn’t matter, because it’s only “acuity” not actual resolution (acuity, really accutance, is edge contrast, and edge contrast is MTF). This is a misleading (and rather silly) claim that ignores both the science of MTF and the way the human visual system works— the human visual system is highly responsive to acutance and it is why a good 'chrome' always looks so good— built-in acuity.
Check in-stock status of the Nikon DSLRs on the Nikon gear page.