Is it better to shoot at diffraction limited aperture or sharpest lens aperture?

Question

In the old cameras like Nikon D80 with 10 Megapixel pixel count. The diffraction limited aperture is F/13 but its known that lenses are generally sharpest at the F/8 so it's a no brainer to shoot at F/8 for the sharpest pictures.

But when I look at the new cameras, they have 24 Megapixel sensors with DLA of F/5.6 or so.

So when shooting at the new cameras which aperture should be taken for the sharpest pictures especially for archive grade shots? The DLA or the sharpest lens aperture? Does focal distance and the distance between the camera and the subject gets into the play too?

Things get even more complicated when you realize lens technology evolved too.

Answer 1

As is usually the case, it depends on your scenario.

The "best" aperture, which I will interpret as being the aperture at which the lens produces its maximum resolution results, will not necessarily be f/8. You'll need to check reviews of each lens that include MTF graphs, to determine the aperture at which your lens will provide the maximum resolution.

Then you have to realize that resolution performance is evaluated on charts, which are flat and have no depth. In practice, most people do not photograph charts, but 3-dimensional subjects.

Therefore, the sharpest aperture of your lens may not provide you the depth of field that you require for your subject. It will give you the best resolution, but most of your subject may end up being out of focus.

If you want to photograph a product like a camera, for example, you may find that filling the frame with it and using the optimum aperture will not allow you to keep both the lens and the body in focus. In such a case, it would be better to trade-off your top lens resolution for more depth of field. I've shot such images at f/22-f/32 and the loss of resolution due to diffraction was far less dramatic than it is spoken of.

Alternatively, if you have the time, patience, and tools, you can take multiple shots at your optimal aperture, focus each differently, and then stack them. See 'focus stacking'. This may give you the best result, but it may be hard to tell the difference from an f/32 shot if your goal is to publish the image in a journal.

Answer 2

"its known that lenses are generally sharpest at the F/8"

That is not correct. Some lenses are sharpest at f/8, most reach their max at one to two stops from wide open, and some are sharpest when wide open. You need to test your lens/combination, or find a good test online.

Strictly in terms of sharpness, you should use the widest aperture that doesn't introduce lens errors (or introduces the fewest errors). In practice, the aperture you should use depends on many factors.

Answer 3

And it gets even worse: ISO and exposure levels are also at play here as well as your personal noise tolerance: diffraction based unsharpness is not distance dependent and can be computationally compensated to a good degree at the cost of amplifying noise.

With my 10MP DSC-R1 with similar size (almost APS-C) sensor to the D80 and a better-than-kit built-in lens, I find that aperture sizes (nominally available up to f/16) are for most purposes dominated by depth of field requirements balanced with light yield. More modern cameras will have much better light yield, but depth of field comes into play more than it already does with 10MP.

I routinely use raw processing with this old camera in order to get best results; more modern cameras will incorporate smarter algorithms that factor aperture diffraction into the image processing algorithms as well as distortion corrections (which also lose a bit of resolution, so an optically well-corrected lens has sharpness advantages).

Diffraction should usually not depend on the focal distance for most optical recipes of lenses, but sharpness, field curvature, and various aberrations may.

So if you are serious about it, nothing but doing your own experiments with the exact lens/camera combination you are using will be reliable.