4
\$\begingroup\$

I recently asked a question (How to increase the quality of photos taken through a DSLR?) on how to increase the sharpness of my images.

One of the answers proposed that even though it is known that for landscapes, in order to have a clear picture, it is recommended to use an aperture like f/10 or f/12, I can use f/2.8 or f/5.6 depending on my lenses.

Is this true? If yes, why does this happen and at what situations can it be applied?

I know that we can use the f/2.8 in landscape astrophotography, in order to gather as much light as possible.

The answers I am referring to is this one by xiota:

Your first image was shot at F10. The high F-number (smaller aperture) produces deeper depth of field at the expense of more visible diffraction. The DOF makes more of the image look in focus, but the diffraction sacrifices fine detail. For crop sensor, F5.6 is a good place to start. Adjust up or down as desired.

and this one by Jonas:

It is often said that the lower the zoom range of a lens, the sharper it is. Maybe try to get a "cheap" (compared to other camera lenses) prime lens (which means that the lens does not zoom; its focal length is fixed) such as a 35mm f/1.8. Then stop down this lens to, let's say f/2.8, and your images will be very sharp.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks! I don't think that one should use f/2.8 for landscape photography. What (I think) Jonas meant to say was that if you have a lens with a maximum aperture of f/1.8, you should stop down a bit to increase sharpness. They used f/2.8 as an example, but I think it's not a good example. I'm not posting it as an answer as I'm not sure. \$\endgroup\$ Commented Nov 16, 2020 at 13:32
  • \$\begingroup\$ As @SaaruLindestøkke assumed, I was only making an example and not specifically referring to landscape photography, where you should (in most cases) indeed use a small aperture/large f-number :) \$\endgroup\$
    – jng224
    Commented Nov 16, 2020 at 17:50

2 Answers 2

3
\$\begingroup\$

Different lenses have different behaviours, but as a rule of thumb: if you need maximum sharpness in the center but don't care about corners (for example you are shooting a portrait with shallow DoF), set your aperture to one step higher than lowest available. If you need sharpness across whole frame (typical landscape) start with f/8 and adjust as necessary - some lenses be sharpest at f/5.6, other will need stepping down to f/11.

While higher f-numbers increase depth of field and reduce optical aberrations in extremes of optical circle (corners and edges), they also introduce diffraction effects across whole frame leading to decrease in overall sharpness. I recommend getting a tripod and shooting the same scene few times with different settings to learn how your lens behaves at different apertures.

On lenses with small image circles (covering only crop sensor size) diffraction effects will kick in faster, this is why @xiota recommended f/5.6 as starting point for you. On the other side of a spectrum are medium and large format cameras - those will often use f/16 or f/32 and will still have good sharpness.

Additional point: getting exposure right and good post-processing is more important than overall sharpness. There are already great answers in your linked previous question. Getting a tripod and better quality prime lens will help (for landscape you can use older, cheaper manual focus lenses - that tree or mountain will not run away as you spend few seconds setting the focus). Another step is using tripod and ND filter for longer exposure times to get that nice, smooth vegetation or water surface while static objects stay razor sharp.

To answer your question in more simple terms: In general it is not a good practice to use a low aperture for landscapes, but is it is better than underexposing or getting a camera shake from long exposure time.

\$\endgroup\$
1
  • \$\begingroup\$ This answer has too many errors... smaller apertures reduce optical aberrations across the entire image circle, not just the periphery. And crop sensor lenses with smaller image circles do not cause diffraction effects any earlier. \$\endgroup\$ Commented Nov 16, 2020 at 15:39
1
\$\begingroup\$

Light itself is sharpest when there is no obstruction/aperture. When the wavefront of light has to pass through a restriction/aperture it creates a new pattern which spreads out... this is called diffraction. An analogy might be water through a garden hose... if you create a smaller restriction by placing your thumb over the opening the water sprays out in a much larger pattern.

If there was a lens that was diffraction limited (perfectly sharp) at f/1 it could theoretically project approximately 1900MP onto a full frame sensor (median/green wavelengths). But due to the restriction, and the larger airy disks resulting from the restriction (increased diffraction), the same lens could only project 4MP onto a full frame sensor at f/22... the dots (airy disks) are larger so fewer fit.

But AFAIK there are no lenses that are diffraction limited (perfectly sharp) at f/1, and most lenses have some optical aberrations/errors at their max aperture setting. So, for most lenses stopping down removes optical errors that you can see/record; while simultaneously increasing diffraction which reduces potential resolution that you can't see/record... at some point/aperture that tradeoff reaches a balance and we tend to call that the lens's sweet spot; where it is as sharp as it can be and lets as much light through as possible at the same time.

It is actually diffraction and the loss of potential resolution/sharpness that *increases the depth of field in an image (zone of acceptable focus). Because, when the size of the in-focus airy disks matches the size of the blur radius of less focused points, they are recorded as equally sharp. When the airy disks become excessively large and matching points that are even farther out of focus, to the degree that it causes a loss of perceived/acceptable sharpness, then it is diffraction limited.

So, it's all tradeoffs... do you want more resolution/sharpness from what is actually in focus (larger apertures), or do you want more that is acceptably sharp (more DoF/smaller apertures)?

(*stopping down also makes out of focus points less out of focus)

\$\endgroup\$
5
  • \$\begingroup\$ Re, "the dots (airy disks) are larger so fewer fit." That could be misleading to readers who don't fully grasp how lenses work. Really, the projected image is composed of an infinite number of "dots." They don't have to "fit" because they overlap. As for, "MP," that's a way of expressing how much detail is in the image. If we say that an image has 4MP of detail, we're saying that a capture with a 4MP sensor contains enough information to reconstruct the same digital image that would have been captured by any larger sensor. \$\endgroup\$ Commented Nov 16, 2020 at 15:40
  • \$\begingroup\$ @SolomonSlow, when the airy disks overlap they combine and cannot be resolved separately, the net result is fewer airy disks recorded (diffraction limited resolution). \$\endgroup\$ Commented Nov 16, 2020 at 15:45
  • \$\begingroup\$ @StevenKersting Deconvolution processing can untangle some of the overlap between airy disks, though. Lenses and the images they project are analog. They do not project "megapixels", nor can their resolution be properly described as "so many megapixels.". \$\endgroup\$
    – Michael C
    Commented Nov 17, 2020 at 7:32
  • \$\begingroup\$ @MichaelC, sharpening can clear up some of the overlap when it is small enough to only be causing a reduction in contrast (unless the sharpening is artificial). Lenses project dots (airy disks) which are recorded digitally by dots (photosites); the lens' resolution can quite properly be described as/correlated to the MP required to record it. This paper was written by, and contributed to, by some very prominent authorities on the subject. www2.uned.es/personal/rosuna/resources/photography/Diffraction/… \$\endgroup\$ Commented Nov 17, 2020 at 14:13
  • \$\begingroup\$ @StevenKersting What part of a digital image is not artificial? What part of a film print, for that matter? But particularly when one is talking about a camera with a Bayer masked digital sensor in which color information is calculated for all three RGB channels (because the Bayer mask uses three colors different than RGB for the filters), everything is interpolated and artificially created to represent, as closely as possible, the image projected by the lens. As the paper at your link repeatedly states, it's much more nuanced than claiming that "this lens has 'x' MPs of resolution." \$\endgroup\$
    – Michael C
    Commented Nov 18, 2020 at 1:32

Not the answer you're looking for? Browse other questions tagged or ask your own question.