Skip to main content
1 of 2
Edgar Bonet
  • 3.5k
  • 21
  • 17

The article you cite is not very good advice if you want great sharpness for landscape photography: it’s based on the concepts of depth of field and hyperfocal distance. These concepts are intended to help the photographer find the required aperture for getting barely acceptable sharpness across the relevant parts of the picture. What the author (like many authors) calls “circle of confusion” is actually the maximum acceptable circle of confusion. In its original sense, the circle of confusion is the size of the blur spot you would get from a point source: if it’s larger than the maximum size you can tolerate, then the image is not acceptably sharp. The key word here is acceptable. If you are after just-acceptable sharpness, these concepts are good. If you are after maximal sharpness (likely for the kind of landscapes shown as examples), they are not.

Assume that there are 3 objects, located at 0.4m, 1.2m, 5.0m.

In such a situation, if you want maximum sharpness, you have to focus at mid distance between the closest and farthest object, where “mid distance” is actually the harmonic mean of the distances:

                             0.4 m × 5 m
optimal focus distance = 2 × ─────────── = 0.741 m
                             0.4 m + 5 m

The harmonic mean is easy to estimate by just looking at the distance scale of your lens: it's exactly halfway between the 0.4 m and the 5 m marks.

Now you can estimate the circle of confusion that you will get at any subject distance with

N × c   │         1                1        │
───── = │ ──────────────── − ────────────── │
  f²    │ subject distance   focus distance │

where c is the circle of confusion, N the aperture number, and the vertical bars mean “absolute value”. This formula shows that, at f/11, both your furthest and closest object will be imaged with a 42 µm circle of confusion, which is not really that sharp. You have to stop down to f/16 if you want to stay below the canonical 30 µm limit. Stopping to f/22 will still increase sharpness, but stopping further to f/32 will reduce sharpness due to diffraction.

If you want maximal sharpness on both these objects, you will have to balance the blurring due to defocusing (gets better as you stop down) with the blurring due to diffraction (gets worse as you stop down). Rather than going through the math, I suggest you read Selecting the Sharpest Aperture by Ken Rockwell. I know the author is controversial, but in this particular instance the article is very sound.

Edgar Bonet
  • 3.5k
  • 21
  • 17