sign in sign up
<<
Home , Circle of confusion, Hyperfocal distance

by Harold M. Merklinger ADJUSTING DEPTH-OF-FIELD as published in Shutterbug Oct. 1991.

One upon a time, in the land of se- to infinity. de- out there—over a mile away—at rious photographers, I too, believed pends upon the of the 1,929.20 meters. Am I supposed to in the myths of depth-of-field. By used, upon the f-stop used and call her up on the cellular phone and the myths of depth-of-field, I mean to upon the criterion used for acceptable say “Would you mind stepping back refer, of course, to the definitions, sharpness. by about three-quarters on an inch? rules, tables, graphs and charts found Before we get down to serious The wind is blowing your hair to- in almost every serious book on the business, let me explain what I be- wards the —it’s just a tad out- subject of . After about lieve to be the three most serious and of-—... Ah, that’s it,... beauti- three decades of failing to achieve erroneous myths associated with the ful!” This is patently ridiculous. She my goal of confidently producing subject of depth-of-field. Number could walk closer to the camera by a really sharp, detailed images, the one is the “one-third rule”. Have you quarter of a mile—or more—and I light finally dawned. And I have ever read that you should focus one- wouldn’t be able to see the differ- lived rather more happily with this third of the way through the field you ence! subject ever since. want to be sharp? It’s in many The fact that tables sometimes I used to switch back and forth be- books; I’m sure you have. Well, it’s specify distances to six significant tween and 35 mm, almost never true. Mathematically it digits tends to make one think that trying to capture that quality I some- is true under precisely one condition: things are that critical. They are not. times achieved, apparently by ac- when the far limit of depth-of-field is The reason the tables have so many cident. If anything, my medium for- precisely twice as far from your cam- figures in them is that it’s easy to cal- mat were less detailed era lens as is the near limit of depth- culate tables to that degree accuracy. than the 35 mm photos, but prettier— of-field. Under any other conditions, But tables are not that accurate in the nicer grays and such. the math (as opposed to the myth) first place. For a start, we generally Then one Friday night it hit me: says you should find some other don’t know the focal length of our the depth-of-field rules told me how point on which to focus. So much for lens that accurately, or it’s actual f- to achieve the “minimum acceptable that one. stop for that matter. Furthermore, re- sharpness”. And that’s exactly what I Number two is: “if you want to sidual aberrations in cause the was getting: minimum acceptable re- maximize your depth-of-field, focus distance at which the lens is actually sults: amateurish results and some- your lens at the hyperfocal distance focused to change with all sorts of times down right fuzzy results. for the f-stop you are using.” This things like f-stop, temperature, and How might a photographer re- rule is mathematically correct, but it even subject contrast. And where the liably achieve super-sharp results all may not give you what you expect. lens is focused will probably affect the time? In this article I’ll try to ex- We’ll study this one in detail a bit lat- our image far more than anything plain what the photography books er, along with its corollary: “if you’re else. For close-ups, tables might be didn’t tell you about depth-of-field not happy with the results you get useful. For anything else, a well de- and sharpness, and I hope this will this way, close the lens down a stop signed depth-of-field scale on the lens help you to achieve the results you or two extra.” It is my view that this is probably more useful. really want. rule was once true, but is today out- So, on with the main thrust of this For those who may not be familiar dated by superior films and lenses. dissertation: what’s wrong with the with the concept of depth-of-field, I Myth number three is that for best time-honoured treatment of depth-of- offer this brief explanation. Ac- results one should not use the depth- field? Why was I unhappy with my cording to the theories of geometrical of-field scale on the lens; rather use photographs? The answer lies in the , a lens focused at distance ‘X’ pre-calculated tables: they are more assumptions that were made. The will require that, for maximum sharp- accurate. It’s hard to say that this one theory, the mathematics: they’re ad- ness, the subject be positioned exact- is wrong, but it’s not as right as you equate. It’s the application of the the- ly at distance ‘X’. In practice, it is might think. And worse, tables tend ory with which I take issue. found that if the subject is a little to be misleading. The depth-of-field Assumption number one is that a closer or a little farther away, even scale on your lens is actually very circle-of-confusion of one-thirtieth of the best scientific instrumentation helpful; it gives you a lot more in- a millimeter diameter or less is ad- will not be able to detect the differ- formation and flexibility than any sin- equate for 35 mm photography. (The ence in sharpness. So, in practice, gle table ever could. circle-of-confusion is the perhaps best there is a zone within which any sub- The main trouble with depth-of- explained in the present context as the ject will be imaged “sharply”, or with field tables is that they give a false out-of-focus disk image produced by acceptable sharpness. That zone of sense of what to expect. For ex- the tiniest of point objects.) This lens-to-subject distances is the depth- ample, the depth-of-field table for the number, one-thirtieth of a millimeter, of-field. Countless books and mag- 200 mm Micro-Nikkor, says that at was derived, I believe by Leitz, in the azine articles will tell you how to de- an f-stop of f/5.6 and with the lens set 1920s or ’30s when that was the best termine how big that zone is. at infinity focus, the depth-of-field resolution (or, in their words “thick- A related useful concept is that of will extend from “1929.22” meters to ness of outline”) that the best films of the hyperfocal distance. If a lens is infinity. Note that the inner limit of the day could produce. focused at its hyperfocal distance, the depth-of-field is specified to one cen- You will also find justification for depth-of-field is claimed to extend timeter, or about three-eighths of an this figure on two other grounds. I from one-half the hyperfocal distance inch. Let’s suppose I have a model have seen articles written in the 1940s a) b) c)

Figure 1: My sister-in-law, June, as photographed under various conditions as follows. a) June is at a distance of about 160 ft; the 50 mm f/8 lens is focused as shown in Figure 2 at its hyperfocal distance. b) As for a) but with the lens focused at infinity. The enlargement factor for both of these photos is about 120 times. c) Here the lens is still focused at infinity, but June has walked forward to within 10 ft. of the camera. For c) the enlargement factor was reduced to result in the same size image as for the other two examples.

feet 100 5 5.5 6 7 8 10 15 25 50 ∞

22 16 11 8 42 2 4 81116 22 5.6 5.6 11

Figure 2: Here’s what the distance and depth-of-field scales might look like on a 50 mm lens focused at its hyperfocal distance for f/8. The depth-of-field scale has been drawn for as large as f/1 and for an allowable circle-of-confusion one-thirtieth of a millimeter in diameter. (This diagram was printed incorrectly in Shutterbug.) which state that the average lens, at look at a 31/2 by 5 in. print, taken us- My next step was to ask: in order its best , can produce a circle- ing the rules for a 1/30 mm circle-of- to achieve a circle-of-confusion seven of-confusion no smaller than one- confusion, it looks slightly fuzzy to times smaller than that assumed by thirtieth of a millimeter. The ex- me. And my eyes are not all that my depth-of-field scale, how many planation most commonly seen these good. Second, when I analyzed those stops would I have to close the lens days is that the human eye, looking at accidental prints with which I was down by? The answer is about six! an 8 inch by 10 inch at happy, I discovered that my desired No wonder closing the lens down by normal viewing distances, can detect standard corresponded not to 1/30 only one or two stops didn’t do the spots no smaller than one-hundredth mm, but rather to 1/150 mm. Taking trick. of an inch. Given that a 35 mm neg- into account the loss in quality due to Another way to express the result ative would have to be enlarged by a the enlarging lens, the diameter of the is as follows. Let’s say I am really factor of eight to produce the print, circle-of-confusion on the using a 50 mm lens at an aperture of the equivalent spot size on the neg- would have to be no larger than about f/8. Is there an mark on my depth-of- ative would have to be one-eight- 1/200 mm. That is, I was demanding a field scale I can use to estimate the hundredth of an inch. That is, ex- sharper negative by a factor of about true depth-of-field for my revised as- pressed in millimeters, one-thirty- seven! And getting it—sometimes. sumption about the maximum accept- second of a millimeter. Clearly today’s films and lenses are able diameter of the circle-of- My problem with this is two-fold. much better than they were in 1930s confusion? The answer is yes: if I Never mind the eight-by-ten, when I and ’40s! want to improve definition by a factor of 7, that is, use 1/210 mm instead of quarter of a mile inside the supposed References: 1/30 mm for the diameter of the circle- inner limit of depth-of-field and my of-confusion, I should divide the f- not being able to see the difference. Vith, Fritz, Leica Handbook, Tech- number I am really using by 7 and It turns out that under many circum- nisch-Pädagogischer Verlag, Scharfes use the depth-of-field marks for that stances we can tolerate the subject Druckereien Kom-Ges., Wetzlar, Ger- aperture. 8 divided by 7 is about 1.1. being well inside the calculated inner many (1933). I should use the depth-of-field marks limit of depth-of-field as shown in for f/1.1! But my lens is an f/2 lens; Figure 1 c). But that’s another story. Zieler, H.W., “ there isn’t a pair of f/1.1 marks on the In summary, it is my contention and Disc” in The Com- scale. OK, the depth-of-field scale— that the standards used for calculating plete Photographer, W.D. Morgan as you might have noticed—is a lin- depth-of-field are no longer appropri- (Ed.), Vol 2, 768-781, National Ed- ear scale. The f/8 depth-of-field ate. Today’s lenses and films are ucation Alliance Inc., New York marks are twice as far from the focus much better than those of the days (1949). pointer as are the f/4 marks. The f/22 when the image quality standards marks are twice as far from the point- were derived. My experience sug- Cook, G.H., “Miniature Camera er as are the f/11 marks, etc. And so gests that today’s films and lenses are Lenses Part I” in Leica Photography the f/1 mark would be half-way be- about seven times better than typical Vol 3, No. 11 E. Leitz Inc. New York tween the focus pointer and the f/2 depth-of-field tables assume. And (1950). marks. that’s the case for 35 mm . Can I revise my hyperfocal dis- Medium format results are probably tance? Yes, multiply it by 7! If for even better. an allowable circle-of-confusion of 1/ If we were to revise our depth-of- 30 mm in diameter, my hyperfocal field scales for seven times the as- distance at f/8 is 32 feet, for a per- sumed resolution, depth-of-field missible circle-of-confusion of 1/210 would seem to just about vanish. Yet mm, the hyperfocal distance would be my tests show that many subjects are 224 ft. quite adequately resolved well inside Seems like I don’t get much depth- the calculated inner limit of depth-of- of-field doesn’t it. Well, it’s not so field. Further adjustment to the the- bad as you might think. Remember ory would seem to be in order. what I said about the model walking a © Harold M. Merklinger, 1991