Luminous Landscape's example of diffraction-limited lens performance

[bdeacon]

OK, apparently my comments were controversial, so allow me to explain. It seems reasonable to me that someone would read Reichman's article and infer that using a wider aperture in the service of greater sharpness is a wise course of action. You are correct that Reichman doesn't say this in the article, so I'm "taking issue" with this idea, not Reichman's argument. My point, which I feel silly for making here for a third time now since it is common sense, is that it is often desirable to use a smaller aperture in the service of increased DOF at the expense of sharpness because the loss in sharpness is likely barely noticeable in the final print. I’m sure we are all in agreement on this so let’s let the matter rest. Frank, I prefer not to stop down to f32 if I can help it but I nevertheless do sometimes because (a) many of my subjects require it, (b) I want as much of the image as possible to appear sharp, and (c) this has little apparent effect on the quality of the drum scanned image. In his book, Jack Dykinga describes routinely shooting at f32 and f45 for the same reasons.

[Kirk Gittings]

I agree with Oren and Brett. I routinely stop down that much. Generally speaking, I would much rather have an image that will not print large than one with off plane, out of focus objects in the foreground.

I had a related discussion a couple of years ago with Mark Citret, one of my favorite photographers, after a joint presentation we were giving on architectural photography. In talking about how some of his images were made he mentioned that he oftentimes shot at f45, as do I. A member of the audience came up after and said because of diffraction you couldn't shoot at those small f stops. Mark and I looked at each other and simultaneously said "really?" You choose your compromises with LF. Generally I would rather have an overall level of sharpness that may only look good up to 11x14.

[Frank Petronio]

Yeah but in achitecture you often have an off-center near object on your left and on your right, and yet you still need to carry DOF back. So you stop down.

But geez, if you're taking a photo of El Capitain in Yosemite, or of some mountain or the side of a barn or rushing waterfall or a sunset or any of the other million genre photos that people make... you really don't need f/32 or smaller...

And if anything, wider open helps stop the wind from messing with ya.

[QT Luong]

Maybe one way to distinguish your genre photos from the other millions available made by dabblers is to use strong compositions that quite often involve having some kind of prominent foreground ? As to whether f32 or smaller is often needed, check the f-stops listed in a John Sexton monograph. For instance, Listen to the Trees: 50% at f32, 25% at f45 and less.

[Ole Tjugen]

Frank - you will need f:32 or smaller if there's any foreground in the picture.
http://www.bruraholo.no/images/Lodalen_GF.jpg is an example - f:32 was necessary to keep some definition in the branches on the right. Even if most of the scene is distant, that foreground is important.

[robc]

I have a theory about sharpest aperture. Diffraction must be partly a function of iris edge area related to aperture area. A rough calculation giving an aperture diameter of 5-7mm gives the approx sharpest aperture. So on 35mm camera a normal lens is approx 50mm so an aperture diameter of 6mm would give approx f8. On a 4x5 camera with 150mm lens and 6mm aperture diameter would give f25. A 10x8 camera with a 300mm lens and 6mm aperture diameter would give f50.
Those numbers seem close to what we usually find are our sharpest apertures on those focal length lenses. So my theory suggests that aperture diameter is fairly constant across lens focal length for optimum sharpness. I know there will always be an exception to the rule.

Now can anybody confirm this with maths. Infact I will do a poll later and we can see what people think is their sharpest aperture on the common focal lengths.

[Dan Fromm]

robc wrote "Infact I will do a poll later and we can see what people think is their sharpest aperture on the common focal lengths."

Rob, your theory is, to be polite, unconventional. Sharpness in the image on film is independent of film size.

By an odd coincidence, I mainly shoot lenses on a format smaller than the ones they were designed to cover, so for me lenses' central sharpness is much more important than sharpness even midway between the center and edge of the fields they were designed for. Short answer to your question, the aperture at which the lens has highest on-film (or, for macro lenses, on subject) resolution depends on the lens and the magnification.

Good macro lenses (designed for shooting above 1:1) are nearly all sharpest wide open. Lenses designed for shooting below 1:1 are all over the place. I have some that are as good wide open as at f/16 on the emulsions I shoot. Others that are not that great wide open, hit their best a couple or three stops down. Understand, now, that I'm talking about central resolution, not resolution in the corners. There are aperture dependent off-axis aberrations, and that is why stopping down improves image quality in the corners.

My lenses don't agree with your rule. Your rule is nonsense, you should do the experiments.

[naturephoto1]

Perhaps I am incorrect, but doesn't diffraction start to be of real concern when we start to get to within about 2 stops of the minimum marked settings for the aperture for a lens? Most of us that shoot LF tend to avoid the use of the smallest apertures unless needed. I know that many of us will go further than that when there is reason as an example for maintainining the near/far relationship in landscape. I do shoot many landcapes in 4 X 5.

Of course the longer the lens the less the depth of field at any aperture at the same distance. As an example, I use the following lenses: Rodenstock f6.8 90mm Grandagon N MC, Rodenstock f5.6 150mm Apo Sironar S, Rodenstock f9 240mm Apo Ronar MC, and Rodenstock f9 300mm Apo Ronar MC. The Grandagon and the 2 Apo Ronar lenses are optimized for f22 and as I recall the Sironar S is optimized for between f11 and 16. The minimum marked apertue settings for the lenses are f45, f64, and f90 for the Grandagon, Sironar S, and Apo Ronars respectively. Whenever possible, I try to not use the smallest or the 2 smallest apertures when possible for the Grandagon (f45), Sironar (f45, f64), Apo Ronars (f64, f90). For the Apo Ronar lenses the sweet spot is probably f22 though f32 is still very very good.

As an example, and whenever possible for the Grandagon lens as mentioned, I try to avoid the usage of f32 or f45. I have had some very good success at f45. Additionally, I have taken with this lens what many people consider to be one of the sharpest photos that they have ever seen printed as 24" X 30" and 30" X 37 1/2" when printed off of a Chromira machine from a drum scan. The original Velvia 50 transparency is very sharp even under a 10X loupe. The image mentioned is described in the following paragraph.

I am attaching the photo Dead Horse Point State Park. No tilts were used in this instance because of the infinite number of planes and as such depth of field was maintained by aperture alone. This image was taken at about 8:30 AM in May of 1995. The image was taken with a Linhof Technikardan 45S, 90mm f6.8 Grandagon N MC at f32, Heliopan Linear Polarizing Filter, Tiffen 812 Filter, Fujichrome Velvia 50, Gitzo 320?, Arca Swiss B1? ball. No Graduated ND was used or available. There is some light fall-off more evident in the sky (much of which corrected in photoshop and not as evident in the large prints as shown). Additionally there is a small amount of loss of sharpness toward the edges. Image was prepared by Bill Nordstrom (Laser Light Photographics), founder of EverColor Fine Art and my printer for the last 12 years.

Rich

[Oren Grad]

Brett: Sorry for the late-night crankitude - nothing like having a good, rip-roaring agreement...

[bdeacon]

No sweat Oren. I'm up for having a heated agreement anytime.