Question on diffraction limiting . . .

[Guest Walt]

Advertisement (gone after registration)

I remember some discussions on the forum, though I was unable to find them with a search. I recently read the article and used the "calculator" at Diffraction Limited Photography: Pixel Size, Aperture and Airy Disks

 

The suggestion here, and in use of their calculator, is that with a 9 x 14 inch print diffraction limits are not reached even at F11. For 16 x 20, F8 seems the minimum aperture size before diffraction limits are reached. This calculator is presumably accounting for print size, viewing distance (25 cm), sensor resolution (10.3), crop factor (1.3), aperture, pixel size (7um), maximum circle of confusion (14.1um) and the diameter of the airy disk (10.7um). I realize that it is not accounting for the design of the Leica lenses (very good correction wide open), variations in definition of sharpness, etc.

 

I am wondering how credible people find this calculator and the conceptualization behind it. My experience is that most of the lenses I am using on the M8 are really best around 5.6. This includes both the 28 ASPHs and both the 35 ASPHs. So I'm thinking that perhaps this has less to do with reaching diffraction limits on these lenses than with other design parameters of the lenses that provide better correction at larger apertures. The 28/2.8 ASPH seems the exception to me, performing well even at F11 in my impression.

 

What are people's thoughts and observed experience on this?

 

Walt

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[gmaurizio]

given the precision of Leica lenses (and Zeiss and some CV ones), I would say that you reach diffraction blurring from F11 and up. In normal photography, you can safely use F16. For architecture, product and editorial, I would try to stay below F8. This is regardless of how large you print

You will 'see' a good picture, even at F16 or in some occasions F22, but the blur is there. In my opinion, this only matters in technical photography. Leica lenses are optimized for wide open, and many reach their peak performance at F5.6 or F4.

[ho_co]

Walt, the problem I have with the calculator you link to is that it uses simple mathematical calculations which simply set upper bounds without regard to the quality of the lens in question (e.g., do you believe all 50/1.4 lenses are diffraction limited at the same aperture?) and doesn't incorporate presence or absence of an AA filter. Seems to me it's a decent illustration of a limited set of criteria, though it would probably have been fun to program.

 

Extending the thought: Do check out the 'Yes it does-No it doesn't' article at Luminous landscape.

 

Some information on the topic is also available in the LFI article "Form Follows Format" from about a year ago.

 

Basically, as Gus implied, stopping most Leica lenses down beyond f/4 or f/5.6 will bring diffraction into the game on the M8. Lens reviews by Erwin Puts on film have long said that f/8 is already too small for most Leica lenses, and several forum members have commented that they have found the same to be the case with the M8.

 

--HC

[mjh]

The suggestion here, and in use of their calculator, is that with a 9 x 14 inch print diffraction limits are not reached even at F11. For 16 x 20, F8 seems the minimum aperture size before diffraction limits are reached. This calculator is presumably accounting for print size, viewing distance (25 cm), sensor resolution (10.3), crop factor (1.3), aperture, pixel size (7um), maximum circle of confusion (14.1um) and the diameter of the airy disk (10.7um). I realize that it is not accounting for the design of the Leica lenses (very good correction wide open), variations in definition of sharpness, etc.

 

I am wondering how credible people find this calculator and the conceptualization behind it.

The term “diffraction limited” has come to be used in slightly different ways, causing some confusion. Traditionally, being diffraction-limited is a term of praise: a lens is said to be diffraction-limited (at a certain f-stop) if its resolution is only limited by diffraction – not by spherical aberration or whatever. In other words, it performs as good as the laws of optics allow.

 

What the calculator calculates is something else, namely at what point the resolution limits imposed by diffraction become noticeable, provided they are not masked by more severe aberrations caused by the lens. I think the calculator is most useful when you set the CoC to twice the pixel size, thus ignoring print size, viewing distance etc.. With this option set, the calculator tells you what mileage to expect from a given sensor, assuming for the moment that lens performance is limited only by diffraction – things can only get worse with lens aberrations taken into account.

 

On the other hand, the effective resolution depends on the correction of the various aberrations just as well as on diffraction, and with some aberrations being reduced by stopping down, the highest resolution may be achieved at different f-stops, depending on the quality of the lens – the better the lens, the bigger the aperture at which the lens will be diffraction-limited. A lens that is diffraction-limited at f2 is superior to a lens that is diffraction-limited only at f5.6.

[sean_reid]

I second Howard's recommendation of the LL article. In practice, I start to see diffraction limiting resolution, with many modern RF lenses, beginning at F/8. So F/1.4 - F/5.6 can, indeed, often be the sweet spot. That's assuming, of course, that those small decreases in resolution (at F8 and moreso at F/11) are of any real consequence to the picture. Often, they are not.

 

Cheers,

 

Sean

[Guest guy_mancuso]

Same here i am not seeing any damage to the files when i go to f11. I'm sure it is doing it but it is not noticable . Most Leica lenses hit there peak at F4 or 5.6 but also Leica lenses are really the best around wide open. Zeiss also has great wide open performance. This is what seperates the men from the boys in lens design is wide open performance. Leica has it going on better than any in my book in a overall lens systems.

[Artichoke]

Advertisement (gone after registration)

Same here i am not seeing any damage to the files when i go to f11. I'm sure it is doing it but it is not noticable . Most Leica lenses hit there peak at F4 or 5.6 but also Leica lenses are really the best around wide open. Zeiss also has great wide open performance. This is what seperates the men from the boys in lens design is wide open performance. Leica has it going on better than any in my book in a overall lens systems.

unless shooting macros

in order to obtain sufficient DOF, you must stop down well beyond f8

I have found my 135 Tele Elmar holds quality up to about f13 & is reasonable at f16 ...my 65 chrome Elmar does not do quite as well when stopped down that far

OTOH my Nikkor 85 pc/tilt Micro seems to do much better with extreme stopping down ...now if I could only find a way to use this stellar macro lens with the M8, I would be in heaven

the M8 is not designed to do macro photography, but I have found it can deliver excellent results with the Viso/bellows combination & have not found the lenses available to this much of a liability, though I do miss tilt/shift

[jaapv]

OTOH my Nikkor 85 pc/tilt Micro seems to do much better with extreme stopping down ...now if I could only find a way to use this stellar macro lens with the M8, I would be in heaven

 

Easy - Novoflex bellows - Visa ring and Nikon adapter.Maybe one could even superglue the ring and the adapter together and eliminate the bellows. You'd lose infinity I suppose, but for macro it should work fine.

[Guest Walt]

Thanks, folks, for the thoughts. This pretty much supports my impression of 5.6 as optimal. I did look at the LL article (I don't know what "LFI" is) and it was both a lot more than I wanted to know, but did bring up some interesting ideas. I asked the original question because I would sometimes like to be able to operate with more depth of field on the M8 without developing that soft look, which I usually don't like. I am now going to do a few tests on using a larger proportion of blue channel in my BW conversions to see what this looks like and if it allows smaller apertures with less softening. The green and red channels, because of wavelength, appear to contribute the most to diffraction limits.

 

Walt

[henning]

As Michael stated, 'diffraction limited' indicates an aperture where the residual aberrations are brought under control to the extent that diffraction now starts being the major factor in determining achievable resolution. Particularly well corrected lenses are diffraction limited at apertures close to maximum. The ASPH lenses are all at their best around two stops from maximum at the center, and possibly a bit more in the corners.

 

Wider angle lenses generally also should not be stopped down as much for optimum performance. The 12 Heliar shows a noticeable falloff at f/11, and is usually best wide open at f/5.6. The 15 is also best at f/5.6. Both these lenses are quite poor at f16, whereas the 135 lenses are still pretty decent at f/16.

 

The 75/2 is one of the best ones; it really is at it's best at f/2.8, but it performs at such a high level that even at f/8 it is still excellent.

 

If you do macro and micro, aperture use becomes even more critical. The Photars should not be stopped down at all if possible, and this becomes even more critical with the shorter lenses. I can see the difference very clearly between f/2.5 and f/4 with the 25mm Photar.

 

With the standard range of M lenses, stopping down for dof is often necessary, but for best quality stopping down 'just in case' should be avoided.

 

In the main, just set and shoot. The quality from the M8 with ASPH lenses will hardly ever be the limiting factor in your (or my) shots, no matter what f-stop is used.

 

Henning

[ho_co]

I did look at the LL article (I don't know what "LFI" is) and it was both a lot more than I wanted to know,

 

LuLa article is good basic info. Article in LFI (Leica Fotografie International) goes into a lot of specifics that relate closely to the Luminous Landscape article, but isn't directly concerned with the topic.

[wparsonsgisnet]

Walt, I started a thread called "F8, last stop for gas" that is full of neat postings.

 

In general, I see the image begin to degrade on the M8 after closing down past f8.

[Guest Walt]

Well, I did the obvious thing, a personal test of the issue. To my surprise, the diffraction issue seems rather inconsequential to me in a direct comparison. I photographed the same outdoor scene with a 35 Summicron ASPH, once at F8 and once at F16. In full frame 13 x 19 inch prints, there is very, very little difference at any viewing distance, so little that I wouldn't give it a thought. I'm not even sure it's detectable. At 100% on screen the difference can be seen, but it is subtle enough that it requires some looking to find. And, of course, the F16 shot benefited from much more DOF and, as a result, gave an impression of better sharpness across the image both on screen and in the prints. Subject matter in the scene ran from about 10 feet to near infinity.

 

These two shots were processed identically in ACR and sharpened identically in PS.

 

This was a big surprise to me--I don't know if this is going to hold across different Leica lenses, but I'm going to use the camera this way until I see a reason not to. I value the ability to use the M8 as a point-and-shoot with a 28 or even 35 and F16 allows that. I'm realizing how little the 100% screen view pertains to photography and I'm glad to know that. I haven't subjected the prints to a loupe and I see no reason to do that.

 

Walt

[gmaurizio]

Walt, I guess it is hard to see diffraction in landscapes or outdoors, due to the rather fuzzy nature of the edges. You will be able to see it with longer focal lenses and more product or technical oriented pictures. F22 might be too much of a stretch.

Also, remember the old large format rule (8x10, remember?), when in doubt, focus at infinity and use the smallest aperture. For instance, using a 28mm lens at F16 focused at infinity, the level of detail you are able to record, even in the very near fields of the picture, is the quotiente of Focal/Fnumber, meaning in this case 28/16mm!!! That is less than 2mm detail of picture space!!!! Considering everything farther than 200 focal lengths (in this case farther than 6 meters or 20 feet) will be very sharp, and you get detail of anything larger than 2mm before that, I think it is worth taking at least one shot.

With digital, people have forgotten this rule scared by diffraction.

Other highly pictorical value can be obtained with a Leica M8 body cap with a very small circular perforation (preferably laser made) equivalent to F200 or more. It makes the M8 behave like a pinhole camera, for less than $50. The dreamy effect is just gorgeous. Try it on for size....

My 2¢