There has been much discussion and debate on this forum about which Leica lenses (at various focal lengths) are sharpest at either their widest or their best apertures. For modern Leica M lenses, the best aperture is usually f5.6 or even wider, with gradual losses at smaller apertures due to (I assume) diffraction. My question is about which Leica M lenses are best at the smaller apertures typically used for landscape shooting: f8 and smaller.

If test reports from the various magazines are to be believed, as their apertures are made smaller, lenses lose sharpness at different rates -- even lenses of the same focal length. For example, the french magazine Chasseur D'Images reports that for both current 90's best image performance comes at f5.6, with the 90 Apo being the sharper lens. But at f8 the Apo has dropped in performance (at the center) from excellent to very good; while the 90 Elmerit has lost no center sharpness at all and has lost only a very little at the edges. Do any Leica shooters have confirmation of this from experience with these two lenses?

Also, why, as the aperture gets smaller, should performance fall off at different rates for lenses of the same focal length? I would have thought that for all well corrected lenses diffraction would affect them about equally, so that there would be little to choose between any two good lenses at, say, f11.

Thanks for any light you can shed on this.

-- David B. Mark (dbmark@ix.netcom.com), April 03, 2002

Answers

Interesting question David. I was thinking something similar about my Tri Elmar, which stops down to f22. My thought was, if the standard lenses at the three focal lengths stop at f16, will the fact that they can squeeze f22 out of the Tri Elmar mean the performance at smaller apertures is proportionately better at f11 or f16? Or does it get even worse using f22?

-- Steve Barnett (barnet@globalnet.co.uk), April 03, 2002.

Sorry not really an answer to your question but In the real world on any lens longer than even a 35mm apertures bellow 5.6 are going to give you quite shallow 2d planes of focus - good for test targets! If I'm shooting a scene with any depth to it then the smaller the aperture the better - the sharpness lost through diffraction is insignificant compared with the improvement in sharpness you get for objects behind and infront of the focusing point through the increase in depth of field. I've never understood this obsession with wide aperture performance - the world is 3D ! ( shoot me down in flames now)

-- John Griffin (john@griffinphoto.u-net.com), April 03, 2002.

The reason we like wide open aperture performance is due to the fact we're not always in strong lighting...hence opening it up to get a good shutter speed. : ) And it's cool to isolate subjects sometimes.

-- James (snodoggydogg@hotmail.com), April 03, 2002.

My old 35mm Summarons (both f3.5 and f/2.8 versions) seemed to have a good ability to be stopped way down without diffraction setting in. Of course, when you factor in how slow they are at the beginning, this makes sense, but I have some nice hyper-focused f/22 shots from both lenses that have no fall-off in performance (and maybe improved performance for the f/3.5 model). On the other hand, I have some ruined hand-held shot that would have been better served by f/2.0 or f/1.4, than the slow apertures of the Summarons. The faster apertures are more useful to me on a hand-held camera.

-- Al Smith (smith58@msn.com), April 03, 2002.

A theory only: long lenses are less diffraction limited because: 1.) the angles of the light path are less severe. 2.) the diameter of the aperature are larger for any given stop.

-- Steve (leitz_not_leica@hotmail.com), April 03, 2002.

I'm also interested in relative performance of the modern lenses (such as 28/2 ASPH, 35/1.4 ASPH, &75/1.4, 90/2 AA and even very late 135/2.8 E55) when stopped down to f/11 or f/16. While favoring the faster lenses for the reasons already mentioned, I'm often faced with brighter conditions when traveling (for example LA night life which is the reason to bring the quick glass, then out during the day at Santa Monica pier, where at 1/1000th I'm forced into f/11 or 16). I seemed to have noticed some lessening of resolution in 12" wide prints compaired to those shot at f/4 to f/8. But this is hardly a scientific comparison given the variables of shooting conditions. Any insight anyone?

-- Marc Williams (mwilliams111313MI@comcast.com), April 03, 2002.

david -- you should, of course, be using your cobweb-infested canham for landscape work off the tripod. however, to answer your question about diffraction effects, remember that diffraction is focal length dependent -- a 90mm lens will show less diffraction at f16 than a 35mm lens also at f16. this is because the degree of diffraction effects (with light, sound, water or anyhing else that travels in waveforms) is dependent upon two factors only: opening diameter and wavelength. since the actual iris opening at f16 on a 90mm lens is larger than the opening at f16 on a 35mm lens (f stops being ratios), and assuming that the wavelength(s) of light passing through both remain(s) constant, there will be more diffraction when the light passes through the tinier hole of the 35mm lens. simple, rite?? well, there is one other factor to consider: magnification. although there is less diffraction with the 90mm lens, it also magnifies more, exacerbating the visual effects of diffraction. however, thanks to calculations contained in howard bond's article in PT (it was bond, rite?), it appears that the mag effect lags behind the increased diffraction effect (the former increase arithmetically, while the latter geometrically), so the 90mm is still the better choice. however, all leica lenses are diffraction limited by f8 or so. if you need the small aps for dof, and your subject isn't going anywhere, you really should be using a camera with movements. but you knew that already!

-- roger michel (michel@tcn.org), April 03, 2002.

You'll be hard pressed getting all the resolution onto film from any of the R APO lenses. I've got the APO-Macro 100mm, amazing lense - extremely sharp. I've also heard good things about the APO 180mm's.

Thing is, to extract the kind of hyper sharpness you want you're going to have to put up with a v.heavy tripod, mirror-lock-up, long floppy cable release, no wind and of course extremely fine-grained, sharp film.

BTW, here's a thought for sharpness fetishists... I recently looked through a book called "Life goes to War", full of PJ shots taken during WWII and Korea and Vietnam. Much to my surprise, every single shot was blurred - either due to camera shake or sloppy focus or a poor/dirty lens! Now if these shots were good enough to go into Life's archives and in some cases re-published 30 years after they were taken, then surely any Leica Summicron at f5.6 or f8 is sharp enough?... ;?)

-- Andrew Nemeth (azn@nemeng.com), April 03, 2002.

Roger, thank you for your thoughtful post. A couple of points in response: 1. There are no cobwebs on my view camera; I dust it every week. 2. Your view that, all other things being equal, diffraction effects at a given aperture diminish as focal length increases sounds plausible -- except that I have never seen any empirical verification. On the contrary, the best small aperture (smaller than f8) test results I have ever seen for any 35 mm lens have been for lenses in the normal to moderate wideangle category. If your view is correct surely we should see long lenses regularly beating shorter lenses at small apertures -- and we don't. 3. In any event, you haven't answered my question: why should two lenses of the same make (Leica) and focal length not perform similarly at small apertures? As I said in my original post, test reports indicate that , e.g., by f8 the 90 elmarit has a discernible sharpness advantage over the renowned 90 Apo.

Andrew, surely it is reasonable to demand a higher level of technical quality in a landscape photograph than in a Photo-journalist photo taken in wartime. If 35 mm landscape photographers are going to compete with those using the larger formats, we must use all our technical skill to get as much as possible out of the small negative.

-- David Mark (dbmark@ix.netcom.com), April 03, 2002.

magnification, my boy, magnification. in my explanation, i should have emphasized that comparison must be between the same picture sections. if unadjusted for picture section, as in mtf testing, of course the superwide, in terms of resolution, will be least affected by diffraction. because the mag factor is so low, you simply can't see it (that well). remember, the effects of distortion remain constant for a give slit width and wavelength. how well you see it depends on how close you look. as for why the elmarit m and summi AA show different degrees of res falloff as you stop down, the answer lies in the fact that the two lenses embody very different optical formulas. sharpness at small apertures is not merely a function of "theoretical perfect lens" minus diffraction at small apertures. all of the other aberrations that beset imperfect (i.e. manmade) lenses still remain in varying degrees. in some lens designs, these continue to be corrected as you stop down, even past the point where diffraction really kicks in, and in some the incremental effect is negligible past that same point. in view of the radically diff formulas in the elmarit m and summi AA, i'm not sure why you would expect resolution plotted against ap to be the same for the two optics. finally, as for the specific case of the elmarit m versus the summi AA, you should scroll down a bit. i created quite a furor a day or so ago by positing exactly what you have observed (i also cited chasseur d'i); namely, that the elmarit m is sharper than the summi, largely due IMO to contrast effects, at all crit aps. and you don't dust.

-- roger michel (michel@tcn.org), April 03, 2002.

david -- vis-a-vis your comment to andrew: there is no way that 35mm will EVER be able to compete with large format for static landscapes shot from a tripod. as you well know, even apart from the mag factors involved in getting a teensy 35mm neg up to 11x14 (as compared to the same "enlargement" from an 8x10 neg), the perspective/focus controls of LF cameras put them in an entirely different league than fixed lens/film plane cameras in terms of suitability for landscape work. i don't think there is any well- known landscape photogapher who shoots in 35mm -- MF at the very least. let's face it, a mamiya 7 is not appreciably larger than an M7, and yet the neg in the mamiya is four times larger. the difference, such as there is, betwen the two cameras lies in handling and response time. these are irrelevant to a lanscape photog. why would any landscape shooter choose the miniature format?

-- roger michel (michel@tcn.org), April 03, 2002.

I've noticed that my 50mm lenses seem to work better stopped down than my wide angles or teles.I dont know, but maybe it has all the benefits of being in the middle focal range with more depth of field than a 90mm and none of the distortion or diffraction of the wide angles.I shoot medium and large format too and seem to notice diffraction more there than with my 35mm shots with 50mm DR Summicron or Noctilux.But this is what I see ...not a scientific test result which can be quite surprising.I recently tested my LF lenses and was surprised at the differences in quality as the lenses changed!You really need to test for yourself.

-- Emile de Leon (knuightpeople@msn.com), April 03, 2002.

Hi David, I think you are fighting a principle of physics, i.e.that light waves tend to diffuse when passing through a small iris....suggest you use a faster shutter speed and larger aperature. This principle is not related to the quality of the lens...it just happens. Good shooting

-- George L. Doolittle (geodoolitt@aol.com), April 03, 2002.

What Steve said. Among M lenses, the 135mm models should suffer the least from diffraction limiting. Since D.L. takes the edge off the lenses' resolution anyhow, maybe an economical old F/4.5 Hektor would fill the bill for $140.00 or so.

-- Bob Fleischman (RFXMAIL@prodigy.net), April 03, 2002.

david -- a response that hews more closely to the precise question you asked: assuming that the summi AA is sharper than the elmarit m at 5.6, but the elmarit m is sharper at 8 and all smaller aps, there is only one possible explanation. at 5.6 all (or more likely almost all) residual aberrations are fully corrected in the AA at 5.6, stopping down will not improve resolution. however, diffraction will diminish resolution as you stop down, so the AA will take the full diffraction hit the more you close up. in the elmarit m, it must be the case that all aberrations are NOT corrected fully at 5.6, and additional resolution is achieved by further stopping down. this further improvement beyond 5.6 mitigates the diffraction effects in the elmarit m (which the elmarit m will suffer to precisely the same degree as the AA because the two lenses have the same focal length and iris design), yielding better numbers at smaller aps than the AA. the fact that the summicron reaches maximal sharpness at a wider ap than the elm m, but then is less sharp than the elm m stopped down is, of course, no surprise. fast lenses are always designed for optimal perf wide open (or at least in the wide range). why bother to build a 2 when you already have a 2.8, unless you optimize it for speed?

-- roger michel (michel@tcn.org), April 04, 2002.

one last thing: i won't debate the point, but i KNOW that i get sharper pics with my leica as compared to slr systems. (a fact that was driven home yesterday when i developed a bunch of negs -- some from my f2/45mmP [a nice combo], some from my iiig/35 asph). however, i don't believe it entirely (or even mostly) due to the fact that leica lenses are sharper than nikkor optics. rather, i think the fact that RF focusing is, in many situations, a lot more accurate than slr focusing has a LOT to do with it. it is the whole leica system -- good rf body coupled to shrap lenses -- that accounts for the "leica advantage" (sorry mr. wildi). anyway, this is a long way of getting to another reason leica is not an especially good choice for landscape. in general, you are shooting at infinity, making the rf focus accuracy advantage irrelevant. even if you focus closer, you have gobs of time to get it rite. this is not to say, of course, that leica will be any worse than an slr for landscape, it's just that the edge that it seems to have over other systems for many types of action/low lite/candid photography will not translate into superior results for landscape shots.

-- roger michel (michel@tcn.org), April 04, 2002.

David: Some pretty good remarks here so far - here's some more food for thought.

The 2 Leica 90s have very different optical designs - 5 elements vs. 4 is just one example of the differences - in addition the elements have different curvatures, different spacing, and are made out of different glasses. So the light passes over very different routes on it's way through the lens to the film in one lens compared to the other. The only real similarity between them is that they cast images the same size.

If you could 'stretch' the Elmarit's elements' diameters wide enough to turn it into an f/2 lens (without changing the curvatures/thickness etc.) then THAT f/2 lens should perform identically at each aperture to the "2.8" version - because the light ray paths are still identical. But that lens's image quality at f/2 would not be particularly like the APO's - and for the same reason: the light paths within the two lenses are just - different.

Closing down a lens doesn't change the routes the light follows through the glass - it just cuts off some routes altogether. The 90 APO is an extremely good lens wide open - so obviously its edge rays are well corrected - maybe even BETTER than the central rays (given the kinds of compromises lens designers have to make). In that case stopping down more and more leaves only the LESS corrected rays. But to know that for sure we'd need a Leica optical designer with his computer to explain what's going on.

It's the same reason one 35 f/2 lens will have a different 'bokeh' than another - how and where the light passes through the various bits of glass may be very different even though the size of the image cast (and the aperture casting it) are the same.

-- Andy Piper (apidens@denver.infi.net), April 04, 2002.

Thanks Andy. Between your post and Roger Michel's, I now have been given a theory that fits the observed data: At f5.6 the 90 Apo (a lens deliberately corrected by its designers to optimize wide aperture performance) reaches its peak of performance. At smaller apertures there is, perhaps for the reasons you suggest, no further improvement in correction of aberrations, but diffraction begins to take its toll, so its performance drops off somewhat. With the Elmarit, stopping down from f5.6 to f8 DOES further reduce aberrations. Here too, of course, diffraction takes its toll, but the performance loss due to diffraction is offset by performance gains from further reduction of abberations. The result is that the Elmarit slightly outperforms the Apo at f8, and very likely this slight advantage for the elmarit continues all the way to minimum aperture.

So it very well could be that the slight small-aperture performance advantage for the 90 Elmarit that is reflected in the Chasseur D'Images test results is a real phenomenon. It would follow that for the subject matter I shoot with my Leica the 90 Elmarit will be a better all-around lens for me than the 90 Apo-Summicron.

Of course this leads naturally to the question: for the reasons we have discussed, is it likely that the old 28 Elmarit has better small-aperture performance than the new 28 Summicron?

David

-- David Mark (dbmark@ix.netcom.com), April 05, 2002.