Macro Lens Tests at 1:1

[sebastianmoran]

<p>I've been using and reading about macro lenses for years. I've been especially interested in shooting at around 1:1 for flower center details, insects, etc. Also for slide copying and digital-camera-as-scanner.</p>

<p>The common wisdom I've accumulated about lenses in this range over the years:<br>

- Lenses are optimized for a certain range of magnifications (image:subject conjugates)<br>

- Most lenses are designed for normal shooting distances, not macro. You can focus close with tubes or bellows, but the performance is compromised.<br>

- Macro lenses are best at 1:3 or 1:5, even though they may focus to 1:2 or 1:1. Much better than "normal" lenses, but not optimal at 1:1.<br>

- Enlarging lenses are designed for typical print sizes. Thus, they make great macro lenses reversed for say 5:1, but not optimal at 1:1<br>

- A few lenses are optimal at 1:1 (my favorites: 75mm APO Rodagon D 1:1, Olympus 80mm bellows lens)</p>

<p>Just now, I've <a href="http://coinimaging.com/Lens_tests.html">discovered a site</a> (coinimaging.com) with tests of a number of lenses from 1:4 to 1:1 and beyond. The site is about lenses for shooting coins. He's run real tests using IMATEST and the results are interesting:<br>

- The big performance change when lenses are used outside their design conjugates is in corner performance. <br>

- The 50mm f/2.8 Micro Nikkor is best at 1:4 or 1:2, so it's great reversed for higher magnification than 1:1. With this lens at 1:1, the APSC corner resolution is down to half the center resolution. Ditto or a little worse for a very good enlarging lens (Schneider Companion-S 50mm f/2.8).<br>

- My two favorite 1:1 lenses are indeed optimized for 1:1 where they have very good sharpness across the APSC frame. The APO Rodagon D is tightly optimized for 1:1; it's corner resolution is down by 50% at 1:2.5 and also at 2.5:1. The Olympus 80mm Bellows lens is engineered for a wider range, down by 20% at 1:2.5 and also at 2.2:1. I don't want to copy his graphs, but take a look at the 3rd graph on <a href="http://coinimaging.com/aporod75-4var.html">this page</a>. </p>

<p>CoinImaging.com uses MTF50 for "Sharpness", MTF10 for "Resolution", and measures corner performance as "% loss is sharpness vs. center". In all this, 2:1 is an image 2x the subject, 1:2 is an image half the size of the subject.</p>

<p>What do you think of this testing? Can you point to other sites with good tests for macro lenses?</p>

[Tim_Lookingbill]

<p>Do you define 1:1 according to how an actual subject such as a coin fits within the format frame size such as 35mm frame or APS-C digital sensor which has a long end length of about 7/8 in. which means the actual coin must be at least 7/8 in. across?</p>

<p>That's how I define 1:1 and I can get that with a Sigma 70-300mm macro zoom and Quantaray 2X teleconverter. I have to be at least 3 feet from the subject.</p><div></div>

[bobatkins]

<p>Rarely, if ever in nature photography are you shooting flat subjects and most of the time you're stopped way down, so corner to corner sharpness at large apertures (or even at small apertures) and an ultra flat field are rarely necessary (especially if they come at a high price!).</p>

<p>I'm sure the coin imaging website is good and their results using imatest should be quite accurate, though they may not be as important for nature photographers as for those shooting images of flat subjects such coins, banknotes or stamps.</p>

<p>I don't know any other websites that specifically concentrate on macro lens performance at high magnification.</p><div></div>

[Gup]

<p>Thanks for the link, Sebastian. <br>

I purchased an AF-S 105mm VR earlier this year as my first macro lens. I'm finding the learning curve much more difficult than I would have thought. The number of 'fails' is higher than I've ever experienced. I start to suspect the lens isn't functioning correctly and then I'm rewarded with a pleasing image and am left wondering about my own technique. It's the focus that always seems to be the problem. I usually use spot metering and focus with that single point and often find later the D800E focused behind or in front of my intended target. It never happens with any other of my lenses.<br>

The lens was purchased used, but only weeks old and with a receipt and warranty from a guy that seemed straight up. There isn't a mark on it so I doubt it experienced a serious bump of any kind. I guess I just need to work with it more. <br>

Has anyone any similar experience with this combination? Or advice? (No, I haven't performed any focus fine-tuning with it.)</p><div></div>

[bobatkins]

<p>For macro work it's often better to manually focus using whatever screen magnification your camera gives you. Since DOF is so small it's easy for AF to lock onto the wrong part of the subject, even if it's accurate.</p>

[Gup]

<p>Another</p><div></div>

[Gup]

<p>I felt that auto focus was probably better than manual when hand-holding to counteract any fore and aft movement on my part, but I guess that movement can still occur after the shutter has locked focus and before it has tripped.</p>

[sebastianmoran]

<p>Tim, yes by 1:1, I mean the image on the sensor is the same size as the subject. Just as you say.</p>

<p>For full frame, that's a 1x1.5 inch subject filling the frame. For APSC, that's a postage stamp or a medium sized button.</p>

[sebastianmoran]

<p>Bob, good point about field flatness. I wonder if the sharpness loss he measures at edges is due to non-flat field or to other sources? </p>

<p>In macro, the DOF is so thin, I'm usually doing some kind of focus stacking, at least for static subjects. So, I am interested in image quality at the edges. With stacking, non-flat field isn't a problem, as long as one of the slices is sharp.</p>

<p>Macro on moving subjects is tougher yet. </p>

[sebastianmoran]

<p>Gup, keep at it, macro is difficult.</p>

<p>I'll bet the focus problem you're experiencing is due to camera movement between focusing and exposure. You can easily check to trying a couple on a tripod. </p>

<p>I have not seen mis-focusing with my slightly different 105 Micro Nikkor (I have the non-VR version).</p>

[stephen t]

<p>Ditto the manual focus. That's what I have found best even with my auto focus macro Nikons. Save the autofocus for more distant subjects.</p>

[Laura Weishaupt]

<p>Gup, it's difficult to hand hold at 1:1 or close to and keep the focus where you want it. Worse yet, effort on a tripod can be thwarted by the slightest breeze. :-) There are some folks on pnet who are very good at hand holding macro work, so it's not impossible. Try using a tripod and manually focus in live view. Fine tune the focus at 5 and 10 x in live view. If you don't have a cable release for the shutter, then set for a delay. Work on static subjects to get the hang of things. Be sure to turn the VR off. </p>

<p>Sebastian, you can check out <a href="http://extreme-macro.co.uk/lenses/">this site</a> from U.K. The link goes to the lens page. I don't know if it's exactly what you're looking for, but they have a lot of information on macro lenses. I agree with Bob in that requirements for photographing coins and flat items are different than those of plants, bugs, crystal formations, fungi and so on. As to your question of what I think of the testing: I don't know enough about it to render an informed opinion. I have 2 dedicated macro lenses. Both are stellar and do what I ask of them in photographing nature. </p>

[sebastianmoran]

<p>Laura, thanks for the comment and the link. I'm more interested in performance around 1:1; your link is about "Extreme Macro", more like 3:1 or 5:1. Site looks like good advice for that. </p>

<p>Enlarging lenses and conventional macro lenses, when reversed, are good at 3:1 to maybe 6:1. Of course, photography gets more difficult above 1:1 (DOF, getting enough light, diffraction, holding anything steady).</p>

<p>Laura, I'm curious what are your two stellar macro lenses?</p>

[bobatkins]

<p>Edge and corner sharpness depends on field flatness, aberrations and aperture. All 3 contribute. They also contribute to center sharpness but in that case field curvature doesn't enter the equation because it's zero at the center of the field by definition.</p>

<p>Edge and corner sharpness are mostly determined by field curvature, coma and astigmatism. Luckily the effects of all three are reduced at smaller apertures. Coma is actually reduced and the increased depth of focus mitigates some of the effects of astigmatism). Chromatic aberration can also be present (and usually is) - but it's not really reduced by stopping down. Field flatness itself is unaffected by aperture, but again stopping down reduces its effect on sharpness by increasing the depth of focus (which is the depth of field translated from the object plane to the image plane). Stopping down also reduces spherical aberration if present. It has no effect on distortion (rarely a problem for nature work).</p>

<p>If you're not shooting flat subjects and you stack exposures for increased DOF then field flatness may not be a concern. Imatest doesn't measure what's causing any image degradation, it just measures resolution. You could attempt to measure field flatness but it's not built into imatest.</p>

<p>Don't forget that if you are stopping down to get greater DOF, resolution will be reduced by diffraction. With a good lens diffraction softening is often detectable at f8 and it's very obvious at f22. If you're at f16 or f22, much of the difference between a really good, very expensive, lens and a modestly priced lens may well be lost. For shooting flat coins and stamps you can take advantage of maximum sharpness (usually at around 5.6) of a really good lens, but for nature macro work that's often not practical due to DOF considerations.</p>

<p>Note that aberrations can change with focus distance (magnification) and may be significantly affected (increased) if extension tubes are used.</p>

<p>The bottom line is that for a lot nature macro work, relatively inexpensive lenses (or even extension tubes and achromatic doublet closeup lenses) can give results close to those you'd get from an expensive 1:1 macro lens. That's not true if you're shooting coins, stamps, banknotes or other small, flat, items.</p>

[Laura Weishaupt]

<p>Sebastian,<br>

I use the Canon EF-S 60mm 2.8 and the Canon MP-E65.</p>

[sebastianmoran]

<p>Bob, thanks for this and for all your many contributions.</p>

[paul_de_ley]

<p>Interesting website! On the downside, the selection of "true" macro lenses is restricted to a few lenses for Nikon mount and doesn't as yet include third-party offerings like the Tamron 60 f2 and 90 f2.8 macros, Sigma 70 f2.8 macro etc. Hopefully those will be addressed in future updates.</p>

<blockquote>

<p>Macro lenses are best at 1:3 or 1:5, even though they may focus to 1:2 or 1:1. Much better than "normal" lenses, but not optimal at 1:1.</p>

</blockquote>

<p>Like Bob writes above, that's not a very meaningful statement if it doesn't also specify which particular optimality criteria are employed. And even if relative corner sharpness in a strictly flat-field application is the main criterion, then it's not universally true that macros perform better at 1:3 through 1:5 or so. Although most of the tested Nikkor micros suffer very noticeable losses of corner sharpness at highest magnifications, Sigma's 50 f2.8 macro does best at 1:1.6 through 1:1 in this regard, and I would expect some of the other 3rd party macros (not to mention macro lenses available only in other mounts) to behave quite differently from each other and from the tested lenses too.<br /> <br /> As to Gup's question and comments, compared to manual focusing I for one have never gotten remotely as good percentages of well-focused shots using AF at magnifications around 1:2 or higher. I suspect AF isn't just inappropriate because of the nearly inevitable slight forward/backward movements by camera and/or subject, but also because the great majority of macro subjects have various types of repetitive surface patterns (eg folds and veins in petals or leaves, punctations hairs or scales on insect bodies etc) that become just contrasty and structured enough to distract AF algorithms at near-1:1 magnifications.</p>

[sebastianmoran]

<p>Paul, thanks for the note. I do agree that macro lenses can differ. </p>

<p>On AF, interesting comment on surface patterns.</p>

<p>For Gup's question, hand-holding with the 105VR macro, I have two suggestions: 1) Try, if you can, to fix the camera and subject positions (tripod for the camera, ? for the subject). 2) The second fix that's worked for me is to shoot many frames, say 10+, then pick the sharpest. </p>

[Gup]

<p>I appreciate what everyone has said and agree with the advice. It's only hand-held shots that I have been disappointed with because I do use manual focus and a cable release when the camera is mounted, and live-view when the screen can actually be seen (I find them useless in the sun). As I mentioned, it has been a learning exercise not to trust the viewfinder image during auto-focusing. <br /> Does anyone use mirror-up for macro work? <br /> I bought the lens on the way to the airport for a vacation in Costa Rica so didn't have any experience with it or practice time prior to that, thus the many discards I found later when I returned. I also haven't used it for any other purpose yet. It's the only lens I own with VR, too, and I haven't even tried that feature. I expect hand-held macro work will be my only complaint and now I realize that may be asking a little too much of us both!</p>

[Gup]

<p>*</p><div></div>

[Ed_Ingold]

<p>There is an whole class of lenses designed for 1:1 (or greater) reproduction that seldom enter the field of general photography. They're called "process lenses", used primarily for engraving printing plates. Before "Xerox", they were used to reproduce documents in general. I used one, mounted on a Speed Graphic, to copy photographs submitted to a newspaper for publication. At that time, our photo engraving was limited to 1:1 contact prints using negatives. Even then, 1:1 was rare, except for wallet-sized pictures submitted for obituaries.</p>

<p>I think the need for 1:1 magnification in general photography is exaggerated. Most natural subjects, including insects, are seldom photographed more than 1:2, and 1:4 is probably the most common ratio. Until I began "scanning" slides with a digital camera, the times I needed 1:1 reproduction over a span of nearly 60 years could be counted on one hand. Aside from coin and stamp collectors, most people could be happy with less.</p>

[sebastianmoran]

<p>Edward, thanks for the note. </p>

<p>I agree that 1:1 is an uncommon need. But somehow, it got my curiosity. 1:2, 1:4 or 1:6 is well-served by macro and enlarging lenses. And, 2:1, 4:1, even 6:1 by reversing these lenses. I got curious about how to do images at 1:1.</p>

[Matthew Currie]

<p>A quick note on process lenses. I have had good luck with "Compugraphic" lenses, which were once used in typesetting machines, and tend to show up from time to time, often in multiples on a turret. It happens that most of these will fit into a microscope adapter, which makes for a quick and dirty macro mount, utterly lacking in focus or aperture control. You simply move the camera. Various powers can be had, and can be varied some by placement in the adapter. In general, a microscope adapter is a very handy thing to have if you like to experiment.</p>

<p>Because it is so light and short, this rig is actually relatively easy to shoot hand held. The aperture seems to be around 5.6 for most of the lenses, which is doable with an SLR. It's still hard to get a sharp image, but the lenses themselves are quite good. Here's a lucky shot done with a D3200 and one of the relatively shorter Compugraphics, "straight from the camera." </p><div></div>