Article on sample-to-sample lens quality variation published by dpreview

<p>I don't know if this article will convince us to worry less or more about we have "a good copy of a lens," but it seems reasonably well done.</p>

<p>http://www.dpreview.com/articles/7333489584/variation-facts-and-fallacies</p>

<p>It's about what I expected. I never worried much about this, and after reading that, even less. Thanks, Hector.</p>

<p>Nice of him to wrap this up like this. I've seen some of his earlier writings, along the way, and it's great to put it all in perspective. Obsessive pixel-peepers, though, frequently to their thing for its own sake, and there's no pleasing them.</p>

<p>I know all along that so called "sample variation" is mostly nonsense. There are some Nikon lenses that I get to test with multiple copies. For example, I have tried 6 different samples of the 24-70mm/f2.8 AF-S, and the are remarkably similar. Earlier this year I accidentally ordered two copies of the TC-20E III; in the few days I had both with me, I tried them on various lenses side by side, and they show essentially identical results under pixel peeping.</p>

<p>In my case, I only checked those lenses with my eyes on images on the computer screen and prints. Even though I have bertter than 20/20 vision wearing glasses, I can only distinguish so much fine detail. Roger Cicala, the author of that article, runs LensRentals.com so that he regularly handles multiple copies of the same lens models and uses machines to analyze minute differences among samples, tiny variations we humans cannot observe.</p>

<p>Another issue he points out is that once you take a lens apart, you need to use precision equipment to caliberate it when you put the lens back together. That is why it is unwise to take any lens apart yourself to clean the middle element or replace a scratched front element. It may seems easy to unscrew the front element and put another piece of glass on for some lenses, but if you lack such equipment and/or knowledge to caliberate the lens afterwards, your lens will likely have problems.</p>

<p>Nice to know that all this pixel-peeping is probably not too relevant. The proof is in the pudding, as the friendly chef might say. In any case, there are probably more differences in the performance of older age lens formulas and many of the better present ones, not to mention all the other aspects than sharpness that affect a lens performance and show up differences between competing same lenses of different manufacturers. In practical terms it is unlikely that great differences occur and the weakest link is not the lens or the camera but the user, and the quality of his practice.</p>

<p>One reporting that did impress me some years ago was Ctein's review in a photography review of enlarging lenses (subsequently updated, it appeared in a chapter of his darkroom manual that I believe was called "Post Exposure"). He showed noticeable differences in 2 or 3 copies of the same lens from each manufacturer and the considerable variation of the different lenses tested (some 10 or 20 different models/makes) in performance, including light uniformity. Getting a lemon, at least some 20 years ago, was not uncommon, according to Ctein. I found the article useful in finding a better lens that later in practice showed considerable improvement in diminishing light fall-off and poorer resolution in the edges of large prints (greater than 10X magnification). Was my new lens as good as others of its contemporary series? No idea, all I know is that it gives fine results for me and that is the bottom line.</p>

<p>One thing I noticed in his article: He mentions that autofocus sysyems are not perfect and I suspect that he considers that to be more important than lens to lens variations. In attempting to overcome that problem (or perhaps minor problem?) he presents the following suggestion that perhaps knowledgeable DSLR users might enlighten me and perhaps others on:</p>

<p><em>"When you buy a lens, and assuming your camera allows you to, you should microfocus adjust it. If you do it properly, using a sensible focus distance, it really does make a difference. Then do some very basic tests to make sure it functions properly, and go take some pictures. If you like the pictures it makes, then keep it."</em></p>

<p>I am not familiar with "microfocus adjustment." Is it something available only on high level DSLRs? (My system cameras are manual focus, and a basic fairly unrefined AF is available to me only on a small digital P&S).</p>

<p>Good comment, Shun, in regard to lens disassembly. Better left to those with sufficient experience/equipment.</p>

<p> </p>

<p>I've always thought how silly we are/I am to base a major lens buying decision on one test of one lens sample in one magazine. Just once I would like to see three samples randomly picked off the assemble line and compared.</p>

<p>Thanks for the link, very informative.</p>

<p>My G1 has a micro focus assist feature that has been or aid many times to me, especially when shooting flowers. It is particularly useful when I test lenses on the body thru adapters. In this way I can determine whether my 50mm Nikon, Canon, or K lens is the sharpest over it's scale and what apertures are better. I set up a leveled steel ruler with markings for millimeters and 64th's and shoot files at the cameras highest files size which I open in Elements and enlarge 200% or more. The camera lens is of course leveled as well. With my eyes, and the fact that you cannot go by EVF or LCD focus, it is a feature I wouldn't go without unless shooting an old film camera, obviously.</p>

<p>I saved the review in a pdf file, but dropped off the comments some of which seemed to be "I know what I know, so don't try to refute it with mere facts" - a line of argument all too familiar these days.<br>

Indeed, thanks for the link.</p>

<p>Have the same thoughts expressed here about lens quality.</p>

<p>This is why I buy cheap (under $100) lenses, shoot Raw and fix in post which is quite easy and quite entertaining to see what can be done with software. I've come across some surprising finds (in image quality=color/sharpness) with legacy film lenses for my Pentax K100D. Frankly the 35-105mm Pentax AF suits me so well I don't even think about what it would be like to buy and use a much more expensive and modern lens.</p>

<p>At least if I drop the darn thing or it just goes south, I'm not out the $1500 Il'd have paid because I equated price to perfect image quality. The majority of the high price goes into materials and workmanship so that lens can be dropped and still be useable. I'm assuming of course because I've never bought or checked out a $1500 lens.</p>

<p>The bit I found most interesting is that the shots using magnified Live View are as good/better than simply re-tripping the AF (focused & locked) shutter and considerably better than repeated camera AF on a perfectly lit target. </p>

<p>Excellent article.</p>

<p>Maybe this article should've been written when they were designing the lens and mirror system of the billion dollar Hubble space telescope. I guess everyone knows how that went down.</p>

<p>And here we're complaining about slight variances on dinky lenses on much less expensive DSLR's by comparison.</p>

<p>How's that for a perspective?</p>

<blockquote>

<p>I am not familiar with "microfocus adjustment." Is it something available only on high level DSLRs? (My system cameras are manual focus, and a basic fairly unrefined AF is available to me only on a small digital P&S).</p>

 

</blockquote>

<p>On cameras with phase detect focusing the focus sensor is located below and in front of the image sensor. if the focus sensor is not exactly in the correct position the camera may focus slithly in front or behind of the focus target. When the camera is calibrated at the factory they adjust the focus sensor. However the adjustment at the factory is done assuming your lenses are also within specification. If a lens is a little out of factory spec the same problem (front or back focus) would occur but it would only happen with one lens while other lenses would work fine.</p>

<p>micro adjustment is a firmware setting available on most DSLR introduced in the last 3 or 4 years (older cameras may not have it). It allows the user to make small adjustment to the camera focus system that will only be applied if one particular lens is attached to the camera. </p>

<p>Most point and shoot cameras use contrast detect auto focus which uses the imaging sensor on the camera and software to determine correct focus, instead of a separate phase detect focusing sensor. Using the imaging sensor completely avoids the calibration issue and therefore microadjustment is typically not possible on point and shoot cameras. Unfortunately contrast detect focusing is typically much slower than phase detect focusing.<br>

Even if your camera focus sensor is properly calibrated at the factory it is possible that small variations in the shape of lenses and manufacturing variation in the lens assembly can cause some small focus variation </p>

<blockquote>

<p>Maybe this article should've been written when they were designing the lens and mirror system of the billion dollar Hubble space telescope. I guess everyone knows how that went down.<br>

And here we're complaining about slight variances on dinky lenses on much less expensive DSLR's by comparison.<br>

How's that for a perspective?<br>

 </p>

</blockquote>

<p>The Hubble space telescope has a  very good manual system No mater how hard they tried they couldn't fix the problem by adjusting the focus.  The hubble problem was not a simple calibration issue.  <br>

 <br>

The main mirror of the telescope was made to the wrong shape due to a problem with the equipment at the company that made the mirror that wasn't discovered until after the launch.  The only way to fix that issue was to install a corrective lens.  That is the equivalent to correcting your eyesight with a pair of glasses.</p>

<blockquote>

<p> <br>

 </p>

</blockquote>

<p> </p>

<blockquote>

<p>I know all along that so called "sample variation" is mostly nonsense.</p>

</blockquote>

<p>Well Shun, maybe everything you've ever purchased has been perfect with no flaws, but for many of us this has not in fact been the case. There are always going to be defects coming out of a factory. I've experienced it with cars, cameras, and camera lenses. This is generally why there is a warranty. </p>

<p>Case in point. I owned two different Nikon 16-85mm VR DX zoom lenses. The first one was purchased by me shortly after they became available. It was a good performer optically, but the focus motor made a little squeaking sound. I later sold it when I bought a full frame sensor camera. Later I bought another 16-85 VR zoom when I had a DX format camera. This second 16-85 had a couple flaws. One was the focus indicator on the top of the lens was off when the lens was set to 16mm and focused to infinity. It indicated 3ft of focus, though the images were sharp. Then also every photo I took with this lens while framing vertically, was slightly out of focus. Why I never knew. That lens has also since been sold. </p>

 

<blockquote>

<p>I have tried 6 different samples of the 24-70mm/f2.8 AF-S, and the are remarkably similar.</p>

</blockquote>

<p>As that is a pro level Nikon zoom lens, I would hope this would be the case! I've never owned a pro level Nikon zoom (don't want to spend the money, don't want to lug the weight around while traveling) but I may someday. <br>

I've also owned multiple Nikon 35mm f2 AI and AIS manual focus lenses, and all have performed identically. Both Nikon 18-70mm DX zoom lenses I have owned performed equally well. </p>

<p>I know the Nikon D300 and D700 have the ability to perform focus adjustments in the menus, but I've not found any need to do this so far. I think I read too that it doesn't work with zoom lenses, as you can only calibrate a lens while focusing at one particular focal length. </p>

<p>Steven, many thanks for the heads-up on the micro-focus feature. Having a manual focus digital camera sysytem means that I had not encountered that terminology.</p>

<p>Roger, the author of the linked to article, actually owns LensRentals.com, he has <a href="http://www.lensrentals.com/blog">a great blog on</a> the site with many fantastic articles, including the one reproduced on preview, and the preceding articles. The series on lens genealogy really are great reading.</p>

<p>One clarification on the hubble mirror. The center of the mirror focus at one distance while the edge of the mirror focused at a second distance. </p>

<p><a href="../photodb/user?user_id=478782">Dave Lee</a> said: "Well Shun, maybe everything you've ever purchased has been perfect with no flaws, but for many of us this has not in fact been the case."<br>

The article claims that sample variation is small, but does not dismiss the fact that some lenses are defective. I see no reason to doubt the claims. Over the years I have only had one bad lens, a Nikon 12 to 14mm zoom, and it was unusable between 12 and 14mm at any aperture. Sample variation will be due to slight variation in the refractive index of glass, the curvature of lenses, the machining of lens cells and so on, whereas defects are due to gross errors such as inserting a lens element in reverse, or omitting a shim.<br>

Something he does not mention (but he does mention in another article) is that in his exensive experience the big names such as Nikon have much better quality control than companies such as Sigma.</p>