Never Do a Lens Test Again?

[Peter_in_PA]

<p>Okay, I'm just kidding. I don't like doing lens tests over and over, but when I get new optics, I test them to be sure they work well, aren't de-centered, and so I can learn how they work best. So, yes... do lens tests to make sure you got a quality copy for sure... but...</p>

<p>I recently was evaluating a couple new optics I got, and I came upon a startling (but probably obvious) revelation.</p>

<p><strong>Every single zoom I've ever tested is at its absolute sharpest and with the best corner performance exactly 2 stops down from whatever wide open is... period.</strong> Lens I've tested include tokina 11-16, Nikon 18-200. 18-70, and 70-300 VR, and now a couple little Olympus µ43 optics, too. There has never been a variation in that, and I'm tempted to just no longer even look at them, except to make sure they work. Every one probably cleans up JUST the same amount 1 stop down, and is perfectly usable there, too. I've tested lenses on 6MP, 12MP and 16MP cameras only.</p>

<p><strong>Every prime I've ever tested</strong> (including Nikon 35mm AF-S G DX, 50mm AF-D f1.8, 55mm f3.5 micro AI, and 105mm f2.5 AI) i<strong>s equally at its best 2 and 3 stops down.</strong></p>

<p>And whether or not it's 2 stops down, <strong>f8 is very good on every lens I've ever tested</strong>. (Yeah, we already knew that, didn't we...) <em>(Also... diffraction is not really a big deal on most images the way we actually use them, but that's another discussion.)</em></p>

<p>I realize that for the most part my tests are of more "consumer grade" optics... but... Is this just the laws of physics? Is this always going to be the case?</p>

[eric_arnold]

<p>actually common characteristics makes evaluating lenses easier. knowing where the sweet spot is on lenses is half the battle.</p>

[LenMarriott]

<p>Peter, In general I agree. A good 'rule of thumb' to follow. While it may not be exactly true in every case, following it is not likely to lead too many of us too far astray. Diffraction is the compromise one must make only if maximum depth of field is desired\necessary and, in my experience, not worth getting too anal about. The masters among us may not agree. Each of us has quality parameters we will not violate. Mine, good, bad, or ugly, are posted for all to see. :-) Best, LM. </p>

[Palouse]

<p>Good observations and advice Peter. Pretty much what I subscribe to, as it keeps me sane and prevents paranoia. If I don't have a great reason to open up wider than 2 stops from max, or go beyond f11, I <strong><em>usually</em></strong> don't. I say usually, because as a mere human, I do screw up a lot! But I do love those large apertures to help me focus and compose.</p>

[Rick Waller]

<p>" <strong>f8 is very good on every lens I've ever tested"</strong><br>

<strong> </strong><br>

<strong><br /></strong>Think that may explain why, when someone asked the famous photojournalist from the 30's and 40's how he managed to <strong>always</strong> get the great shot, Weegee is reputed to have replied, "F8 and be there".<br>

:)</p>

[sjmurray]

<p>I was more than a little surprised when I began testing my Nikon 18-105 kit lens that came with my D7100. It performs best at all focal lengths, even in the edges, when wide open! I'm glad for this because I shoot a lot of natural light situations, requiring as large an aperture as possible. I also recently did some careful examinations of distance shots with my older Nikon 18-70 kit lens on the D7100. With my D70 and D80 I always used f 8 or so for landscape type shots. Detail wide open at 18mm on distant scenes, was lower in contrast than at f 8, but the detail was just as good. The fine detail in some shots was almost better, but as mentioned, the contrast was lower wide open. The D7100 does have a much greater pixel density than my older cameras and allows closer examination of sharpness, I believe. Both lenses have their "sweet spots" at various focal lengths and distances, which is not surprising considering the complexity of the design and individual variation coming out of the factory. My experience with many prime lenses over the years is similar to the OP's.</p>

[alastair_anderson]

<p>Some modern lenses are designed to perform best wide open. (Actually I'm mainly thinking of Leica F2 primes).</p>

[phule]

<p>[[Never Do a Lens Test Again?]]</p>

<p>This is the best advice I've read on photo.net in years.</p>

[Wouter Willemse]

<p>Peter, I think in general your rules of thumb are about right, but there sure are exceptions, and I think for primes you are a bit too conservative. Especially many modern primes are already great much much earlier for 2-3 stops down - more 1 stop down max (for example the Sigma 35 f/1.4 Art, Nikon 58 f/1.4G).<br>

Even some older ones - my 2 copies of the 105 f/2.5 do not get seriously better going from f/4 to f/8. My 50 f/1.2 is great at f/2 and (to me) does not get a lot better afterwards. My 35 f/1.4 AiS does get worse at f/8 and is better at f/4 or f/5.6... So, it's not law of physics, and I think especially with the more specialised tools (which fast primes are) you really need to get to know your gear. Not shooting brick walls, but actual real-life scenarios, study the results and get grips with how these lenses behave. Just as important it can be to know where their sweet spot is, it's also pretty useful to know how they behave at the other apertures (especially the superwide apertures).<br>

That all said, your simpler rules make an excellent starting point to start learning that behaviour as you'll more quickly get an idea of the best performance, making comparison easier.</p>

[Peter_in_PA]

<blockquote>

<p>So, it's not law of physics, and I think especially with the more specialised tools (which fast primes are) you really need to get to know your gear. Not shooting brick walls, but actual real-life scenarios, study the results and get grips with how these lenses behave.</p>

</blockquote>

<p>Could NOT agree more.</p>

<p>I spent a morning doing lens tests... I know guys who spend a week shooting charts and brick walls. I plan to shoot real world stuff and see how it goes.</p>

[ShunCheung]

<p>The problem is that I frequently do not use apertures where the lenses are best optically.</p>

<p>A very simple example is the 80-400mm AF-S VR. I almost always use that lens wide open at 400mm, f5.6 to photograph birds. Even so, I am frequently forced to use ISO 400 or 800 to get a shutter speed faster than 1/1000 sec to freeze action under plenty of daylight. Using that lens at f8 simply doesn't make sense because my shutter speed would be too slow and/or my ISO would be high.</p>

<p>I also tend to use the 85mm/f1.4 AF-S at f2 a lot, not f4 or so where it is sharper, in order to get the shallow depth of field I want to achieve.</p>

[Peter_in_PA]

<p>Shun,</p>

<p>Yup... those are good reasons to test and learn. I also shoot my 70-300 VR wide open at f5.6 all the time. But I learned it worked well by shooting real stuff (not test charts) with it, too.</p>

<p>As far as the primes thing goes, I've also learned that all the primes I've tested are really very good one stop down. Very usable.</p>

<p>But the whole "soft wide open thing is often a smokescreen for bad technique or the simple physical fact that depth of field is razor thin. And is pushing ISO to 400 or 800 on the cameras we have now really even "pushing" anymore? My old D90 looked just as good (once you size for viewing or print) at 400 or 800 as it did at 200 anyway.</p>

<p>I guess the most important thing to learn anymore might be where diffraction sets in. For me, since I rarely am using stuff at full-pixel viewing, I've even stopped all the way down to f22 and gotten useable results.</p>

<p>I guess I'm saying we worry too much.</p>

[ShunCheung]

<p>For example, I posted the image below to Nikon Wednesday last week, and I used the 85mm/f1.4 AF-S at f2 on purpose. I used spot metering to favor the subject, but I couldn't completely avoid the guy in shorts on the right side of the background.</p>

<p>Sometimes with macro, I use f11, f16 or even f22 to compensate for the very shallow depth of field due to the close distance/large magnification, even though I know very well that diffraction is going to rob some of the sharpness.</p>

<p>A good, experienced photographer will know how to compromise with the camera and lighting settings to achieve the desired effect. "f8 and be there" might work a century ago when cameras were simple, ISO 32 (or even lower) was the norm, and action photography was mostly not possible. Today, a lot of people can capture decent, acceptable images with a smart phone to post on social media. To capture great images, it still requires hard work and experience.</p>

<P>

Plenty of junk lenses can work well enough at f8, hence testing is not really necessary. When we check out an expensive lens, I tend to test it at the extremes, i.e. wide open, the longest and shortest focal lengths for zooms ..., as people pay a lot of money for their performance at extreme settings.

</P>

<CENTER>

<P>

<IMG SRC="http://static.photo.net/attachments/bboard/00c/00ch8X-549629884.jpg">

</P>

</CENTER>

[keirst]

<p>I’ve had this misfortune of owning a couple optics that were not good at f/8, at least not into the corners. The 35mm/2 AF Nikkor comes to mind. Mine had not even 1 good corner at f/13 unless focused near its minimum focus distance. Only a few lenses are really good center to corner wide open, like the 55mm/1.4 Otus and the 135mm/2 Apo-Sonnar, but depth of field can be a battle with my 135mm, so I rarely use it wide open. I’m not interested in the Zeiss 55mm.</p>

[elliot1]

<p>While typically your observation is quite accurate, I have found at least one exception - the Lumix 100-300mm lens (a m43 lens). After careful testing, I found my copy to be as sharp wide open as when stopped down a couple of stops. </p>

<p>Years ago I purchased DXO software with the hope that is lens specific softness correction feature would provide significant improvement for the consumer zoom lenses I owned when shot wide open so I would not have to stop them down to f8 and sure enough, their software offers noticeable improvement in this area.</p>

[john_narsuitus]

<blockquote>

<p>"... I know guys who spend a week shooting charts and brick walls. I plan to shoot real world stuff and see how it goes."</p>

</blockquote>

<p>I normally spend a lot of time shooting charts and brick walls because I do not like to be surprised while I am shooting real world stuff.</p>

<p>One lens I tested, a Tamron zoom, produced images that were warmer and one-stop darker than my Nikons lenses. However, I was surprised when I discovered that the real world images shot at the 200mm focal length were actually comparable to images shot at 150mm with my Nikon lenses.</p>

<p>Now, I have added a focal length test to my brick wall and chart tests.</p>

<p> </div>

[dj_soroka]

<p>I found my 180 mm f/2.8 AIS lens best one stop down, at f/4.</p>

[sjmurray]

<p>Shun makes a good point. Regardless of the lens's sharpest aperture, I use any lens for a total effect, and not just "sharpness." This includes softness of background, depth of field, low light needs, etc. Often these overall needs have nothing to do with absolute sharpness. In the 60's I shot a lot of indoor documentary type shots without a flash. My usual formula was Tri-X, D-76 1:1 developer and my Nikkor 50mm f 1.4 wide open and 1/30 sec. It never let me down. Now, I knew this was not the optimal for sharpness, but I got my shots, and they have a certain look and feel to them because of my formula. </p>

[sjmurray]

<blockquote>

<p>Now, I have added a focal length test to my brick wall and chart tests.</p>

</blockquote>

<p>I agree. I have found that my zoom lenses can have different characteristics at different focal lengths. Simply shooting a wall chart doesn't tell me what a lens will do, say, at 18mm at different apertures shooting a woods from a distance, compared to shooting indoors at living room distances. At 35mm, my kit lens becomes a stellar performer wide open at any distance, and at infinity it is sharp at all focal lengths. At 70mm it is slightly softer on the edges at any aperture. My 18-105 has completely different characteristics.</p>

[glen_h]

<blockquote>

<p><strong>Every single zoom I've ever tested is at its absolute sharpest and with the best corner performance exactly 2 stops down from whatever wide open is... period.</strong></p>

</blockquote>

<p><strong><br /></strong>Sounds about right. For one, as you stop down more you run into diffraction limits. <br>

Also, there are a lot of compromises in lens design, and they tend to show up more at larger apertures. <br>

<br />Reminds me of a rule on reading graphs on technical papers. Ignore the first and last point on the graph. If the data was really good at those points, then they should have been able to get another point on each end, that wasn't quite as good. (And watch out for graphs with only three points.)</p>

<p><strong> </strong></p>

[bob_bill]

<p>Funny, only yesterday a client asked me why I didn't close down a stop for added sharpness... not. </p>