[Michael Piziak]

Do you test a new lens when you receive it? If so, how? Do you use the same subject, being very objective, as so you can compare your previous lens with, or do you take a more subjective approach and simply take the camera outside and shoot away at some things?


I have dabbled with testing my lens, on a very unscientific basis. I find myself "testing" new (or new to me) lens more and more when I receive them lately.


For a conversation starter, check out this article and please comment in this thread: How to Test Your Lens | B&H Explora

[Paul the Sunman]

No. Though I do try it out with different types of shots, to get a feel for it. If something appears amiss, then I might test it for front or back focussing, but that's very rare (contrary to the views of some people on PF who are convinced that all their lenses need major in-camera adjustments).

[Michael Piziak]

No. Though I do try it out with different types of shots, to get a feel for it. If something appears amiss, then I might test it for front or back focussing, but that's very rare (contrary to the views of some people on PF who are convinced that all their lenses need major in-camera adjustments).

I've never done the test for front/back focussing. With more modern lens, does the camera "remember" the setting for each lens?

[StiffLegged]

No, I don’t test a new lens; I go out and take photos with it. I’ll do any testing later should something be noticed in the photos taken.

FA and DA lenses have their individual af settings remembered by the camera (if you have to tune the af for a lens)..

[microlight]

Yes, cameras back to the K-7 of 2009 will remember up to 20 DA or FA lenses individually. There’s a debug mode in some cameras before this that also allow some fine tuning.

If you tend to shoot at maximum aperture a lot, you’d probably need to do this, for example with a macro lens like my FA100/2.8. If you shoot a wide angle prime or zoom at around f8 or smaller then the depth of field will likely tend to hide any focus error.

[tibbitts]

I've increased the testing I do with new-to-me lenses over the years. In the film days I did none at all, and it wasn't until I'd had my K100 and several lenses for a while that I found out it might be important. What brought me to testing was acquiring a new Sigma zoom that was giving images obviously soft on one side. I didn't notice that in a test, just in using the lens.

Since then I've acquired a number of both new and used lenses, and now I do take photos of test targets as well as larger subjects (brick walls, distant scenes with signs in them, etc.) My two DA* zooms have been very good. Well, except both of them were at or beyond adjustment range in PDAF, so 10 or more off (and that's identical on multiple bodies with different technology AF, so it's definitive.) One of them was new and under warranty, and after Pentax serviced it (without me providing a body to test on or saying in what direction it was off), it came back at 0 (well, not at every focal length or distance, but that's another story.) But I like the lenses because they're consistent across the frame at all focal lengths, and there's always CDAF (now that I have K5 or newer bodies.) I've kind of given up on PDAF with zooms in general.

I lost track of exactly how many new 16-85mm and 55-300mm (DA) lenses I went through trying to find decent copies. Some were truly horrible, soft on one side/corner or another, but always at only some focal lengths or distances. That's a huge frustration about testing zoom especially - you can practically wear a body out just figuring out if your lenses are any good. And after only a couple of exchanges you wear out your welcome at retailers.

I have a full-frame 70-300 Tamron, with the legendary purple fringing, and very inexpensive. I only bought one copy, used, and it was far more consistent than all but one of the 55-300s. Longer lengths don't give the most amazing performance, and it really wants to be stopped down, but that usually fits with how I've used it. My first Sigma 10-20mm f4 was horrible on one side at any of the wider settings. I now have two Sigma 10-20mms, one a replacement f4 and one the f3.5 version. Performance isn't perfectly symmetrical across the frame but they're pretty good - infinitely better than that first 10-20mm. I have three Sigma EX macro lenses (full frame of course) in various focal lengths that are close enough to being correct, at least on the APS I have, for me. I have a 17-50 Sigma that's not quite as consistent as the macros but it's a very good lens and I'm happy with it - much more consistent than the Pentax 16-85. The 16-85 is actually good except up close, so it's better for most real-life uses than shows on test targets. I did eventually test my original K100 lenses, 16-45 and 50-200, but stopped using them once I had better alternatives.

About procedures, although I've usually tested with a very heavy tripod, no SR, etc. I have to say that in almost every case I could just aim at a subject handheld, either use a very fast shutter speed or flash, and get identical results to what more controlled tests would show. Maybe not if I tried to count lines per mm or something but I don't do that. Really, it's fairly obvious if a lens is pretty good by my standards or not; the only difficult part is having to take SO MANY pictures, at so many distances and focal lengths. Lenses can look fine at some settings (focal lengths, subject distances) and horrible (as in decentered) at others. If I want to verify a test quickly I can always swap on one of my similar focal length macros or the DA* lenses. In every case they'll be pretty sharp across the frame; with those 55-300s and 16-85s I've sent back, not even close (at some settings, that is.) It's also possible to do things like flip the body over to see if a problem area follows, etc., as another way to verify quickly.

[Paul the Sunman]

I've never done the test for front/back focussing. With more modern lens, does the camera "remember" the setting for each lens?

I think any AF lens can be fine tuned in-camera, and up to 20 adjustments stored. You can also set a global adjustment that applies to all lenses, though that is unlikely to be a good idea across the board. The process is problematic for zooms. if you tended to use a particular FL most with a lens, I guess that's the one you'd compensate at.

[garywakeling]

Testing, not really, I do tend to use the lens in the monthly single in challenge, at least one photo a day and only using the single lens, it makes me try the lens with different subjects that the lens may not usually be used for. I try it on bellows, reversed, on extension tubes, with converters, crop and full frame, adapted to the pentax Q etc. Do I take lens charts, nope, but I certainly try to give the lens a thorough workout, sometimes you discover interesting uses for a lens way away from what your originally thought. Good luck with your own "testing"

[Papa_Joe]

I do test newly aquired lenses, but with a quite casual approach.

First thing is to test if the AF with the central point is spot on with my K-70 which has a very consistent and spot on AF. Lucky me. I look at those pictures in 100% magnification. I have never ever set a correction value for any of my lenses. Probably I am just to forgiving.

Next I mount the lens onto the camera and do as many pictures as possible with open aperture preferable. I watch those in full screen on my 24" monitor and if they seem to be OK every thing is fine. The reason for doing so is, that in my experience there are so many things that influence what we see as "in focus" that you can only go wrong with the 100% view. So often I was wondering if the lens was in focus and if I had to set a correction value. And when I then tested under controlled conditions everything was fine or I was not able to reproduce the of focus effect. As my pictures are never presented in a way where the viewer would get closer then the average viewing distance the full screen test is technically sufficient.

I have to admit that this test procedure does not show every flaw at once. In March last year I bought a Sigma 10-20 and tested it in the way described above. I just did not do al lot of pictures in the live test and found it a very good lens. In August we went into vacation an I brought back some hundred pcitures. I was still very satisfied until in full screen mode I recognized a smear in the outer left third of some pictures. Further testing (brick wall - acctually hut wall ) proved a probably severe decentering of the lens. My dealer sent it in for guarantee repair, but sigma was not abel to repair the lens, due to ceased spare parts.


Let me tell you something else about comparability of tests.
I have recently bought an other 10-20 from a member of our german forum. He has stated, he selected the lens after thorough testing from several pieces. My first center point AF tested showed acceptable results. Full screen test where decent. The decentered lens in contrary was crisp high contrast and colorfull from the first picture on I was stunned by this lens. While the replacement I would say is OK.
But there is one major differencen in the tests. The decentered version was allways tested in sunny conditions, while we now have very dull wether. So, I wait for the sun to come out an hope the new lens will be as breathtaking as the first one.

So far about testing under controlled conditions.

[Fogel70]

I just try a new lens out to see that it perform like expected.
But it is just some random shots of various subjects.

If I notice anything out of the ordinary I will do a deeper analysis of the performance. To see if it need adjustments or if it is defect.
I usually buy lenses used from shops with 3-6 month warranty so I try to verify the lens performance as soon as possible after the purchase.

But I'm not much of pixel peeper so I rarely find any problem with my lenses, or that they need adjustment.

[Lord Lucan]

.
Not a scientific test, but I take a picture or two of an old lens test chart

[CristiC]

Yes, I always test my new lenses. On my cat. Because I have more pictures of my cat than of anything else. Kidding. Take pictures of a grass lawn, on a hill.

The need for testing your lenses by yourself wasn't something I thought of, until I bought the Pentax DA-15mm F4 (without HD) and was not happy with the results. I think I have a less than stellar copy, unable to obtain focus at infinity. I settled on using it just for close-ups, not for landscapes. I read about field curvature and how to use it with live-view. I still hate that lens, as much as I love the starburst effect it gives.

Some lenses have character, something special about their rendering, a certain je-ne-sais-quoi, magic pixiedust that shows in particular situations and settings. It has to be explored, found, identified, then used on purpose. You need some testing work to find it.

Some lenses don't match your camera's autofocus defaults, you get front-focus or back-focus although you thought the picture was in-focus when you heard that confirmation beep. It's always something specific to a lens-camera combo. Then you test the new lens to find out correct settings for your camera's AF fine adjustment, specific for that lens. See "lens calibration" and Roger Cicala's way of testing here.

[Lowell Goudge]

Yes and no.

Once upon a time, especially with older lenses and cameras (especially the K10 and legacy lenses) I tested exposure

What I found was sometimes interesting, for example
- the focusing screen of the K10 induced a lot of exposure errors in green button metering, where it was nightly non linear, but predictable,
- the k10 also, due to the same non linearity had errors when using teleconverters that did not correct the native aperture value reported to the camera. Leading to 1 stop over exposure with a 1.4x TC and 2 stops over exposure with 2x tc on my 70-200/2.8
- my tamron XR dii 28-75/2.8 had an aperture which under exposed wide open and over exposed when stopped down, it went from +1 wide open to -1 at F32, but this exposure issue was an absolute straight line as a function of F-Stop

This helped quite a bit early on with getting good shots because I knew the irregularities and could adjust for them. Lesson learned, know your gear

I really should do more testing, on other lenses and bodies but stopped when I got my K5.

[RICHARD L.]

First I inspect thorougly any "new to me" optical instrument I acquire : smoothness of focusing, diaphragm action, absence of fungus, etc. Then I clean the item carefully : careful wiping of the glass, vigourous cleaning of the mount, car polish on metal painted surfaces (mostly for 645 and 67 lenses). After that I check how it behaves on the camera body : AF precision, exposure accuracy, etc.

Finally I go to my "Benchmark" scenes :

1) a RIVER north of where I live for general color and contrast rendering,

2) a SUNSET on the Saint Lawrence River to check for resistance to flare and veiling,

3) the SHINGLE ROOF of an old Flour Mill that serves as my "brick wall" to evaluate far-away resolution,

4) a CLOSE-UP to check for sharpness and bokeh (though I'm not a "bokeh freak").


1) RIVER


2) SUNSET


3) SHINGLE ROOF


4) CLOSE-UP


I shoot a landscape picture on tripod at every aperture and try to determine the OPTIMUM for edge-to-edge sharpness. Ex. 645 FA* 300 mm f/4 : OPTIMUM = f/16

f/32

f/22

f/16

f/11

f/8

f/5.6

I never use a lens wide-open. When a lens succeeds at all these tests, I can have full confidence it will deliver good images.

[MossyRocks]

I do test lenses that I get but there it is more so that I know where to run them when doing astrophotography. Once of the first things I do is ensure that they focus slightly past infinity ensuring that they will always be able to reach infinity. After that I they get a ride on my equatorial mount pointed at some bright stars and I shoot a bunch of different f-stops to find out where I get results I like.

For normal shooting I am less concerned but before buying a lens at the store I will put it through its paces with some quick shots to ensure I'm not getting a used one with problems I don't want to deal with. After that I will spend some time learning it but not so much really testing it.