How to check how decentered your lens is

How to check if your lens is decentered?

check lens decentering centered centering well bad sample variation

You may have already heard of sample variation or decentered lenses. In this article we will show you how to easily test your lenses for proper centering. If you know how it works it will take you less than 5 minutes per lens!

Where this test originates from

This test was published by the German Gletscherbruch homepage, but as this website is only available in German we decided to write an English version.

How does it work?

What you will need:

  • Tripod
  • Some building far away (chimney, antenna, church tower etc.)

How to set your camera:

  • use the widest aperture of your lens (some lenses have very bad corner performance, stop these down until there is visible detail in the corners).
  • manual focus
  • manual exposure (or use AEL toggle)
  • IBIS/OSS off
  • electronic first curtain shutter off (if you have a stable tripod)

What to do:

  1. You set up your camera on the tripod.
  2. You put your target in the center of the frame and you focus on it.

    Focus is on the chimney in the center of the frame
  3. Do not change focus!
  4. Move your camera on the ballhead until your target is in one of the extreme corners of the frame and take the shot

    Chimney in upper left corner
  5. Repeat step 4 for the other 3 corners.
    Chimney in upper right corner
    Chimney in lower left corner

    Chimney in lower right corner

How to rate your shots:

Before you rate your shots there are a few things to consider:

  • This is the most important one: do not send us your shots and ask if we think your lens is okay. We are neither your free-lens-checking-service nor your lens manufacturer’s quality assurance team. You check for yourself to see if you are okay with it and not to see what we think. Seriously don’t send them to us.
  • A perfectly centered lens does not exist and a very slight decentering you will only rarely if ever notice in your pictures.
  • There is no objective standard. You will have to decide if what you see is an issue for you or not and if you can expect to find a better copy.
  • Every manufacturer ships decentered lenses, but some are worse than others. Zeiss for example has significantly better quality control than Sony, but no manufacturer is perfect.
  • Usually fast and wide lenses are more prone to decentering.
  • Zooms are more prone to decentering, compared to primes you should lower your standard a bit.
  • When testing zooms for decentering you should do it at the extremes and in the middle.
  • If your lens has field curvature the corners will look worse than the center. To check centering you need to compare the corners to each other, not compare the corners to the center.

The “perfect” lens

check lens decentering centered centering well bad sample variation

This is pretty much as good as it gets, there is almost no variation. You just won the lottery.

The “almost perfect” lens

check lens decentering centered centering well bad sample variation

One corner is very slightly worse, but this will be pretty much unnoticeable. This is a perfectly usable lens.

The decentered lens

check lens decentering centered centering well bad sample variation

This one is badly decentered in a way that it might actually have a visible influence on your results. You might want to consider returning this or have it serviced.

We can also highly recommend Roger Cicala’s article on optical quality assurance. It is a more complicated business than you might have guessed.

Further Reading

Support Us

Did you find this article useful or just liked reading it? Treat us to a coffee!
via Paypal

The following two tabs change content below.
My name is Bastian and I am your expert here when it comes to ultra wide angle lenses, super fast portrait lenses (ranging from a 50mm f/0.95 to a 200mm f/1.8) and I also have reviewed way too many 35mm lenses. Don't ask me anything about macro or wildlife shooting though.

157 thoughts on “How to check how decentered your lens is”

  1. Hi Bastian
    Thanks for this (as usual) usefull article.
    I have one silly question : how to easily put the four 100 % crop next to each other for better comparison ?
    Thanks again and I look forward to your next post.

  2. Hi, for me, it is not clear, why one set of pictures is better than the other, since it is not mentioned, what your should be looking out for.

    Could you please update the article to make this more clear.

    Thanks a lot, also for all the other great articles!

  3. Where would I service such decentered lens when it’s out of warranty (used, vintage, second hand)? And can or do the usual manufacturers repair such decentering? And round about what costs are we talking about?

    I’m thinking that it’s quite difficult to find some one who can fix decentering on a vintage lens. And one modern lenses, it might be too difficult or expensive and the service just rejects the case (all within acceptance). But I never tried …

    1. Depending on where you live, there should be choice of “old school” repairpersons you can send your lens to.

      Before camera lenses started being considered “disposable” electronic items (!!), it was common practice to have lenses Cleaned, Lubricated and Adjusted (CLA’d) every few years (or decades), especially if a working pro.

      1. Yes, I know. But decentering is another level, isn’t it? Most 1st party manufacturers refuse to touch a decentered lens and send you right away a replacement. Micro-aligning such modern, complex lens designs is very, very difficult or even possible.

        1. Yes, not all lenses can actually be aligned.
          Yet not every optical design is prone to decentering in the same way.
          With more thick elements it is less of an issue compared to a few thin ones.

  4. Your post is right on time, I got my Zony 35 1.4 yesterday. Have been reading up all the posts on decentered copies of that lens at Fred Miranda and now I am a bit scared to do that test …

        1. Thanks for the image, now I know what the folks at FM where referring to when they spoke of clearly visible decentering of the 35/1.4.

          Mine seems to be Ok (at least within my tolerances):

  5. Thanks for this very useful article. Have heard about it but didn’t know much on how to test it. I’ll now get down to test all my lenses.

    Many thanks

  6. Very useful article especially with the example images. However, the first bullet on how to setup the camera is a bit unclear as the parenthetical seems contrary:
    – use the *widest aperture* of your lens (some lenses have very bad corner performance, *stop these down* until there is visible detail in the corners).
    – use the widest aperture of your lens with reasonable corner sharpness (some lenses have very bad corner performance, stop these down until there is visible detail in the corners).
    Also, a sentence under the decentered lens pointing out the differences in sharpness would probably help clarify things for people new to decentered lens testing.

  7. Hi this is Lee

    last sunday

    I had tested my batis 18 mm with your test mehtod but i can not tell mine is good or not.

    could you check my result??

    if you can please send me a mail than i will reply to you with my shots

    thanks for your tip.

  8. Okay it seems i got unlucky with all my lenses 🙁
    I have the 24-105mm F4, 16-35mm F4 and the 55mm 1.8

    Could you guys tell me if i should give them back?
    I didn’t notice any of that during my everyday use.

    It would really be appreciated!

  9. Good article, pretty easy to follow. Received a copy of the FE 85/1,8 today and did one test on an infinity object (mast) and one on a closer one:!AlCoUQ0jvInBgYp-nNmK_9arjrQnCA

    To me the bottom corners look slightly worse than the top ones but it is slight. I’m pretty happy one side or corner is a smear but then this lens has been described as ‘sharp into the corners wide open’ so I wonder what others think?

  10. Thanks for this interesting paper. In your experience, did you see badly ‘decentered’ lenses that were actually fine, but the ‘decentering’ effect was due to a bad adapter with non-parallel mount planes?

      1. Oh, I know, you are right! I forgot to mention, but I did a test based on your post too. The result was soo bad on the left side that I was sure I must did something wrong. Here it is:
        (just bottom left and bottom right corners)

        The guy who selling this lens did some test shots too (he mentioned me this blog actually, for which I was very grateful) at f4 or 5.6 and it was pretty good. Could this be true that a lens show almost nothing at all at f4 and on its widest f2 it is so bad?

        1. I don’t think you did anything wrong, this sample of the lens just sucks.
          Decentering goes away when you stop down enough.
          I would not buy that lens though, especially if you intend to use it for e.g. astrophotography.

          1. Thank you very much for your help! The seller is a very kind man, he is willing to take the lens back after these tests.
            I am just wondering now, if I should try out more samyang(rokinon) 12mm f2 or the samyang 21mm f1.4 could be a better lens? (I saw too many decentered samyang 12mm f2.) With fuji x I don’t have many choices for astro.

          2. There isn’t really much choice, try another Samyang 2/12 or have a look at Laowa 9mm 2.8 instead.
            With the 21mm you will be doing panorama for milky way all the time, not sure whether that fits your workflow.

  11. Hi! Thank you very much for this helpful article. I am also one of those having a hard time judging my results. These are some samples from my new Tamron 15-30 G2 on a Nikon D850. I am not so convinced, one corner seems to be visibly sharper than the others, even at f/8. But what do you think?
    Thank you!

    At 15 mm, f/2.8:

    At 15mm, f/8:

    At 30mm, f/2.8:

    At 30mm, f/7.1:

      1. I’ve sent the lens to CPS repair center for testing this problem. Answer is – in specification….

    1. Very often when a lens is decentered that means the field is tilted and the bad parts look better when focus is adjusted.
      I cannot tell you whether you are happy with this lens or not, but as you decided to make a post that probably means you are not happy so you should return it I think.

  12. Hi! Maybe I’m a little bit stupid.. but I don’t really understand how I should take the four photos… I don’t have a good tripod and no ballhead, is it possible to do this handheld? Don’t you change the focus if you move the camera while handheld? For even more clarity in the description it would be great if you add photos of the camera positions while taking the four photos.
    Kind regards,

  13. Hi Bastian,

    I just got my copy of Sigma FE 35 f/1,2. While I’m very happy with its performance and the images I get (people), the de-centering test seems to show that I might have a problem. This wouldn’t be a huge deal were this lens cheaper, but at this price I think it should be less obvious.
    Stopped down everything is fine across the frame, from let’s say f/6,3.
    Could you asses please? The upper right corner is the one troubling me.

    Thank you!

  14. Nice and easy and short. Thanks Bastiak, I had read so many times but not occurred to do it with a tripod only a flat target

  15. Checked my Sony 16-35GM; lucky me–there are no visible differences among the 4 corners at 16mm or 35mm. 🙂

    Thank you very much for the well-explained, simple system.

  16. Great article as usual, Bastian. I am a bit confused by step 2 to 4. Why focusing on a subject in the center of the frame first, then moving it to the corner positions, instead of putting the subject in the corner positions directly, then focusing on it?



    1. It is possible, that an element is very slightly tilted, so one corner is best focused slightly in front of the center and the opposite corner is best focused slightly behind the center.
      It is further possible, that this very slight tilt may still be in your tolerances, but if you focus on the corner the deviation between those corners will be double compared to what you get when you focused on the center before.
      I hope I could make this somewhat clear.
      For a perfectly centered lens it should indeed not make a difference.

      1. Hi Bastian. If you focus on the center and one corner is tilted forward while the other is tilted backward by the same amount, the two corners will look equally out-of-focus, and you will wrongly conclude the lens is perfectly “centered” (untilted)

        The doubled difference you mention instead always correctly shows you how tilted the lens is.

        Fred at FM also promotes center focus in his method, and both your page and his are referenced frequently, so I thought I would send you this note for consideration.

        Cheers, Richard.

        1. In case it wasn’t clear in my comment above, I thought I better add that what I am proposing therefore is that focusing on one of the corners rather than the center is preferable because it is both more reliable and more sensitive.

          1. You will have a hard time finding a perfectly centered lens, they all show some degree of tilt.
            When you focus on the corners you will end up returning many lenses that are indeed perfectly usable for any task.
            This is not about finding the perfect lens that doesn’t exist, this is about sorting out the real duds.
            Therefore focus on center is preferable for 99% of people.

            As you already know about these things you are probably also not the target audience of this article.

  17. “This is pretty much as good as it gets, there is almost no variation. You just won the lottery.”

    In the first example, I wouldn’t necessarily conclude that four simultaneously sharp corners means you got the best possible sample. It’s still possible that the maximum centre MTF wasn’t necessarily achieved. One could conclude they received a sample that exhibited even behaviour across the field and possibly a relatively flat field.

  18. Hi,
    Thanks for posting such a helpful article.
    I have a lens Pentax FA 24-90mm which for sure is decentered at 24mm. But starting from 28mm the decentering effect becomes negligible and totally dissapears between 35 and 90mm.

    I wonder how this is possible. If decentering effect is visble at one FL, should it not be visible at all other FLs?

    Can you explain such phenomenom?


  19. Hi Bastian. I bought a used M-mount CV 35mm 1.7 to adapt to my A7 and the top two corners are much softer than the bottom two. The top left corner is very bad, much worse than all the rest. The bottom right corner is decent. Here’s the test:

    I know this lens has a focus shift problem in the corners due to the Sony sensor, would the problem show like this? I thought each corner would be more equally affected. Please reply if you can, I’m trying to determine if this lens is defective.

    Unfortunately I won’t be getting a correction filter for some time, Opto-sigma seems to have a delay for now due to the virus.

  20. Hi Bastian,

    Thanks a lot for posting this very usefull article !!!
    Could you have a look on these lenses (shooted at 42Mp)?

    I am not sure about my analysis :
    The 85mm seems decentered. The 55mm also but little less.
    The 15mm & the 24mm seems not so bad (even if a little something on the 24mm)

  21. Hi everyone. I am testing a couple of lenses at moment using your method. Is the way you do it at a distance in corners the same regardless of focal length? Will it test decentering of a macro lens as well as a standard 50mm lens using same method? Is this a way of standard testing to see if any problems? Thanks dean

    1. If you want to use your macro lens mainly for taking pictures of stamps or something like that it may not be the best method, otherwise it works for every focal length.

  22. Hello everyone. Thank you very much for this article.

    I got a Tamron 28-75mm F2.8 Di III RXD this week and I am trying to decide if I should keep it or not. I am really worried about decentering. I have never done this test and I don’t know if I am doing it properly. I followed your instructions on a Sony A7ii. I tried from 24mm, 35mm, 50mm and 75mm.
    Could you have a look, please? Am I doing something wrong?

    Thank you

  23. A question: if I have a tilted bayonet/tilted sensor, will I have similar poor performing corners on each lens/focal length? Or would it be more prominent with a particular lens? I have had three copies of a “bad” 24mm lenses and want to rule out sensor/bayonet tilt but I can only see noticeable differences on those 24mm lens, not at 35mm, 50mm, or even 14mm…

    1. The faster and wider your lens the more pronounced it shoud be.
      Lenses like 24mm 1.4, 20mm 1.8, 15mm 2.0 or 35mm 1.2 will most likely show issues whereas with longer/slower lenses a tilted bayonet or sensor will be harder to detect.

      1. Makes sense to me and confirms to me that I coincidentally had three “bad” lenses. Corner-to-corner differences were obvious at 24mm 1.4 on each copy and nearly no distinguishable difference corner to corner at 35mm 1.4. Thanks for your help!

    1. Is there a standard for what magnification the crops should be viewed for comparison (100%, 200%). It seems with greater magnification decentering seems more of an issue.

    2. this is bad for a sony 24 1.4, i assume it’s Sony GM 24 1.4…bad centering for such an expensive glass, i’d definetely return that one for a great copy and keep changing untill i find it.

    1. If you can rule out camera shake and antenna movement for the top right one,
      IBIS was turned off, shutter speed fast enough or you used a stable enough tripod
      and if there isn’t a general issue with your sensor being tilted (meaning not most lenses show a weaker top right corner on your camera) I might be tempted to replace this one for my peace of mind, if easily possible.

          1. I completely agree.
            The 85mm top right is still great but the 35mm top left is awful. I should buy new Sigma 35mm dg dn later.
            Thank you, Have a nice day.

        1. Hello Iman,
          I had two Sigma 85DG DN art for Sony e-mount and i can say 1st one is badly centered, upper right corner was bad, then tried 2nd one, still not perfect but was clearly better than 1st one and i kept 2nd one and quit trying…I think this lens is low production quality, i dont think you can find perfect one.

  24. How does this method compare to the Zeiss Siemens Star Chart test (as described by Lensrentals)? I have a Batis 18 where I suspect a slightly weaker corner however when I test via the Star Chart it is flawless.

    For wide angle lenses the light entrance angle would be more extreme so more difficult to get it relatively similar 4 times?

    1. I don’t think the Siemens Star test is one that should be carried out by non professionals. This one is way easier to get right, also for an 18mm lens.

      1. If I understand correctly, the star chart test only looks for lenses where the midpoints of the lens elements do not perfectly align. If one of the lens elements is “only” tilted/skewed, that test wouldn’t find it.

        The test from this page (the Gletscherbruch test) on the other hand will find ANY defects which result in a decentered image focus (so decentered lens elements AND lens tilt). So indeed the more complete test.

        I suspect that some sort of star chart/collimator tests can be executed automated in the production process but tilts are much more difficult to spot. So in modern lenses you might see rather tilted elements than actual midpoint decentering (as QC might catch those easier)?

  25. I’m confused. I did the most repeatable test with the most accurate corner mapping. The distance is not set to infinity.I have a question, how is the central object taken into account during the test? Because I have noticed that some lenses have less contrast and are sharper even when slightly off-center, the texture of the glass is much more visible, I have no idea why this is. Is it AEL or the incident light or simply the focus is more precisely set. I have done the test several times and the results are similar. Below is an example of a Rokinnon 45mm. Please help as you rate. Rokinnon is my 4th and 5th copy. Sonnar 55m, I hit it the first time, and the test was performed during snowfall, which is a plus for this lens. Sony A7RII.

    Sonnar 55mm 1.8 Sample

    Rokinon 45mm 1.8 Lens 1 Sample 1

    Rokinon 45mm 1.8 Lens 1 Sample 2

    Rokinon 45mm 1.8 Lens 2 Sample 1

    Rokinon 45mm 1.8 Lens 2 Sample 2

  26. new to a camera body that has IBIS. when i originally did these tests with a DSLR, i made sure i had the stabilization off on the lens, just did this test with 2 new lenses on a mirrorless and accidentally left IBIS on.

    should IBIS be turned off because it COULD make some corners softer giving a false positive?

    or should it be turned off because it can HIDE the negative effects of a decentered lens?


  27. Does this test work on lenses with significant filed curvature? I have such a lens which has completely failed this test. The lens is an excellent performer especially when it comes to center sharpness wide open so I was surprised. I did a different test for centering using a Zeiss Siemens Star Chart and that test shows no centering issues. It looks perfect actually. I am trying to make sense of the discrepancy.

      1. The Zeiss Siemens Star test is something from Roger Cicala at lensrentals, although he admits it isn’t perfect. The other issue is that I displayed the chart on a 27 inch monitor and took a picture of the monitor and not the paper. Other than the circle in the middle there are rays and of course you can resolve the pixels in the monitor. In that picture the lens seems fine. You can see the extreme corners are pretty much the same with some corner softness in the extreme corners but at the same level. Perhaps I can post some pictures. I can also attempt to redo your test again in case I didn’t do it correctly. However, I already did it twice. I will share the pictures from both tests and maybe you can provide your feedback?

          1. The lens is a Mitakon 35mm f/0.95 lens. Tests were conducted wide open at f/0.95.

            Chart test on monitor. I see no issues. Corners appear the same at all corners. Full resolution jpeg.


            The test from this blog. Corners are pretty soft, but the softness is not the same indicating decentering. Right has visibly more smearing.


            I did the same test at f/2. Not shown here and pretty much the same. Right side is softer than the left.

            Actual pictures taken with this lens. Doesn’t seem like it has any obvious performance/decentering issues.

            This picture taken at f/4.


            This picture I think is either wide open at f/0.95 or f/1.4


          2. Siemens star carried out like that is rather useless.
            Wrong distance, too little detail in the corners, way too hard to properly align.

            With “our” test there clearly is diagonal tilt visible.
            Being a cheap f/0.95 lens I doubt you will find a better sample though.

          3. Thank you for your input. I appreciate it.

            I only posted an in focus picture of the chart. However, I did two other pictures with the lens defocused to different degrees and those also looked fine. There was no overlapping of circles.

            Shouldn’t a decentered lens also be affected with central sharpness? The two normal pictures I posted don’t show any such issues. They are very sharp pictures. They don’t even show corner issues unless you see something I missed.

            In fact, the picture of the candle holder was a test with three 35mm lenses. First, the candle holder stretches all the way from the left to the right and there is no visible softness on the right side. Second, the lens was up against some serious competition. One was a modern Fuji prime that at that aperture should be approaching 80lpmm. The other was a vintage prime tested long ago in a photography magazine that should be exceeding 80lpmm at that aperture. The Mitakon was the winner.

            Either way I think I will move on. It’s not all bad news. I tested 3 more lenses. Two of them are perfect. What you called lottery winners: Fuji 35mm f/1.4 and Fuji 56mm f/1.2. They also surprisingly have excellent corner sharpness even wide open. The other one a Fuji 27mm f/2.8 is acceptable. Not the best copy but useable. You can’t win them all.

          4. I think I have an answer. The test in this blog is a more general optical test while the star chart is only for decentering. People are using decentered as a general term when in fact lens elements can be decentered, tilted or spaced incorrectly. The Mitakon does not have any decentering, but it does have tilt. This is why it passes the star chart test and fails the Gletscherbruch test. Since the lens is sharp and not a poor performer I wonder if the tilt has been done on purpose as a compromise or if is is a manufacturing defect.

            Overall, I don’t think it matters. Mitakon claims to have a 5 year warranty on their lenses but their support have never emailed me back. Like you stated before it’s unclear if I can get a better copy of the lens. At close and portrait distances the lens produces excellent images so I will just use it.

  28. Hello Bastian,
    I hope you’re fine. I just bought Sigma 105 F2.8 DG DN macro for Sony E camera. Here are a few test shots. What do you think? Last three photos are shot at minimum focus distance , i can see focus plane but dunno how it is to you,
    or they’re usefull at all…
    Thanks for looking and evaluating.Best regards.
    (you have to click full resolution button to load completely)
    last three

  29. hello bastian
    i am not trying to be obtuse, but i have a question about the process. when you say, “move your camera on the ballhead until your target is in one of the extreme corners” do you mean, a) unlock the pan/tilt to reframe the target, or b) leave the pan/tilt locked and loosen the ball head leveling feature to reframe the target? i assume the former (a), but want to be certain. thank you in advance for the clarification.

  30. This might be a dumb question, but will a lens be equally decentered throughout a zoom range? I’m getting a 24-105 soon, and wondering if I only have to test one focal length, or if I should do wide, mid, and long.

    Thanks for the write up!

  31. Hi everybody!
    This is the first check I do, so I’m looking for confirmations.
    I plotted the degrees on the image for easier assessment:

    The lens is a SEL35F14GM
    Thanks for helping!

    PS: in degrees, which deviation do you consider acceptable?

    1. If I’ve understood decentiering correct, decentering doesn’t cause geometric distortion to differ in the corners(although this is probably possible as well) but the resolution. In my eyes the right side of your sample image looks a bit better than the left and so there could be some decentering there, but nothing too major.

    2. Maybe you are making a joke?

      Decentering causes loss of resolution: the geometric effects in the outer field are are a combination of perfectly normal rectilinear wide angle lens effects combinedwith not holding the camera plumb to the ground.

  32. Hello,
    I have tested 24mm lens and it’s decentered. Do you have any idea what element could be shifted. When I focus on the infinity both right corners are sharp almost as the center but both left corners are out of focus (when I focus a bit closer they get sharper). Also the left side is rendering sharper horizontal lines so when I change the focus they transform 1.3x more in vertical way (oval shape bokeh like anamorphic hahaha)

Leave a Reply

Your email address will not be published. Required fields are marked *