Do This for Every Lens You Buy

Imagine heading to a photo shoot and noticing afterward that most images you took are partially out of focus. You are sure you set the focus right, and the depth of field in the photos is also sufficient. The problem might be a decentered lens.

Half of the lenses I bought in the past 15 years were decentered. At first, I didn't know of that problem. The first wide angle lens I got from Canon required focus-stacking all the time. Since I mostly shot landscapes, it wasn't clear to me that this was due to a problem with the lens and not because of the topography of the landscape or the way I composed my photos. Even if I had no element near the lens, I couldn't get everything acceptably sharp. The problem got worse with time.

Interestingly, the out-of-focus areas appeared mostly on one side of my photos. Getting everything into focus in a single image was nearly impossible. So, I kept stacking my images when what I should have done was get the lens fixed.

The problem of decentered lenses is not a problem of cheap lenses, either. When I bought the Canon RF 15-35mm f/2.8 a few years ago, I had to return it twice until I got a perfect copy.

If you're new to photography, you might not have the confidence to return new equipment that doesn't work correctly, thinking it might be something you're doing wrong. I was there myself. But I'll show you now how to test different lenses and build the confidence to return a lens if it's not working flawlessly. When you buy new gear, you typically have a two-week cancellation period you should use if you are not 100% convinced by your purchase.

How To Test Wide and Normal Lenses

In the feature video, I show how I tested the first copy of the Canon RF 15-35mm f/2.8 lens I bought. For lenses up to 100mm focal length, you can perform an indoor test using a target. For longer focal lengths, this can become impractical unless you live in a big apartment.

For a long time, I used a detailed map of the world mounted on the wall of my office. A shelf with some books is also a perfect subject for such a test.

This map of the world served as my test target for new lenses for a long time. The vertical and horizontal lines helped to properly align the camera by avoiding perspective distortions.

With such a flat subject, you must position the camera so the focal plane perfectly aligns with the target. Otherwise, you can run into depth-of-field issues. To get perfect alignment, a subject containing straight lines is beneficial. Those help you spot perspective distortions, which indicate your camera is not set up appropriately. Use a tripod to allow fine-tuning.

Once you've finished the setup, go through the different focal lengths of your lens and take test shots wide open and stopped down. Don't just check the extreme focal lengths because decentering is not exclusive to those.

After you've taken the test photos, assess them in your editing software. Start with the corners, but don't stop there. My second Canon RF 15-35mm f/2.8 copy looked perfect for the corners, but it turned out to be soft in the area to the right of the center.

While the indoor test can help you spot imperfections in a lens, you should also photograph a 3D subject. Head out into the landscape or find some architecture to photograph. It doesn't have to be a photogenic scene. Just make sure you have enough details filling the frame. You should also avoid foreground elements too close to the lens. The goal is to find a scene you can capture with acceptable sharpness when using the widest aperture of your lens. This hyperfocal distance table can help you with that. Sometimes, it can be as simple as taking photos out of the window of your apartment.

How To Test Telephoto Lenses

The longer the lens, the harder it is to find an adequate test subject. At 500mm, for example, an indoor test will be difficult unless you live in a mansion. But even outdoors, it can be hard to find a flat and detailed subject.

There is a simple solution, though, which I show in the video above. When I bought the Canon RF 100-500 f/4.5-7.1 lens, I mounted a target on a tripod, which I positioned 10 meters from my camera.

Instead of just taking a photo wide open and stopped down for different focal lengths, I took five photos each time. I put my target into the center and the corners of the frame.

Good examples of targets are church towers and other prominent, free-standing buildings. Whatever subject you choose, make sure it's far away from your camera. If it's too close, depth of field can become a problem and cause the corner images to look softer than the center image or vice versa.

Conclusion

I recommend using the first two weeks after your purchase to do some of your typical photography in addition to the tests. Find scenes you'd photograph on a photo tour and use the lens how you'd do in the field. When you return home, do pixel-peeping. If you find the lens doesn't perform as expected after the first round of tests, repeat them. After the two-week cancellation period, you should be confident to use the lens on your next photo trip and not worry about soft photos. And if you have doubts, get a replacement right away.

It's also possible that an initially perfect copy gets decentered over time. If you begin noticing it, do another test and get the lens fixed. A few years ago, I paid around $200 to get my Canon EF 16-35mm f/4 lens fixed by Canon service. After I got it back, I repeated the test and confirmed the improvement. To compare the performance of the lens before and after the repair, make sure your test is repeatable. An indoor test is perfect for that.

Michael Breitung's picture

Michael Breitung is a freelance landscape and travel photographer from Germany. In the past 10 years he visited close to 30 countries to build his high quality portfolio and hone his skills as a photographer. He also has a growing Youtube channel, in which he shares the behind the scenes of his travels as well as his knowledge about photo editing.

Log in or register to post comments
51 Comments

this is an excellent article. In the beginning of my photographic activities in 1959, I returned two nikkor lenses for exchange and got no argument at all even though I was more or less still a kid. but I brought 17x20 prints with me to show the salesman what I meant along with another made from a photo taken with the 3.5 Elmar which was my sharpest lens at that time. the replacements were better.

does the author think any of the modern lenses can be recentered if that problem is detected - like the Sony line?

Hey John, I had one of my Canon lenses fixed by Canon service, cost me 200$ -> see conclusion part of article. I'm not sure if it always works. A colleagues of mine had sent in a "decentered" lens once and just got it back as before. I guess service will decide if it's within their defined "margin of error", and if it is try to tell you that it's fine. But I'm a pixel peeper and I like to print big, so I need perfect lenses. It is why it's even more important to find problems within the first two weeks.

I had that issue years ago with a first version of the Tamron 28-75 f2.8. Some of the images looked weird, almost like a tilt shift(ish) effect, and I thought it was just me seeing things because of the angle or something. Then, on a photoshoot where the subject was on a swing about 3 feet from me and a ferris wheel about 30 yards away from us, at 35mm f2.8, the subject was in focus and pretty much so was the ferris wheel. That's when I've confirmed something was wrong with the lens. I never got it fixed. I was mostly shooting with primes and wasn't motivated to put on that Tamron. I ended just selling it to KEH.

In the last 12 years, I've never experienced anything like that. But, since that Tamron, I check all my lenses.

So you just passed a defective lens off as good and hoped the next person who bought it wouldn't notice?

Do you not know who KEH is?

Yes. I've done business with them, both via courier and walking in to their physical location in suburban Atlanta. I would never sell them a lens I knew to be defective without informing them of such.

They base their grading on the cosmetic condition of the gear. They do not do any optical testing before buying anything.

--- "They base their grading on the cosmetic condition of the gear. They do not do any optical testing before buying anything."

They wouldn't be a trusted source for used equipment if they didn't do any testing.

https://www.keh.com/shop/grading-system

I've walked in their door, handed a representative two lenses, waited less than five minutes while he visually inspected them at the counter without using any optical diagnostic equipment, since there is none in their lobby, and walked out with a check. There was no optical testing done prior to buying the lenses from me. One was a Sigma 120-300mm f/2.8 Sports, the other was a Canon EF 100mm f/2.

And, you think they just wiped them clean and put it on the shelf to be sold without inspecting them after you left. Is that what you really think? Let's put on our thinking hat.

That has nothing to do with my statement; "They do not do any optical testing before buying anything."

However, based on my own and others I know experience, I'd guess whatever optical testing they *might* do is not very in-depth. A fair percentage of lenses bought from KEH do get returned during the grace period because they have alignment issues.

It's far more efficient for them to buy a lens back in the first two weeks when someone who buys a lens notices it's not performing up to spec than to spend the money doing that kind of testing on every lens that goes through there.

--- "A fair percentage of lenses bought from KEH do get returned during the grace period because they have alignment issues."

Sounds like you are just grasping at straws and making things up as you go along. Do you have source(s) of the percentage and reasons for the returns?

Sounds like you're mad because someone called you out on selling a defective lens without telling the buyer it was defective. I'm not making anything up. On the other hand, you admitted to unethical behavior in your original comment above and are trying to deflect away from that.

Nope. Big companies can take care of themselves that's why I didn't go the eBay route.

Sounds like you are pivoting because someone caught you in a lie.

Keep living in denial if that's what it takes to like yourself. Selling a lens you know to be defective without disclosure of such is unethical, regardless of how "big" or "small" the buyer is. I know what my dealings with KEH have been. You can project your dishonesty onto others all you want, that doesn't make it so.

Oh, cry me a river. They've been in business for 45 years. You don't do that by constantly crying, "Boo hoo. Not fair." When dealing used, you take the risks. They're willing buy equipment from some random person off the street that just walks in and then hands them a check. They have the means fix or write off.

Below are a few images from that lens. For all you know, the buyer is happy with the purchase for the price. If not, KEH will make them whole. At the end of the day, everyone is happy, except those projecting the skeletons in their closet.

Most of them don't seem to test gear. They don't even do a shutter count. The last purchase I had to do the cleaning internally myself. I had the tools for that but that was my first and last time buying from that local store.

Who is "most of them"?

Shutter counts on lenses?

Which "local store"?

--- "The last purchase I had to do the cleaning internally myself."

Did you not inform them?

When you speak in general terms and can't name names, everything you say sounds hyperbole.

I won't trash a store on an open forum. It was also clearly sold "as is". The price was good and I know how to service cameras (film too). I'm 71, I was glad to increase my DSLR chip by more then 2x for very little money. Apparently you have never been sued.

I might just be the king of hyperbole...

Looking through the list of tests/checks they do, they will not detect a de-centred lens.

Correct. Everything they look at is based on cosmetic condition and if the thing turns on and off properly, in the case of items with an on/off switch.

This is excellent. I particularly like that you included very detailed info on how to test lenses.

Also, I've often wondered how much it would cost to fix lenses with issues like this. I appreciate different manufacturers will charge different prices, that lenses have different warranty periods, etc. But it's good to have an example to give us a ballpark number.

Thank you for share.

Actually I have sent a new Canon 24-70 F2.8 RF to Canon service because it has like a front-focus microAF problem (I thought that was not possible on mirrorless). I have had to prove that shooting a lines template (focus test chart V2)

Several months with this problem because I never believed that a lens of this level from this brand could have these type of problems. It's really annoying.

That's not a front focusing problem, it's a user error problem.

Focus "points" aren't really points, they're areas. The areas of sensitivity for each focus "point" are much larger than the little squares, rectangles, or dots in the viewfinder. The camera will focus on the highest amount of contrast anywhere within the active AF area. Most camera's logic system is also biased towards focusing on the nearest area of high contrast within the active focus area when contrast at several distances is the same. To properly test your camera's AF system, you need the target potion of the test device to be parallel to the plane of the camera's sensor, not at an angle to it.

I attach a screenshot of the DPP where you can see the focused area. For me the focus is before that area. What's more, I did tests with the same type of focus and another lens, and the focus came out exactly as expected.

Regarding the focus tests, they are done with a 45º angle to determine if it focuses before or after. If you did them with a parallel plane, everything would be in focus or out of focus but you wouldn't be able to validate it.

That little square is not the actual active AF area, though. The actual area is much larger than that little bitty square. It's probably closer to something like this:

Do you have any official documentation that confirms that?

PD: In any case, as I said before, the same test with another lens focuses correctly and no front-focus is observed. So, it invalidates your hypothesis.

No, it shows that the other lens is back focusing by the same amount as your user error in aiming it on front of the actual target.

Go read Roger Cicala's Lensrentals blog entry titled: Autofocus Reality Part 3B. Also Andre's Blog entry titled: Cross type AF points in EOS 7D.

This is an officially published "map" of the EOS 7D AF system. I got it from Chuck Westfall's (RIP) Flickr page years ago. Other Canon DSLRs are similar, but the 7D chart is easier to understand with only 19 AF 'points" instead of 45-61-65 AF points like later DSLR models had.

The top right chart represents what you see in the viewfinder (without the numbers next to each AF "point"). The middle right chart shows the setup of the PDAF sensor array. (Note that micro-lenses on the entrance to the PDAF array aim light from various parts of the unfocused light field projected by the lens onto different areas of the PDAF array. There is not a 1:1 correspondence between position in the frame and position on the PDAF array.) The lower right chart shows the actual area of sensitivity for each AF "point". As you can see, the areas of sensitivity are much larger than the little squares in the viewfinder, and they are often shared by adjacent AF 'points'. The chart on the left tells which lines on the PDAF array in the chart at middle right are used by each AF "point" labeled with that number in the chart at top right.

Awesome! Now, show one for mirrorless because that's what he has.

The idea is the same. Areas of sensitivity are larger than the little indicators in the viewfinder. Especially with AF systems now having 65-100+ AF "points" that overlap, if they showed the actual map in the VF you wouldn't be able to compose due to not being able to see past the AF indicators.

That's like saying that combustion engines of old perform just like the combustion engines of today because the idea is the same. They don't.

Even if I were to entertain your idea that mirrorless (ML) is designed the same way, clearly the performance and accuracy won't be the same. We're talking 19 AF points vs 1000+ (Canon R5). And, that's not even taking into account the processor. There's a reason why even at f1.2, ML can hit the eye with high accuracy and consistency.

As for Tony's issue, the lens is front focusing. It's not user error.

These aren't internal combustion engines. They're cameras that use the same design philosophy: Make the indicator small enough in the viewfinder for the viewfinder to still be usable while making the areas of sensitivity they denote much larger to make the AF system usable for the typical user.

You're arguing against straw opponents when all one has to do is actually go take pictures with these cameras and test them to see that the areas of sensitivity are significantly larger than the indicators in the viewfinder. This is just as true for MILCs as it was for DSLRs.

With 1000 points on the same amount of real estate, there's going to be even more overlap than with 19...

If Tony were to use a better target, his results would very likely be different. With MILCs, there's never a difference between the distance of the focus detector and the imaging sensor - because it's the same part of the camera. Aiming at a tilted target almost always produces "front focused" results. That is, results that appear to be front focused to those who don't understand how AF systems work.

--- "You're arguing against straw opponents"

The only straw are those things you are grasping at, again.

--- "These aren't internal combustion engines.”

No duh. It's what we call in the intellectual community, an "analogy". And, it fits this scenario perfectly.

--- "They're cameras that use the same design philosopy"

Wow, look at that. Similar to combustion engines. However, the common sense logic that seem to escape you is this thing called "advancement in technology".

--- "With 1000 points on the same amount of real estate, there's going to be even more overlap than with 19"

And, therein lies the problem. You're to hung up on DinoSLR tech. If you think that 1000 points is just as accurate (or inaccurate) as 19 points, you might want to run along to KEH and sell the rest of you camera equipment.

--- "If Tony were to use a better target, his results would very likely be different."

No, it would have been the same because it's front focusing severely. And, if he's mentioned it 2 times, he's mentioned it 2 million times, his other lens is focusing as expected using the same target.

--- "Aiming at a tilted target almost always produces "front focused" results."

Wrong. There's this thing called "plane of focus". Read it up on it when you get a chance. No matter what angle that target is (horizontal line in this case), straight across, left to right should be in perfect focus. If anything, because it's just a simple horizontal line, there's nothing in front or back of the line to cause front/back focusing.

I would like this article to actually go into detail about what a decentered lens is, like what actually happens with the elements or whatever that results in decentering. I would also like a lot of different examples of decentered images, from several different lens models.

I will not watched linked or embedded videos. so having all of the info written out in the article here on Fstoppers would be great!

If you Google "Photography Life Decentered Lens" the very first result is for a very good article giving a great explanation complete with diagrams and images.

Good! Thank you. I would have liked that kind of information to be included here in this article. But I will have to go off-site for it.

Why do you say the lens is decentered when the problem you describe is almost certainly caused by a tilted, not decentered, element?

The very definition of a decentered lens includes tilting of an element:

Decentered lens: A decentered lens contains one or more optical lens elements that are either moved or tilted from the principal axis of the lens.

Nope. Uncle Roger explains the difference.

https://www.lensrentals.com/blog/2012/05/testing-for-a-decentered-lens-a...

"Strictly speaking, decentering would involve one or more of the lens elements being off of the central axis of the lens."

"An element can also be tilted to one side or another. Strictly speaking this is not decentering, but it can have similar effects, so people often say a lens is decentered when in fact it’s tilted."

"The third problem that can occur with lens elements is spacing. If elements aren’t the proper distance apart the lens may not focus the image sharply, or might not focus all the way to infinity. But the lens is not decentered and the tests we’re describing would be normal. "

Nope. The term "decentered lens" covers more than just decentering. A term is often more encompassing than its literal meaning .... as is the case here.

I can find web sites that say the U.S. never put any men on the moon. Since it's on the internet it's got to be true! Right?

I'll go with Uncle Roger and scores of printed textbooks on this one.

Hard Pass. The number of lenses that are completely fine and within manufacturer spec that overly obsessive pixel peeping photographers send back (after using) is a joke. One can EASILY find flaws in lenses when doing these tests (speaking from experience) as no lens is perfect.

Yes there are the rare instance where a lens is so badly decentered (likely from being dropped somewhere along the way) where it makes sense to run a test and swap it out but to claim that half the lenses you purchased are decentered and have to go back is pretty lame. I've purchased a handful of brand new lenses from B&H that were NOT brand new. Most likely these are the rejected test samples by over zealous pixel peepers. At least Pictureline charges a restocking fee to dissuade people from doing this.

I'm sorry to contradict you but photography is about PRECISION tools. (It's not like being a bricklayer where 1 cm error is fine.)
Obviously there are more tolerable cases and more severe cases, as in everything.
But for you to say in general that people are very exaggerated seems to me to be exaggeratedly free.

This is very good. I did this back in the 1980s when trying to match slide projector lenses for a six projector show. Similar problem with focus across the frame.

That is not how to test for de-centering.

More comments