How to Tell If You Have a Good or Bad Copy of a Lens

After buying an expensive zoom lens, it’s normal to wonder if you got a good or bad sample. However , if you don’t have got multiple copies of the lens to test, answering that issue may seem difficult. That’s what I hope to change with the procedure outlined in today’s post.

The biggest issue that differentiates a good compared to bad copy of a lens is decentering . In other words, one or more elements in the lens are tilted or off-center, which can lead to blurriness in certain regions of the frame. The process in this article is designed to figure out whether your lens is decentered, without needing a second copy of the zoom lens to compare against.

Before I tell you the particular steps, keep in mind that no lens is perfect. When we test lenses at Photography Life, we regularly sort through multiple copies from the lens to test for small sample variation, and there are always some differences from duplicate to copy. Usually, right after are minor and may just be visible in diagnostic tests. You don’t need to worry about minor decentering or small differences between copies of a lens.

That said, occasionally the results on one duplicate are bad enough they can harm real-world photos. Is that true for your lens? Here is the process I recommend following to find that out:

  1. Place your camera on a stable tripod in a non-windy region.
  2. Find a subject at infinity, such as a mountain on the horizon or anything in the distance seen from an overlook. Look for something that has lots of little details. Put that issue at the center of your structure.
  3. Now it’s time to focus. Switch your lens towards the widest aperture. Focus in live view; autofocus is okay, but you can also focus by hand at a high magnification.
  4. Enable a two-second self timer as well as electronic shutter or electronic front-curtain shutter when available on your camera. Ensure you’re shooting raw, not JPEG .
  5. Without refocusing , recompose your shot therefore the same subject is in around the corner of your photo. Take an image. Then, recompose three more times and take three more photos, each time with the subject in a different corner.
  6. Bring the photos back to your computer for analysis. How razor-sharp is the subject in all 4 photos? If it looks about the same in each corner, your own lens is very likely a good copy. If one or more sides is  clearly less sharp than the others, repeat this test. If your results continue to show an absence of sharpness in the same corner(s), your lens is very probably decentered.

Here’s how this looks in practice. First, focus on a subject in the center of the image. Below, I find the peak of the mountain because it was at infinity together lots of little details. Remember to take these photos at your lens’s widest aperture:

Uncropped-Center
Put a faraway, detailed subject in the center of your own image to focus.

Then, without refocusing, recompose so that the subject (the mountain peak) is in all separate corners:

Uncropped-Lower-Right
Bottom right
Uncropped-Lower-Left
Bottom left
Uncropped-Upper-Left
Top left
Uncropped-Upper-Right
Top right

In your post-processing software, plants all the images to magnify your subject and evaluate the sharpness of each picture. Here, there was noticeable decentering, with the top right corner being less sharp than the others. This is what a lens with moderate decentering appears like:

Cropped-Lower-Right
Bottom right (100% crop)
Cropped-Lower-Left
Bottom left (100% crop)
Cropped-Upper-Left
Top left (100% crop)
Cropped-Upper-Right
Best right (100% crop). Take note the blurriness in this corner compared to the others.

There could always be the fluke with an individual photo that makes it look unsharp. Do it again the test with different subjects. If you clearly and consistently view the same corner or sides looking worse than the other people, your lens is most likely decentered.

Again, many lenses are decentered to some degree. Don’t worry about sending back again your lens unless the blurriness is so bad it will cause problems with your work. In case your type of photography involves out-of-focus corners (such as portraiture) or shooting at filter apertures like f/11 most of the time (such as landscape photography), you may be able to tolerate some decentering without any worry. Some other genres like astrophotography can require stricter standards.

What if all four edges look bad, but regarding equally so? Most likely, this doesn’t indicate that you have a poor copy of your lens. All corners may be slightly blurry when you run the test over simply because most lenses are not particularly sharp in the edges wide open – not a problem worth worrying about. It could also be that field curvature   is causing the corners to be slightly from focus and therefore less razor-sharp – again, not an indication that you have a bad copy of your lens. Some unsharpness will be normal; what isn’t normal is if each corner has dramatically different levels of blur.

If you happen to have got multiple copies of the zoom lens at your disposal, I recommend running the test above with every copy. Then, you can make sure that all are relatively similar to one another in sharpness. Keep in mind that there is always several sample variation from lens to lens. It is not really worth worrying about minor differences among copies. I’d only be concerned if one copy of the lens is clearly and consistently worse than the others.

At the end of the day, your odds of getting a bad copy of a lens are fairly low. Cheaper lenses usually have more sample variation than expensive lenses, but also then, it’s rare. We test dozens of lenses each year at Photography Life, usually with multiple copies of each lens, and only about 1 or 2 per year are so decentered that we’d return them.

If you’re at all worried, follow the steps above plus you’ll figure out whether there is a good copy or not! And if you’re not concerned, stay that way. Decentering certainly can occur, but bad copies of the lens are pretty rare and usually very apparent when they occur – actually without performing the test over.

Leave a Reply

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