Torture Test: Do Small Scratches in Lenses Prevent Sharp Images? We Gouge a Zoom to Find Out

Does a 5mm scratch in the center of the front element ruin a lens? We gouged a brand new, straight-from-the-box, $499 Tamron 28-75mm f/2.8 zoom to find out. Then we shot some test subjects. Was the defect visible in the images? See for yourself…

I’ve done some difficult and painful things in my life, but digging into the front element of a new $500 lens with a tungsten carbide awl gave me pause. It felt so wrong. But in the pure pursuit of knowledge, and with a heartfelt thanks to Tamron for providing the victim, I forced the tool to strike.

And strike it did. It was more difficult than I’d anticipated, but I managed to cut a furrow about 1/20th of an inch, or roughly 5mm, near the center.

Before committing this mayhem I shot a selection of images at different combinations of f/stops and focal lengths. Experience (and the laws of physics) suggested that if the flaw were to be detected it would most likely occur at the widest focal length, in this case 28mm, and at the smallest aperture, herein f/32. So I needed reference images.

After the act of destruction I shot the same regimen, again using small and large f/stops at wideangle and telephoto settings. Then I set about shooting general, miscellaneous subjects just to see if the scratch would appear under normal, everyday circumstances.

The rest of the story you can see for yourself. Every image in this article—except for the two images of the lens itself—was shot with the damaged lens. 

From this angle it’s hard to tell that the scratch is more-or-less in the center, but you can see it’s for real. ©Jon Sienkiewicz

This image was captured with the defaced lens. Even if you knew exactly where the damage was located, you wouldn’t be able to find it. ©Jon Sienkiewicz

Same as above, shot with the scratched Tamron 28-75mm f/2.8 zoom (which is a very nice lens, BTW). ©Jon Sienkiewicz

I thought the scratch might show up as a blip on this white wall, especially since the exposure was 1/400 sec at f/10. So far, I cannot see any evidence that the lens has been brutalized. ©Jon Sienkiewicz

Fence and sign look fine. ©Jon Sienkiewicz

No problem with these yellow and white ranunculus flowers. ©Jon Sienkiewicz

As you can read in the metadata, this was shot at f/16 in the 28mm wideangle position. The flaw is visible, but only if you know where to look. For the record, the blob did not appear when I shot similar scenes at larger apertures. ©Jon Sienkiewicz

At f/32 the aberration is clearly visible, and there is no denying that the lens is damaged. However, based on this experiment, I’m prepared to defend the assertion that unless you shoot at a small aperture and at wideangle, a single scratch, even one as large as 5mm, is virtually unnoticeable. ©Jon Sienkiewicz

Here are the fence and ranunculus again, only this time I’ve pasted a section from the previous image so you can see where you should be able to see the flaw. Can you see a flaw? I cannot. ©Jon Sienkiewicz

—Jon Sienkiewicz