Yesterday I bit the bullet and purchased an Olympus d-400Z. I took some shots around the office and quickly transferred them to my PC. I was very suprised to see highly visible distortion in my wide angle pics, enough to be distracting.

I have read that this Oly had *very* good distortion qualities, so this came as a suprise. I even thought that maybe I had gotten a defective camera. Then I noticed that the Camedia software had a 'distortion' filter, which fixed the problem - this led me to believe this is a wide-spread issue, and not something specific to the camera.

Whats going on? I see nothing like this on the 'sample images' page here...

Carlos

-- Carlos Morales (morales@slip.net), February 12, 1999

Answers

The 400Z does have some barrel distortion at the wide-angle end of its range, particularly right at the edges of the image (top and bottom are much more noticeable than the sides). At the telephoto end, there's an almost imperceptible amount of pincushion. Overall, the distortion is at the lower end of the range we've observed in most digicams. (With very few exceptions, they all have barrel distortion at the wide-angle end of the zoom range, many much worse than the 400Z, a few much better.)

If you've got straight lines close to the top or bottom of the frame, you're definitely going to see barrel distortion. Also, of course, there's the normal perspective distortion of a wide-angle lens, for objects close to the camera. If that's what you're seeing, it's characteristic of wide-angle, rather than a lens defect. There is indeed a "distortion" filter in the Camedia software, which is intended to correct for barrel distortion.

As a check on what you're seeing, check out the "viewfinder" test images on the Comparometer - they should give you a very good idea of what's normal for the camera (or others) at both wide-angle and telephoto focal lengths.

-- Dave Etchells (web@imaging-resource.com), February 13, 1999.

Another comment, and idea that just occurred to me: We should probably *measure* the distortion on the viewfinder test target, and present it in the review as an objective number - percent deviation across the frame, or something equivalent. That way, people would know *exactly* what they're getting, without us resorting to vague adjectives, etc. Make sense, worth the hassle of doing?

-- Dave Etchells (web@imaging-resource.com), February 13, 1999.