Testing a lens

[Gibmeister]

I would like to take each of my lenses and see what there capabilities are. How well they perform at different combinations of f-stops, shutter speeds and focal lengths. I was hoping that someone could give me some guidelines to use in the process. Two of the lenses I would like to work with are a Sigma 28-70 EX 2.8 Aspherical and a Sigma 70-200 2.8 APO. Any suggestions would be helpful.

Gib

[rdenney]

I usually take a picture of a smooth brick wall with slanting sunlight. Point the camera straight on to the wall, so that the lines are parallel. Focus carefully. When you enlarge the image, you'll be able to tell how sharp the corners are with respect to the center. Any barrel distortion will be seen as curved lines. Vignetting will be obvious by the darker corners.

Then, put a gray card in a scene, and shoot that card using slide film over the entire range of f-stops. Any inconsistencies in the aperture operation will turn up, but I would not really expect any. You use slide film so that the differences will be visible and not automatically corrected by the machine doing the prints at the lab. Any colorcasts will be visible on this one, too, but you must compare against a known-good lens to control for errors in setting the white balance (or to control for color casts in the film). You can do the same thing on a digital by using manual exposure and not changing the exposure between shots. Then view on a monitor with no corrections or post-processing. The thumnails should tell the story.

Stand a series of wood screws or nails on their heads at an angle to the line of sight on the lens, and focus on the center nail. Take a picture with the aperture wide open. If the center nail isn't the one in focus, then you have a focus error.

Then, make some images of people, and see if you like the way the images look. Yes, it's subjective, but if you don't like the way they look, then the numbers don't matter.

Rick "who is doing this right now for a couple of lenses" Denney

[Gibmeister]

Thanks Rick. I am using a Canon 10D to test these lenses. I will give these a try. Would you do anything with the ISO setting on the 10D and would you return all of the settings back to the original factory settings?

Gib

[rdenney]

I'd use ISO 100 and a tripod to remove any non-lens variables. And I would use RAW to avoid mistaking JPEG artifacts for lens faults, though I have to confess that with minimal compression, the large JPEG settings I've tried don't seem to produce any visible artifacts than I can find.

I would definitely shoot with appropriate white balance, but again, I'd compare to a lens of known color cast to make sure. I'd leave the image processing settings (i.e. sharpening, contrast...) at their 0 positions. I do this anyway because I use the Adobe 1998 color space to simplify my work flow with Photoshop, and that setting forces all the post processing to their default settings.

But I'm not sure any of that matters. You can probably tell what there is to tell no matter what your settings.

Bottom line is that if you like the prints you get, then it's a good lens. The tests I suggested are just designed to expose faults quickly.

Rick "who thinks the most important thing is not to change settings in the middle of the testing" Denney