Canon 28-105 vs. Canon 28-135

Introduction.This page shows a comparison between the Canon EF 28-105/3.5-4.5 USM and Canon EF 28-135/3.5-5.6 IS USM lenses. Both of these lenses are popular choices for the Canon D60 and Canon D30 digital cameras, and seem to be regarded as good consumer lenses. I wasn't sure what sort of quality to expect from these lenses, but I've been pretty disappointed. Below are the results from my "newspaper" resolution test.

Methodology. All images were taken using a Canon D60 digital camera, mounted on a tripod, using mirror lockup and self timer to eliminate any camera shake. The target was a newspaper page, taped to a wall, illuminated by a mixture of daylight and halogen lighting. The chart below shows cropped portions (from the center and edges) of several test images taken with the two lenses at different focal lengths and different apertures. (Because the lighting was slightly uneven, I applied some brightness correction to the crops taken from the edges of the images. For this reason, this test is not good for judging light falloff or vignetting. My feeling is that Canon 28-135 shows slightly more vignetting, but these pictures should not be used to judge this.)

Results. I encourage you to look at the images below for yourself, but here is how I interpret these images.

• Both lenses are pretty blurry at wide apertures at all focal lengths. For my purposes, neither of these lenses are usable at wide apertures. In fact, I was unable to find any combination of aperture and focal length where the edges were as sharp as the center.
• The Canon 28-135 seems to perform slightly better at 28mm, particularly at wide apertures, but still isn't very good. Given that I wouldn't want to use either lens at wide apertures, I consider this superiority to be of academic interest only.
• The Canon 28-105 is clearly better at 105mm. The edges in particular are much sharper using the Canon 28-105 lens than the Canon 28-135 lens. Even at F11, the edges of the Canon 28-135mm lens are considerably softer than the center.
• At 50mm, things seem pretty even. Both are soft at the edges wide open, although the center of the 28-135mm lens is slightly sharper at wide apertures.
• The canon 28-105 shows less contrast, and the images are slightly "warmer" (more of a red tint).
• Although this is not apparent from these test shots, I felt that the Canon 28-135 was not uniform in its softness. At wide angle, the right edge of the image appears softer than the left. At longer focal lengths, the top of the image is softer than the the bottom. In contrast, the softness of the edges of the Canon 28-105 was roughly uniform. For these crops I picked areas from the softer edges of the Canon 28-135, so these may represent a "worst-case" scenario.

Conclusions. For me, image sharpness is important. Because I wouldn't want to use these lenses at much less than F11, the image stabilization (IS) feature (although it seems to work remarkably well) of the Canon 28-135 is of limited use. Even with IS, a tripod is a necessity at F11 to produce really crisp images in anything but brilliant sunlight. My conclusion is that the $200 Canon 28-105 is a better buy than the $400 Canon 28-135.

Final Thoughts. After reading almost uniformly good user-reviews of these lenses (particularly the Canon 28-135), I was pretty surprised to see the results. I'm still not sure what to make of all the people who claim that they produce "fantastic" images with these lenses. I guess expectations differ wildly, or perhaps Canon's quality control is such that some lenses are much better than others. Either way, I don't think either of these lenses come close to bringing the best out of a camera like the D60. I guess there really is a reason why the Canon "L" lenses cost 5 times as much and cover half the focal length.

I've also tested the Canon 28-105mm lens against the Canon 50mm/1.8 lens.