Canon EF 28-135mm IS Lens
Compared to Other Consumer Zoom Lenses
Primes and Canon L-Zooms

Note: This is a dated review based mostly on film camera performance. A more recent review of the EF 28-135mm f/3.5-5.6 IS comparing it's digital camera performance with that of the EF 24-105mm f/4L IS and the EF 24-70mm f/2.8L lenses is HERE.

This report briefly summarizes my review of four moderate range consumer zooms for Canon autofocus cameras. I compared the performance of several consumer zooms lenses against each other as well as Canon L-zooms and Canon prime lenses.

Consumer zooms compared from left to right:
  • Canon EF 28-105mm f/3.5-4.5 II USM (US grey market cost $210 USD)
  • Canon EF 24-85mm f/3.5-4.5 USM (US grey market cost $280 USD)
  • Canon EF 28-135mm f/3.5-5.6 IS USM (US grey market cost $385 USD)
  • Tamron SP 24-135mm f/3.5-5.6 Aspherical AD (US cost with rebate: $380 USD)

Lenses were tested for resolution, distortion, color and contrast. Color images of similar subjects - mostly buildings and landscape subjects were taken with transparency film. Resolution was tested on a USAF 1951 resolution pattern as described with either Technical Pan film developed in Technidol or with Provia F 100 film. Lenses intitially were tested at focal lengths of 28mm, 50mm, 85mm, and 135mm. A later test was conducted at 24mm after I had acquired a second EF 24-85 lens.

Testing results in common. All of the lenses tested gave excellent contrast and color in transparencies in the focal length range from 24 or 28mm to 85mm. Image resolution was good to superb in the portrait and landscape application tests that follow. Image contrast decreased noticeably in the EF 28-135mm and more so in the Tamron 24-135mm lens at 135mm. Linear distortion was substantial in all of the consumer zooms and most severe at their longest and shortest focal lengths.

Lens Resolution Tests

Results at 24mm:
All of the lenses tested performed moderately well. The prime lens predictably performed the best of the three lens tested when shot wide open. The Tamron lens was not tested. I was surprised to find that the 24-85mm lens outperformed the 17-35mm f/2.8 L zoom at this focal length.
Results at 28mm:
All of the consumer lenses gave reasonably good performance when stopped down to f/5.6 or f/8. Excellent resolution performance was achieved with the EF 28-135mm lens shot wide open. The EF 28-105mm and Tamron 24-135 had to be stopped down 2 and 1 stops respectively to achieve good resolving power.
Performance of all of the consumer lenses paled in comparison to that of Canon's outstanding 28-70mm f/2.8 L zoom lens. Barrel distortion was noted with all of the zoom lenses and was objectionable in images with straight lines produced by the EF 28-105mm and the EF 28-135mm lenses at 28mm.
Results at 50mm:
All of the consumer zooms have good resolving power as 50mm lenses shot wide open except the EF 28-105 which needs to be stopped down. Surprisingly, the EF 28-135mm lens outperformed the 28-70 lens in resolution from f/4 on, although linear distortion was still a problem with the 28-135.
Results at 85mm:
Resolution performance is disappointing for all of the consumer zooms wide open at 85mm compared to a prime lens. Performance improves if stopped down 1 or 2 stops.
Results at 135mm:
Resolution and contast leaves much to be desired in the consumer zooms at 135mm. However, in portrait application the Canon 28-135 still produced excellent images when stopped down and images were tweaked in Photoshop. .

Comparison of Image Quality

How well do these lenses perform when you take photographs of people, landscape and general photography? Amazingly well. See the illustrations below.

Lens Resolution Test in Portrait Application

In photographic tests with a mannequin with either film or digital camera, the consumer lenses did an outstanding job at f/8 for the most part. You could see greater softness in the film images with the consumer lenses shot wide open, but you had to look close. With digital cameras, the differences were negligible because of the lower resolving power of CCD/CMOS chips versus transparency film. My wife looked at the series of test photos and indicated that she only wanted me to photograph her with the EF 24-85 and digital cameras. If you look at the mannequin photos, I took with film and digital cameras and the various lenses here and here, you can see what I mean. The consumer zooms at f/5.6 and f/8 and digital cameras produce beautifully soft photos with excellent color. I don't know many women who want to you to take photos of them showing skin pores and other details of skin texture.

The major area where the consumer zooms don't perform in portrait application is in the capacity to create a shallow depth of field and wash out background clutter. See tests with portrait lens at 85mm and 135mm.

Lens Resolution Test in Landscape/Architecture Application

Other than the problem of linear distortion, these lenses produce images that are indistinguishable from more expensive zooms and primes when shot at apertures around f/16

Lens Resolution Test in Landscape/Architecture Application

Summary Opinion

It is unrealistic to expect great optical performance out of a variable aperture consumer zoom lens designed as a compromise. Less than excellent resolution at wide open apertures, substantial linear distortion, and contrast drop off at long focal lengths are to be expected. Nevertheless, there are times when you can only take one lens, and each of the consumer lenses tested has its good points.

Canon EF 28-135mm f/3.5-5.6 IS
This was the overall best performer of the consumer zoom lenses tested. All things considered, I would have preferred having a lens that was 24-105 than 28-135 because I always want wider angle perspective (especially with a D30 and the 1.6x magnification effect of the small CMOS chip). Image stabilization is a handy feature, although I have rarely needed it. This lens appears to be well built, and it is the lens I usually have stuffed in my briefcase or suitcase when I can only take one lens.

Tamron SP 24-135mm f/3.5-5.6 Aspherical AD
I really liked this lens. It appeared to be well built and had good contrast and color. I loved the 24mm perspective with film or digital cameras. Why did I sell it?
It has horribly slow and hesitant autofocus function with Canon EF bodies. This wasn't a significant problem with a good film camera like the EOS-1v. However, coupled with the pathetically deficient autofocus function of the Canon D-30 in low ambient light, the lens became almost unusable. I couldn't even use it to take snapshots in my living room with the D-30. It was very slow and hesitant to establish focus, and I was missing too many spontaneous moments. Too bad....

Canon EF 24-85mm f/3.5-4.5
Construction quality is very poor on the first sample of this lens I had. The first one I bought was so poorly constructed , lens barrel segments woobled when extended to 85mm. Optically, it outperformed the 28-105 except at 85mm. The first finally broke with normal use and from carrying it around in a bag with my D-30. The second one seems slightly better. The lens offers good contrast, color, fast autofocus function. If you own a D-30/D-60 or a EOS-1D, the wider angle perspective of a 24mm lens is really handy. On a D-60 it offers an equivalent perspective to a 38mm lens. On a 1D it offers a perspective equivalent to 31mm. I take this lens along instead of the 28-135 when I think I might be shooting people in confined spaces with a digital camera and won't really need anything longer than 85mm.

Canon EF 28-105mm f/3.5-4.5 II
This is a lot of lens for $210 USD. It needs to be stopped down to get a modestly sharp image. It is still a great lens for snapshots that you don't plan to enlarge past 8x10 inches or if you are photographing landscapes at f/16 and higher on a tripod. The lens I had was well constructed.

Equipment Review Index


Updated 8/9/02; 8/17/02

© 2002, William L. Castleman