Canon lens test

mercphoto

Well, I attempted a lens test. Not sure if my approach was correct or not, but I wanted to compare the Canon 28-135/IS, the Canon 24-70/2.8L, and the much less expensive Tamron 28-75/2.8 lens.

Results are here. Two pictures per lens, one at each end of the zoom range, and always wide open. The gallery is not protected and original size is viewable and downloadable.

http://mercphoto.smugmug.com/gallery/424595

I haven't viewed them enough to form my own opinion yet. Would be interested in what others think (including if my method was flawed).

pathfinder

mercphoto wrote:

Well, I attempted a lens test. Not sure if my approach was correct or not, but I wanted to compare the Canon 28-135/IS, the Canon 24-70/2.8L, and the much less expensive Tamron 28-75/2.8 lens.

Results are here. Two pictures per lens, one at each end of the zoom range, and always wide open. The gallery is not protected and original size is viewable and downloadable.

http://mercphoto.smugmug.com/gallery/424595

I haven't viewed them enough to form my own opinion yet. Would be interested in what others think (including if my method was flawed).

Interesting series merc. Three useful and widely owned lenses you are comparing.

Were these shot handheld? What was the shutter speed? Were all the RAW files processed exactly the same?

Short of using a tripod and mirror lock up, I am leery of ascribing too much significance to these kinds of tests. As lenses and cameras get better optically, it is harder and harder to discern differences withut using meticulous camera handling technique - which means fast shutter speeds, good tripods, mirror lock up or timed release ( not shutter depressed ).

The real difference I notice between the Tamron 28-75 DI and the Canon 24-70 L is with backlit subjects. The Tamron I own is very sharp, I doubt I could tell which lens I used for most images with without looking at the EXIF data. BUT the Tamron definitely does have more of a green flare when shooting back lit subjects .

The Tamron is also smaller and lighter than the Canon as well as cheaper, and I prefer to use the Tamron on my 20D because the size and weight seems more balanced for the 20D.

I prefer the 24-70 L on my 1DMkll for exactly the same reason - ergonomics and balance. So I use them almost interchangeably as far as optics go, but ergonomics does influence my usage.

mercphoto

pathfinder wrote:

Were these shot handheld? What was the shutter speed? Were all the RAW files processed exactly the same?

Yes, because I will use them mostly hand-held. Shutter speeds were 1/50-1/30 for the slow lens (with IS turned on). For the 2.8 lenses it was 1/80 to 1/100. ISO 400 for all. The converter was Canon's EVU, set to the equivalent to Parameters 1, sRGB, JPG level 4, for all files. No exposure or white balance changes, etc. I could have shot them in-camera JPG. I only shot RAW in case the white balance was off. I seldom shoot RAW, so again I processed the files in the same way the camera would have.

I understand your leariness due to the lack of tri-pod. Part of my rational was "how will these look in the manner I will actually use them?". And it might be that when handheld there isn't a huge difference in them.

My salesman pointed out that many prefer the Tamron due to price and size, as you pointed out. He did think the 24-70/2.8L was a bit more contrasty and had better shadow detail than the Tamron. Whether its worth the cost is up to the user.

DoctorIt

I just pulled up the original size at 28mm for all three lenses. Um, is it me or is the Tamron lens HORRIBLE!?! Didn't have to look too hard to notice the sharpness was just not there on that lens. I was really expecting the 28-135 to be the worst of that bunch, but suprisingly not.

I don't have a huge screen, so I just viewed the top of the image, mainly the Canon logo and the sticker to the right of it with the barcode.

pathfinder

DoctorIt wrote:

I just pulled up the original size at 28mm for all three lenses. Um, is it me or is the Tamron lens HORRIBLE!?! Didn't have to look too hard to notice the sharpness was just not there on that lens. I was really expecting the 28-135 to be the worst of that bunch, but suprisingly not.

I don't have a huge screen, so I just viewed the top of the image, mainly the Canon logo and the sticker to the right of it with the barcode.

Doc - I opened the original files of the 28-75 Tamron f2.8 and the 24-70 L and compared them on my desktop - the images must be about 2x3 feet in size. The Tamron looks softer in parts than the Canon, but then the lower right corner looks better in the Tamron version. I would not call the Tamron horrible - look at the size of the image - it would make a very nice 13x19. I have dozens of 13x19 inch prints shot with a 10D and the Tamron 28-75.

I think there is a difference in how much reflection is captured off the shiny cardboard printed surface. The image of the black camera itself is MUCH better with the EOS lens than the Tamron, but the lower right red print is blurrier with the Canon . That is the difficulty with these kinds of tests unless the are rigorously carried out. I think either one would make pretty fair images - as I said I own and use them both


I went back and examined the two images again - the Canon is MUCH sharper along the top than the Tamron, but the reverse is true along the bottom - makes me wonder if the top is in better focus for the Canon , and the bottom in better focus for the Tamron. That would explain a lot and it looks like the sign is not perfectly vertical, but slightly angled.....You have to be so VERY careful about lens tests - OH, I think I said that already

bkriete

Pathfinder-I saw the same thing you did. On the 24-70@70 the detail on the box is slightly better, while on the 28-75@75, the detail on the box is worse but the camera is unbelievably more clean and clear. I think it must be an artifact of focusing (I'm viewing both at original, so it's not from compression).

DoctorIt

bkriete wrote:

Pathfinder-I saw the same thing you did. On the 24-70@70 the detail on the box is slightly better, while on the 28-75@75, the detail on the box is worse but the camera is unbelievably more clean and clear. I think it must be an artifact of focusing (I'm viewing both at original, so it's not from compression).

i was playing it up to focus as well, but it's weird. Why would it be so different? I know the Tamron has great reviews, but this test just doesn't show it. Regardless, I'm pretty happy that the 28-135 is clearly capable of taking photos that are pretty darn close in quality to L glass.

Andy

mercphoto wrote:

Yes, because I will use them mostly hand-held. Shutter speeds were 1/50-1/30 for the slow lens (with IS turned on). For the 2.8 lenses it was 1/80 to 1/100. ISO 400 for all. The converter was Canon's EVU, set to the equivalent to Parameters 1, sRGB, JPG level 4, for all files. No exposure or white balance changes, etc. I could have shot them in-camera JPG. I only shot RAW in case the white balance was off. I seldom shoot RAW, so again I processed the files in the same way the camera would have.

I understand your leariness due to the lack of tri-pod. Part of my rational was "how will these look in the manner I will actually use them?". And it might be that when handheld there isn't a huge difference in them.

My salesman pointed out that many prefer the Tamron due to price and size, as you pointed out. He did think the 24-70/2.8L was a bit more contrasty and had better shadow detail than the Tamron. Whether its worth the cost is up to the user.

merc

your tests work for your style - but imo they aren't conclusive for comparision purposes - i believe that a lens test s/b done on a tripod, with release, mlu, etc. too many variables in handheld sh ooting

this from a guy who doesn't test lenses i just buy 'em, and shoot. if the results in production aren't right, i return it to b&h for another.

thanks for your post though, it's very informative

Andy

DoctorIt wrote:

i was playing it up to focus as well, but it's weird. Why would it be so different? I know the Tamron has great reviews, but this test just doesn't show it. Regardless, I'm pretty happy that the 28-135 is clearly capable of taking photos that are pretty darn close in quality to L glass.

yup, especially that copy you got

of all the canon lenses, there seems to be the most variation in quality wiht that darn 28-135 ... people either love it or hate it! i highly recommend it on a 1.6x body.