I recently picked up a 16-35L on the forums and had the opportunity to take a few test shots with the 17-40L and 10-22, while I still had all three in my possession since the 17-40L will be sold shortly. I originally started out with the Sigma 18-50 f/2.8, but no longer have it, so I'm unable to throw it into the mix.
I took the following tests using tripod, remote release, mirror lockup on a 20D. The raw files were converted straight with ACR without altering. Took a set at the wide end (16mm or 17mm) and another set at the long end (35mm or 40mm). Took crops from the centre and from the corner. The corner crops on the wide tests are right in the very corner because 16mm and 17mm are close enough. The crops from the long test are not in the extreme corner because there was a substantial difference between 35mm and 40mm, but I'd say the crop was taken from the near corner.
These tests were only for my informational purposes, while I still had all three lenses in my possession. I just thought I'd share them since a lot of people do ask about all three lenses.
The more pics I see of the 10-22's performance, the more I am reassured I made the correct move trading my 17-40L for it. Thanks for doing these tests klam!
Thanks for posting the test results. I know how much time that takes.
Based on my experiments with 16-35 and 17-40 lenses, there is more inter-copy variation than there is any absolute "average" difference in image quality between the two.
There seems to be some sort of issue with at least a few of your test pics. The 17-40 at 17mm f/4 should not be that soft. --c
Agreed. I think there was a misfocus or something there. I'll delete those tests with the 10-22. charlesk wrote:
There seems to be some sort of issue with at least a few of your test pics. The 17-40 at 17mm f/4 should not be that soft. --c
Have found that my copy of a 16-35 produces great results on a 1D Mark I, D30, D60, 20D, and 30D. It doesn't do nearly as well on a 5D, especially in the corners. My 16-35 on a 5D performs worse than my 10-22 does on any of the Canon EF-S-compatible cameras.
Sarsfield wrote:
I was about to say that about his 17-40. Mine is way sharper than that at f/4.0
I'd say you have a good copy of the 16-35L and a bad copy of the 17-40L.
I had all of these lenses. My copy of the 16-35L was worse in the corners. The 17-40L was the same as the 10-22.
Date code on the 16-35L is UQ0101. Date code on the 17-40L is UT0303. Liquidpics wrote:
Looks like you have a lousy copy of the 16-35mm in my opinion. Mine was sharp wide open. What's the date code?
Maybe instead of uncovering bad copies of lenses your test illustrates what has already been said by many out there: the 16-35 is twice the cost for an extra stop and not necessarily as sharp, and that the 10-22 is (gasp, don't say this about an EF-S lens--since everyone knows they'll be obsolete in no more than six months) a great performer. If I had full frame I'd get a 17-40. If I was a reporter and needed the extra stop and the company was paying for it, I'd get the 16-35. If I was a professional and shot landscapes I'd get the 17-40. If I owned a 20D/30D and did anything I'd get the 10-22 and *definitely* not look back.
I think the results of your test speak for themselves. Turn this thread into a sticky.
From the MTF's, the 10-22 > 17-40L > 16-35L. Which is interesting, given that Canon's own data indicates that the most expensive lens is the poorest performer.
I think for many people the high cost of the 16-35L is what puts them off. It costs twice as much as the 17-40L but is not any better or only slightly or maybe even slightly worse. I think my tests shots show that. However, despite that I will keep the 16-35L. I find the f/2.8 invaluable and makes this lens so much more usable for me than an f/4 lens.
I think its the same arguement that is made between the 24-70L and 24-105L. You are paying a lot for the extra stop. Whether that's worth it to you or not is up to each individual. I don't think an extra stop is worth double the cost, but what can we do that's how Canon priced it...
bikewithadam wrote:
Maybe instead of uncovering bad copies of lenses your test illustrates what has already been said by many out there: the 16-35 is twice the cost for an extra stop and not necessarily as sharp.
I think the results of your test speak for themselves. Turn this thread into a sticky.
-Ridiculous.(frustrated here) That's is exactly what's being done- uncovering bad copies. The 16-35mm prior to date code UT08 was hit or miss. The copies after UTO8 (just read the reviews here) were sharp wide open throughout the entire focal range. The test is simply inaccurate and no way reflects a good copy of the 16-35mm which out performs both the 17-40L and the 10-22. It's funny how the people that haven't actually tried the lens or can't afford to obtain one always tend to bash it. They dismiss any positive reviews, just dwelling in the negative. The glass will always be half empty for them.
Edited by Liquidpics on May 06, 2006 at 06:29 AM GMT
If you get standard quality lenses which factually peform within their design envelope, the three lenses are not far apart wrt optical performance. However, in my experience the 16-35 rules, closely followed by the 17-40 and then the 10-22.
Overall, those three lenses are sufficiently different....that's why I have all three and I do use them for different purposes. True, the 17-40 emulates the 16-35 IQ rather well, yet obviously there is one stop difference. And that's not merely any old difference between those two lenses: lest we forget, the nominal lens aperture is what breaks the bank, not its build, AF drive or IS.
Liquidpics, I've tried the 16-35; I owned a used copy for a short period of time. Why did I get rid of it? The 17-40 shot images that were just as sharp--if not moreso--for half the monetary investment. Don't get me wrong, the 16-35 is a great lens if you need that extra stop. But I didn't, and if the lens is being used for landscapes, then you're using a tripod, and an extra stop doesn't matter, and even if it did, you wouldn't be shooting landscapes wide open. Plus the 16-35 is significantly larger than the 17-40, and while that may be good for gear envy, I count that as a slight mark against when it comes to practical usage because I travel a lot. But for others who don't it may not matter.
The 16-35 is worth it if you're someone like a reporter shooting wide open, handheld and at wide angles. Then the extra stop is indispensible. It's not ridiculous, it's just logical.
