Luminous Landscape Forum
Equipment & Techniques => Cameras, Lenses and Shooting gear => Topic started by: jimhuber on April 22, 2007, 05:43:47 pm
-
I ordered the new Canon 16-35 f/2.8L USM II as soon as I could find one in stock and immediately evaluated it against my "old" 17-40 f/4L USM on my 5D body*. At very wide focal lengths (16mm to 20mm) and wide apertures (f/2.8 to about f/5) the extreme corners were better on the new lens, but between the center and the corner the new lens wasn't as good as the old one. At very wide focal lengths (16mm to 20mm) at f/5.6 and smaller the lenses were practically identical. In my tests at focal lengths of 24mm, 28mm, and 35mm the old lens beats the new one at all apertures. Neither lens has much chromatic aberration, but the new 16-35 has a little less. Distortion seems pretty equal.
I was in a quandary for a while because I also have the 24-105 f/4L, which is excellent at it's wider focal lengths except for distortion at 24mm. So generally I'll use the 24-105 lens unless I need to go wider than 28mm, and wider than 24mm is where the new lens is strong. But most of my use of focal lengths that short is outdoor landscapes mounted on a tripod with apertures smaller than f/5.6 so the new lens wasn't of any benefit to me except for the slightly wider field of view of 16mm versus 17mm (98 versus 93 degrees horizontally). That alone isn't worth $1,700 USD to me (lens plus B+W filter). Nor do I like having a zoom lens that's only usable through half of it's focal length (the new lens is terrible at 35mm and not good at 28mm). I also wasn't impressed with good corners when the area between the center and corners was mediocre. Overall the 17-40 f/4L USM is, in my mind and based on my use, a better lens than the 16-35 f/2.8L USM II without even considering the price difference.
These tests are very close to my results:
The-Digital-Picture (http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?FLI=0&API=0&FLIComp=0&APIComp=0&Lens=100&Camera=9&LensComp=412)
* on a Gitzo 1348 tripod with Really Right Stuff BH-40L ball head, Really Right Stuff L plate on the camera, with mirror lock up and self timer.
-
Mine just arrived as well and so far it's looking good.
The corners are much better and the center looks about the same as my 16-35 I, the distortion is also the same.
It's getting it's first run on a job tomorrow, i'll let you know how it goes..
-
i'm not inclined to buy an early production lens from Canon (or anyone else)
and the tolerance variations even in later production lenses appears to be significant
from my own testing, the lenses i have appear to be average
now i'm interested in doing better - does anyone have a good strategy for multiple lens purchase, test, and return? same vendor? multiple vendors??
i sincerely hope that Canon is not relying on the new 16-35 to solve their wide angle problems -- i'd be really happy with a 19 or 20mm f4 with real optical excellance (why should i have to buy a Zeiss with adapter and lose autofocus?)
in the mean time, if i have to buy a 17-40, i'd like a really good one
-
Why not just buy one and if it is iffy send it to canon for adjustment?
-
Why not just buy one and if it is iffy send it to canon for adjustment?
[a href=\"index.php?act=findpost&pid=113729\"][{POST_SNAPBACK}][/a]
Because you can't adjust bad glass. The only thing they will do is adjust things like autofocus and the like. If one of the elements is inherently off a bit they dont replace them. I have been down that road and it absolutely amazes me. I ordered a 180 Canon macro. I went to shoot newspaper on the wall and the lens had a terrible dark gradient from the top edge of the frame to about 15 percent in. It was an obvious and glaring defect. Nothing was loose inside, it was shipped in this condition! I bounced it back to the dealer for another one. I did not even consider a repair. I felt more like throwing it through Canons quality control office window
I don't know about the rest of the folks here but I would gladly pay an extra $100 per copy of their lenses to get quality control up to where is should be. When we spend $1500 on a lens we should not have to spend a day performing tests to make sure its a good copy
-
Ah. I've had tamron adjust a lens for me. Worked great. Was no more expensive than paying shipping back to the place I bought it from. I take it canon isn't so desperate for our business?
-
I have never bought into this bad copy stuff. It is a bit of a myth and misinterpretation. What has happened on some lenses is that an element is de-centered in assembly or the auto-focus is set wrong to create a back or front focus -- these can be adjusted by Canon. This has happened to me in the past and it was corrected, (both issues have happened to me). Ideally it should not happen at all.
I did receive the new 16-35f2.8LII and it is better in the corners than the original -- particularly at the wider lengths. It is corner, and not center, improvement that Canon claims. At the long end there is not much difference.
I see the improvement all the way as I stop down -- the old lens never catches up.
-
There is such a thing as variable performance from one copy to another of the same lens model. Canon has ranges of quality tolerances even for L lenses, and from my experience dealing with them here in Canada they will not adjust or exchange lenses that fall within the range.
-
"I have never bought into this bad copy stuff" I tested five 17-40 lenses before buying one. Three were about the same, one was 10% better and the one I bought was 20% better. I don't care what causes this, it shouldn't happen.
-
"I have never bought into this bad copy stuff" I tested five 17-40 lenses before buying one. Three were about the same, one was 10% better and the one I bought was 20% better. I don't care what causes this, it shouldn't happen.
[a href=\"index.php?act=findpost&pid=113879\"][{POST_SNAPBACK}][/a]
Dennis, while you may think it shouldn't happen, as you've seen it does and it will continue to do so. The reason is not wholly technical, but primarily commercial. The best quality requires the best possible materials (not only glass, but also moving parts), very tight tolerances on manufacturing and very tight standards that guide the rate of rejects (remember each reject costs them a few lenses to recuperate). The more severe all this is, the better and more consistent the quality, the higher the cost and the higher the price. They make judgments about this trade-off with an eye to the market. The 17~40 L (and I own one - not so sure how proud I am of that fact) is the cheapest L lens Canon makes, and the results reflect the price. People lucky enough to get the 10% or so that are "spot on" get a tremendous bargain. The rest of us got what we paid for - maybe. That's how it works. Because of the letter "L" there can be mismatches between expectations and outcomes which frustrates customers, but Canon is not a company that cares about frustrated customers - unless they are CPS - and even then. I'm not saying any of this is good - I'm just calling the shots from my experience and what I've learned from that experience.
-
The other part of the issue with lens disappointments is that it is almost always regarding retro-focus wide angles for SLR's, particularly ones with faster apertures, and seldom for longer lenses.
The technology for this type of optic simply lags behind what we can get in say a 135f2L or a 300f2.8. These are lenses that we hear few negative comments on or concerns over a "bad copy". Not just prime lenses but also longer fast zooms are much less often criticized.
You must still expect a degree of compromise with retro-focus wide angles and unfortunately can not yet expect the same level of optical excellence we have become accustom to in longer lenses. There was a time when most lenses over 200mm were at least a bit soft before advanced glasses became so common.
This is why a rangefinder based short lens has held an edge by avoiding or minimizing the need for a retro-focus design.
-
Saw the tests in the link posted (digital picture) compared to a few lenses at different settings and the 16-35II looks good. The 17-40 is an excellent lens also but it isnt 2.8, sometimes for creative or technical resons the larger apperture is needed. The new lens doesnt sacrifice much if anything at all for the extra stop. Looks good.
Canons f4 zooms are a great value and sharp. I have 2, the 17-40 and the 70-200 non is , use them constantly on my 5D and the results are really good. Even so the 5D shines with normal or short tele primes like the 50mm f1.4 and the 85mm f1.2L both will wow you at times. Those lenses are just MUCH more simple to make.
-
I hear you... I tested two of the new 16-35 zooms and neither of them impressed me. Back to square one.
-
At very wide focal lengths (16mm to 20mm) and wide apertures (f/2.8 to about f/5) the extreme corners were better on the new lens
[a href=\"index.php?act=findpost&pid=113676\"][{POST_SNAPBACK}][/a]
Forgive my humour but I'm impressed that the 17~40 f/4 was even comparable at f/2.8 ..... or am I missing something? I come from a school where lenses were selected for their overall uses and simply expecting a newer, faster designed to be better than a pretty recent but slower design seems a little odd to me. If you need f/2.8 then it seems to me that a comparison against the old version would be more useful. If f/4 is sufficient then its available at a bargain price. There are other aspects of using a lens which might be of great importance (such as a brighter viewfinder image) and the holy grail of 'ultimate' performance is not the be all and end all of lens design. As has also been said wide-angle performance still lags behind that of short telephotos. With the exception of the 20/2.8 (poor corners on FF digital) I own no poor Canon lenses and haven't seen any of the bad examples discussed so much on the web. From what has been said, it sounds like the 16~35 II is a pretty reasonable lens.....
-
Well mine passed it's first day at work with flying colors, the lens is very sharp across the whole frame, the corners are a big improvement over the last model...
Very happy..
-
I have never bought into this bad copy stuff. It is a bit of a myth and misinterpretation. What has happened on some lenses is that an element is de-centered in assembly or the auto-focus is set wrong to create a back or front focus -- these can be adjusted by Canon. This has happened to me in the past and it was corrected, (both issues have happened to me). Ideally it should not happen at all.
I did receive the new 16-35f2.8LII and it is better in the corners than the original -- particularly at the wider lengths. It is corner, and not center, improvement that Canon claims. At the long end there is not much difference.
I see the improvement all the way as I stop down -- the old lens never catches up.
[a href=\"index.php?act=findpost&pid=113869\"][{POST_SNAPBACK}][/a]
Aren't you kind of contradicting yourself? You don't believe in "this bad copy stuff" but yet you have first hand experience with a lens that you had to send in.
I'm curious as to whether you've ever done any comparisons between lenses of the same model because if you haven't, it's more of an opinion than anything else.
Along with some of the other posters in this thread, I've actually tested two different copies of the same lens and found significant differences between the two. I can't comment on how often this happens, but I'm sure it exists. L lenses are marketed as professional lenses and cost a significant premium over "normal" lenses so I, personally, expect every copy to be sharp. Consumers shouldn't have to test for copy variation...that's the job for Canon QA.
-
"Aren't you kind of contradicting yourself? You don't believe in "this bad copy stuff" but yet you have first hand experience with a lens that you had to send in."
By rejecting this "bad copy" issue, I am suggesting that the glass elements themselves are ground to ultra high tolerances and refractive specifications, but that on relatively infrequent occasions the assembly of an element group is what is called "off center". This can be rectified where as an element poorly shaped or out of tolerance with it's refractive index is not. I have had one lens off center and it was fixed. I have also had one lens where the auto-focus point was miss adjusted which was also fixed. These assembly issues should ideally be reduced as much as possible but it is not like they are passing off coke bottles for lenses on every second production run. Further, we are not paying Leica prices for these optics -- these are produced to be price competitive. I wish my ultra costly Blad lenses were as sharp as most of my Canon optics.
I have had experience with many Canon lenses having gone through at least $100,000 in purchases of their optics over the years -- and have had and used multiple copies of the same optic. They have been surprisingly consistent.
One problem that people do get caught up on is expecting a fast aperture ultra wide to perform like a well regarded mid range prime, and when it does not live up to this unrealistic expectation they call it a bad copy.
-
Photoguydon, I think what your saying is essentially correct: especially with zoom lenses it has been explained to me by technicians concerned that the problem is more an assembly issue than a glass issue, and there are limits to what one should expect with wide angle lenses at wide apertures. Also, it may true that roughly comparable Canon lenses are somewhat less costly than their Leica counterparts - if such comparisons can be made legitimately (I'm not sure - it's seldom an apples to apples comparison). But there is performance variation within the same model of an *L* lens, and it is noticeable, even within a range that Canon QA considers *acceptable*. They ARE doing the quality control - the issue is the range considered acceptable. Unless the lens falls outside that range, based on my experience dealing with them, Canon WILL NOT re-assemble your lens to get it to match the best of the range for that particular spec. Their philosophy is that once it's acceptable by their standards, its acceptable and if it's acceptable they don't tinker with it. What the customer thinks or wants is irrelevant.
-
one word....PRIMES !
-
I agree with you Doug about the primes. However......sometimes you just need a good zoom, particularly in the dusty environment of the SW where we both are-less lens changing. I was expecting something more from this new lens, a dramatic improvement over past ones, but I am just not impressed enough (for my uses anyway) to switch from the 17-40 f4 L. I am able to use lenses primarily at the mid apertures (on tripod) and there is simply very very little difference at f8 between the two.
Considering the price difference and the fact that these series II lenses are "rumored" to be designed to take advantage of the "rumored" coming 22 mp body resolution abilities. I don't see it. 22 MP would far exceed the capabilities of this and all L series lenses. Obviously the rumors are wrong in one respect or another.
-
dennishh, how did you determine 10% and 20% difference between lenses? test chart - if so what chart did you use?
did all 5 lenses have similar serial numbers? overall, the distribution of lens "quality" will be statistical, but within that there may be periods of time with parts at tolerance limits, machines and tolls at edge of wear limits, "monday morning" syndrome, etc. etc.
i'd be happy to hear anyone's recommendations for lens testing
-
one word....PRIMES !
[a href=\"index.php?act=findpost&pid=114355\"][{POST_SNAPBACK}][/a]
At 24mm the 16-35II is better then my 24 tilt or 24 1.4 ..
-
from what i've read, "better than the Canon 24 or 24TS" is not a great recommendation
if a variety of rangefinder lenses of demonstrably better quality can be adapted to Canon full frame cameras, and Canon can make a very good to excellent 24mm, why can't (won't) they make at least one excellent rectilinear 18, 19, or 20mm prime? it doesn't have to be f1.4 or f2, it "just" needs low distortion, good color, and sharpness compared to the rangefinder lenses
i think most photograpers paying for a lot of pixels will be willing to pay for glass that can make use of those pixels - offer a "L+" line with Leica quality and quality control
-
At 24mm the 16-35II is better then my 24 tilt or 24 1.4 ..
[a href=\"index.php?act=findpost&pid=114465\"][{POST_SNAPBACK}][/a]
At 18mm, an old Nikkor 18mm f3.5 is better than ANY Canon zoom I've compared it to at the same focal. And better than a Canon 24 prime is not saying much...
,
-
"At 18mm, an old Nikkor 18mm f3.5 is better than ANY Canon zoom I've compared it to at the same focal. And better than a Canon 24 prime is not saying much... "
I too use and love my old 18mm 3.5. I'm going to Samy's in the morning and will compare it to the new 16-35 but I'm not expecting much.
-
Here is a pretty though review:
www.16-9.net/lens_tests/index.html
-
Perhaps it's just time to accept that if you want an ultra wide zoom (like the 16-35) you are going to have to sacrifice some quality, especially with increasingly advanced digital sensors. At least if you don't want to break the bank and/or go crazy in obsession.
-
Hi guys,
My first post here, so please be gentle
I have the 16-35 2.8L II on test from Canon at the moment and have been comparing it to my 17-40 f4 L
From my field tests I pretty much concluded that the 17-40 was the better lens and that any advantage that the 2.8 "may" offer, was not work the extra coin.
Then I discovered something really weird with the 16-35 and I am starting to think that it's a dud and that I need to test another.
This image is a crop from a 16mm @ f4 sample. It looks like camera movement, but this is 1/1000 sec, tripod mounted, mirror locked up and timer delayed.
(http://www.potd.com.au/samples/lenstests/16-35_16mm_f4_TopMiddleCrop.jpg)
and just to prove that there is no movement, here is a section from the middle of the same image...
(http://www.potd.com.au/samples/lenstests/16-35_16mm_f4_Crop.jpg)
More samples can be seen here ADPOTD Forum (http://www.potd.com.au/forum/viewtopic.php?t=2631)