Daan -- From the additional shots you've supplied, I'd say you have an excellent copy of this lens. I don't mean to lower your standards, but I don't think mine does any better wide open, and clears up almost completely stopped down one stop, like yours.
I agree also that it is connected with flaring of the brightest light, not CA or SA (not a main component, but possibly contributing).
If it doesn't suit your needs, go ahead return it, as I doubt you will find much better performance in any zoom, and most decent quality primes in that range.
Daan -- From the additional shots you've supplied, I'd say you have an excellent copy of this lens. I don't mean to lower your standards, but I don't think mine does any better wide open, and clears up almost completely stopped down one stop, like yours.
I agree also that it is connected with flaring of the brightest light, not CA or SA (not a main component, but possibly contributing).
If it doesn't suit your needs, go ahead return it, as I doubt you will find much better performance in any zoom, and most decent quality primes in that range.
Yes, I was afraid of that. It seems I have a 'normal' copy of this lens and the glowing is part of the deal
Maybe I have unrealistic expectations. I never owned a Canon WA zoom before. I was expecting performance more along the lines of the 24-105L and 24-70L at comparable FL's. These lenses don't show this glowing. well, maybe my 24-105L does at 105mm f/4 a bit.
On the practical side: when I need 2.8 there isn't much light, so the glowing is kept to a minimum. I actually have 27 2.8 shots that hardly show any glowing. These aren't taken in hard contrasty light. When I shoot in the sun there is plenty of light to stop it down, so the 'issue' can be avoided that way.
You are probably right that other lenses will show similar behavior. I do wonder how the Tokina 16-28 2.8 holds up against the Canon 16-35L II for this. Or maybe the new Tamron 24-70 2.8 is better suited for my needs. I have to think about it.
Back in 2007 when this lens was released I tested it against the previous version and it fared better, though IMO, my copy leaves a lot to be desired for across the frame sharpness. Anyway, the previous version showed this edge glow even more than your samples do, and IIRC with the previous version the edge glow was also bad at 35mm until stopped down one stop. I really believe Canon optimized this lens for 16-20mm, though 35mm seems better than the previous version.
If you can, it might be worth testing other copies. Back in 2007 I went through three before I settled for the one I have now, though it has never been excellent IMO. I've had it at Canon several times, which gave me the opportunity to test the loaner copies, which all tested better overall than mine (though I was more interested in edge to edge and left side vs. right side sharpness consistency)...
This lens is one reason I diversified into a rangefinder set-up for wide angle work...
S Dilworth wrote:
Spherical aberration causes focus shift when stopping down.
From what I can tell focus shifts. Most pronounced at around 28mm. 2.8 shows a bit of a front focus, 3.2 has focus shifted to the back being right on target. f/3.2 also seems to resolve any glowing / softness issues at 28mm.
Daan B wrote:
The lens is brand new with a pristine front end. I don't know if any elements are out of alignment. How do you check for that?
The simplest way is by side-by-side comparison with another copy in a careful test on a tripod.
Failing that, one could set up a test from two users separated by long distance with a repeatable light source like an electronic flash at a specified ISO, shutterspeed, and apertures and focal lengths of a scene that both can duplicate (perhaps lens boxes or common consumables) at a carefully measured distance.
At this point, I think a comparison is required, since others have reported some symptoms more or less the same, so degree of symptom needs comparing. A comparison should be less bother than shipping it to Canon for evaluation.
Daan B wrote:
Maybe I have unrealistic expectations.
I think that your assessment is likely correct. When I looked at that photograph, I saw a fine photograph. When you asked about a specific problem and shared a 100% crop, I looked closely and thought, "Well, OK, if I look really closely I can see the think he is seeing, but it is a complete non issue in an actual photograph."
If you were to print this at 18 x 24, no one would ever notice or complain about this "issue."
An assignment that I often give to people who become overly obsessed with what they regard as being "perfection" in a lens is to go to some galleries or shows displaying excellent prints by fine photographers. Go and enjoy the photography as it should be enjoyed, but in addition to that go ahead and take a close technical look at the work. You will go away relieved, to a large extent, of this notion that all lenses must be "perfect" in all cases.
At 100% magnification, you are looking at the equivalent on your screen of a portion of a print that would be 5 or 6 feet wide. That is a huge print size. Huge prints are essentially never sharper than (and often not as sharp as) what you are getting from this lens.
Just because you can find one thing (one of those other lenses) that may be "better" in some way than this excellent lens, it does not mean that this lens is poor. It is actually working very well, and aside from poring over 100% magnification crops on the screen - an unusual way to display real photographs! - this lens will produce excellent and quite large photographs.
Dan
By the way, if the only way that a small difference between two lenses can be detected is to conduct a test in which every single variable but the one you are looking at is completely controlled and to then have viewers switch back and forth several times between side-by-side 100% crops while trying to find a flaw... there is no difference worth worrying about. It is utterly insignificant in real photography.
