I've been looking to pick up an ultra wide angle lens for two upcoming trips, one to DC and one to the Boundary Waters Canoe Area Wilderness in Minnesota. Both have photographic opportunities that would require an ultra wide to fully utilize. My two main lens candidates were the Sigma 10-20 and the Tokina 12-24, already reviewed here. I went with the Sigma because I could get it from a local store so I could easily return it if I wasn't happy with it, plus it had the potential to be superior to the Tokina in a few ways, in particular having less of a problem with chromatic aberrations. Part of the reason I was so set on being able to return it is that Sigma has a reputation for poor quality control with the 10-20. Reports on Fred Miranda frequently mention one specific problem, a centering defect that means the left side of the image is softer than the right. Always the left. However the word was that if you got a good one, it was a REALLY good lens. First off, as far as physical build goes the lens feels very good. This was my first experience with a Sigma EX series lens, and it feels roughly comparable to a Canon L lens (mind you I've never used an L lens in that size, so take that comparison with a grain of salt). The zoom and focus rings are both nicely damped, and the lens barrel has a nice texture that makes it satisfying to hold. I haven't seen the Tokina in person so I can't compare the two, but I read several reports that said that the Tokina was superior in build quality. That seems difficult to believe after I've had my hands on the Sigma, it left little to be desired. It was at least superior to the Tokina in that it employs Sigma's HSM (hyper sonic motor), apparently more or less the same thing that Canon calls USM. A high speed focusing motor that's nearly silent. Look, HSM doesn't take better pictures, although the faster focus could theoretically give one an edge. But it just makes the lens feel nicer, also the quieter lens can be less obtrusive if you're in quiet surroundings. And like USM it allows you to focus manually while the lens is in autofocus mode. I spent my first days with the lens both trying to test the image quality and also going out and trying out the style of ultrawide photography, something I really had no experience with previously. My initial results were mixed. Some shots looked very good, others... not so good. As an example I offer this picture of our "Millennium carillon Tower"... [attachment=1] a building I've never really warmed to, it was supposed to be our salute to the millennium but it was expensive and never fully funded, in 2008 it's still not completed, they're still trying to get the rest of the money to finish it and in the mean time I just think it sounds awful. You can play the thing like an organ, they have a keyboard that can control it and planned to have special performances. But I just don't think that a bunch of enormous bells do not sound good, instead of harmony it sounds more like a cacophony. But enough about the tower, about the picture. I have two 100% crops of specific sections of the image to accompany it. Keep in mind that the shot was taken in portrait orientation, so top and bottom are left and right. First off we have this crop of a section near the center of the frame: [attachment=2] If it looks a little soft, keep in mind I'm using a custom picture style with low sharpness. But just compare it with a crop from near the top of the tower: [attachment=3] Noticeably nastier, isn't it? That was my first sign of trouble. Yet it wasn't all bad, I got some really good results too. [attachment=4] I spent a while trying to figure it out. My testing seemed to show that if I took one shot of a scene (my standard test was a long fence, it was a straight line with lots of hard edges that I could use to check for sharpness) and then flipped my camera upside down and took another shot and then rotated the images so they both looked rightside up, the left side would look sharper in the upside down shot (suggesting the centering problem was present) but the right side looked the same in both. My reaction to that was basically "WHAT????". I returned the first copy of the lens and got another to see if the problem improved any. Instead both sides just became evenly soft. But I'm not sure if that was really a step backwards or not. In any case, around that time I realized what was wrong. It was front focusing like mad. Basically, it never came close to focusing to infinity, and also wouldn't focus to any specific distance, the same object at the same distance could get a variety of focus settings. If I put it into manual focus mode and focused almost to infinity (I left a little leeway because the depth of field had to be pretty significant on a lens that wide focused that far) then the soft edges mostly went away. I also got some weird shots where the left side seemed to be focusing on close objects and the center and right side seemed to be focusing farther away. Again, I know not if that really meant what it looked like or not. My mind was made up for me when I went into Chicago and did some testing in the big city, where I had previously completely failed at photography before because 28mm just wasn't wide enough. When I manually focused to infinity I got some decent results (although with occasional apparent softness anyway) but I must have knocked the focus off and didn't check it, when I got into Union Station and took a picture of the cavernous hall that serves little purpose now that everything is underground I got a big blurry picture. Shrunken down to web size it's not all that noticeable, but it wasn't good. [attachment=5] And here's one last bonus shot from outside Union Station: [attachment=6] In the end, as I may have alluded to by referring to the lens in the past tense, I returned the second copy as well. It has the potential to be a good lens, especially if put it on manual and focused to infinity. But part of the reason I got it was that it had the HSM focusing system, and what good is that if I leave it on manual all the time? Knock a few hundred bucks off the price tag and make it a manual focusing lens and then we can talk, but for $500 I'd like it to autofocus correctly. It also had a fairly significant barrel distortion. Horizontal lines suddenly bend near the upper and lower edges of the frame. It's not noticeable in pictures with no strong horizontal lines, it only showed up in test pictures, but I still wasn't happy with it being there. All ultra wides have some degree of distortion, but I think I'd have preferred a more regular sort of distortion that was even over the whole frame. I'd guess that that sort of problem would be easier to correct. It wasn't an easy decision to make, it still had potential to be a good lens. But I was frustrated at the front focusing and apparent inconsistent performance (I still don't know if that was my fault or the lens's fault) and I also had the fact that I'd just seen a review of the Tokina 11-16 F2.8 to content with. Photozone.de rated it very highly, they called it the best ultrawide zoom available for Canon APS-C cameras. In terms of sharpness it beat the Tokina 12-24 and Sigma 10-20, even wide open at F2.8. It costs a little more and has a comparatively restricted zoom range, but I still think it sounds like a better alternative, even without considering the wider maximum aperture. Which might help in focusing precision as well. My camera has an extra high precision focus point that only works with F2.8 or wider lenses, besides which reducing the depth of field has to make focusing a bit more precise anyway. Even with my focusing problems and the inconsistent softness problem I'm not quite able to give the lens a thumbs down, I don't know that the same problems would happen to others, perhaps my camera in particular just didn't like it, but my conclusion is that the Tokina 11-16 sounds like a better option for me at this point. My only hesitation is that I'm uneasy at the limit of 16mm at the long end, I probably did spend most of my time at 10mm but I know I got at least a few good shots at 20mm. It should be a reasonable tradeoff, though. PS: I'm using a new feature I found in the forum software that lets me attach pictures but then have them appear in the middle of the post rather than just at the end. However it seems to make them also appear at the end anyway. If that's annoying let me know and I'll stop doing it and go back to just referring to the pictures as they appear at the end. [This attachment has been purged. Older attachments are purged from time to time to conserve disk space. Please feel free to repost your image.]
Dan, Interesting post. I was just looking into getting a telephoto and wideangle lens for my Canon A630 (instead of upgrading to a DSLR that I probably dont need ;D). I was doing some research and am not impressed at all with my findings. Seems like everyone is saying that unless the camera lens is zoomed to full length, there is some vignetting present. Also there is kind of bubbling effect like someone put a magnifying glass in front of the camera. Ugh. Technology.
Oh that's funny. See, I was going to jump on to that post talking about the difference between dSLRs and point and shoots, and one of the things I was going to say was that with dSLRs you can always get a new lens to give the camera new capabilities, but a point and shoot is pretty much limited to the lens it comes with. I was going to say that while you can get add on converters that make the lens wider or longer, I had doubts about the quality of the results. dSLRs have the same thing, you can add, say, a 1.4x teleconverter to add magnification to a telephoto lens (although the converter goes between the lens and the camera, not on the end of the lens) but not all lenses respond well to that. Some do, though, the 400mm lens I'm thinking of getting is supposed to do well with a 1.4x and even possibly with a 2x. And I was going to counter balance that by pointing out that many dSLR lenses can cost more than an entire point and shoot camera. In my case, the lenses that I REALLY want cost many times more than any point and shoot camera. But we're talking $5000 or more, I'll be able to seriously consider getting them when I win the lottery or find out that a long lost relative died and left me the family fortune. All that flexibility certainly doesn't come for free.
I've read mixed reviews of this Sigma lens, too. Which is why I rented the Tokina 12-24 last fall. I am planning on getting the Tokina 11-16 based on my experiences with the 12-24. Good review!
dan, that is one of the main reasons that i am a snob for 1st party lenses. there just seems to be better quality controls there. i do agree that once you get a good 3rd party lens you can get a great bang for the buck, but it doesn't seem worth the hassle of hoping you get a good one. just my 2 cents.
I can think of a bunch of people that would disagree with that statement. Perhaps in the past, but not now. It seems all the manufacturers have let their lens manufacturing controls go downhill.
Well, my Tamron 28-75mm has been a decent lens. I know some people have had trouble with it, but it worked fine for me (I bought it in a local shop so I had the return option then too, I just didn't need it). The Tokinas don't have the same reputation for poor quality control, which is why I'm more willing to consider picking one up mail order (whenever B&H or whoever start stocking them). The Canon 100-400 IS has a reputation for variable quality control too. For example, Photozone tested two copies of the 100-400, one was below average and one was quite good. On dpreview you find some people showing off their extremely high quality copies of the 100-400 that perform nearly as good (or rather 100% crops of pictures taken with them) as the 400mm F5.6, sometimes potentially beating it. But it's typically stressed that those are the cream of the crop of that lens, and you have no guarantee of finding one. It's problematic for me, because if I could get one of those golden 100-400s I'd get it, I'd very much like the IS and the additional flexibility of zoom, but I want the absolute best, including compatibility with teleconverters, and I'd be trying my luck with the 100-400. I've actually considered bringing my laptop to the store and doing image tests while I'm there if they have more than one copy available. The same goes for photozone's review of the 300mm F4 IS. They mention testing two copies, and stress that the second copy that tested well had to be serviced TWICE to get to that point.
dan, that 100-400 L is next on the old wish list. glad to see it got some of the great reviews that it did.
I do have a 3rd Party Tamron lens in my bag. After my experience with it, I'm having a hard time trying out another 3rd Party lens. To me buying 3rd party lenses is like buying a CD. You just never know whether or not you are going to like the whole thing or not.