Sigma 8-16mm Review (Long with many images)

Started Jun 8, 2010 | Discussions thread
OP Severian The Lame Contributing Member • Posts: 550
Sigma 8-16mm Review: Centering and Focusing

Centering and Focusing

  • Two other copies of this lens featured in threads in the Canon lens forum exhibited symptoms of centering problems as did the copy reviewed by photozone. One other forum member posted a thread with photos from his gallery but there was no mention of centering and I haven’t examined them to find out. (Also some threads on this lens in the Nikon lens forum but I haven't paid much attention to them.) So that’s at least three out of four with centering issues –not a good sign. You can now make that at least four out of five; though in my case (and apparently that of LSHorwitz1) the lens’ symptoms are fairly benign and mostly cured by stopping down to f/8 plus I've discovered a work around when shooting at infinity as in the following examples.

  • In the examples below I shot at 8mm at f/4.5, f/8, and f/11. The first shot at each aperture is with the 7D’s battery grip in normal portrait position and the second with it upside down. (As a curious side note, in my tests the 7D’s “evaluative” metering overexposes 1/3 of a stop when upside down! And in case you're wondering about the obviously incorrect time stamps on the photos: the camera's clock is set to UTC.)

  • I had a hell of a time trying to get all of these handheld shots (almost) identically framed. Trying to do the same trick with the camera mounted on a tripod presents its own difficulties. But rotating the camera isn’t really necessary; all you actually need is a frame filling subject with a proper degree of rotational symmetry and fine detail. So I tried next to shoot an ISO 12233 test chart ( ). To fill the frame at 8mm I had to stick the front of the lens about 6 inches from a chart printed at 19x13 inches. And that caused me all kinds of lighting difficulties. Not to mention problems getting everything aligned, etc. etc. If you look carefully at the resultant test shots you’ll soon see that not everything is kosher. Hey, best I could do, especially given that the whole procedure was so mind-numbingly tedious that I just couldn’t face doing it again. My hat’s off to all of you that do this kind of thing for a living. (Question: how do you stay awake?! )

  • All of the aforementioned problems and limitations of the test methodology mean that my results are not even remotely definitive, though they do show to my satisfaction that a problem exists. As a lawyer might say, my results are indicative but not probative.

  • I won’t post all that tedious mess here, but the curious can view it in all its glory at

  • One thing I have found is that (on my copy at least) this problem is focus related. I have shot many landscapes where I don’t notice any issue at all but have virtually identical shots of the same subject where the problem is clear. The only difference I could discern by examining EXIF data and fiddling around with camera and lens settings has to do with the recorded subject distance, i.e., exact position of the focus ring. I would have thought at 8mm and f/8 for subjects mostly at infinity that the DOF would make precise focusing not that critical. But my copy’s defect means that’s not true. Yeah, it’s true for the center, but not for the periphery, especially the upper right corner in landscape orientation. Taking a close look at the lens internals as seen through the front element and referring to Sigma’s published diagram ( ) you can see how the inner focus groups of elements work and how they move back and forth and rotate.


Beyond Infinity

  • Given that the center stays sharp while one side of the image varies my guess is that the problem lies squarely with these elements and is related to their rotation and not to their travel and that in the full image circle the misfocusing is still present; it just gets rotated off of the upper right corner of the rectangular sensor. It would be really interesting to test this: all I would need is a full frame camera and a replacement for the lens that I and my Dremel tool would destroy! Do I have any sponsors? A few test shots can be found here:


  • Sigma lenses often make autofocus an adventure. So how does this lens autofocus on the 7D? It’s hard to say for sure but at close focus distances of a meter or less it seems to work OK. At longer distances? In one sense it’s all over the place. I use either single point or spot focus mode on the 7D and I can get what I’m focusing on in focus with any of the 19 points. (OK, OK. I mostly use the center point and shift to others as necessary, so I haven’t methodically checked every point.) But refocusing on the same point in the same scene can get the lens hopping back and forth as shown in the focus window, sometimes quite significantly. And the indicated focus distance can differ substantially between normal AF and Live View AF (contrast) even though the results in terms of subject sharpness are the same.

 Severian The Lame's gear list:Severian The Lame's gear list
Canon EOS 7D Canon EOS 5D Mark III Canon EOS M Canon EOS 7D Mark II Canon EF 100mm f/2.8L Macro IS USM +11 more
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow