WhyFi wrote in post #17264295
These are some horribly inconclusive test shots. Go outside, shoot it at various apertures and try bracketing some shots around the hyperfocal for the given aperture - once you find out if your distance scale is off (and how much it's off by), you won't need to worry so much about focusing. Oh, and make sure that they're properly exposed.
yup, these are really bad test shots. need to shoot at iso 100, and as whyfi suggested, do it somewhere outside, under good lighting. personally I like to use AF under 10x zoom--the computer is usually a tad better than you can be at micro focusing.
that said, op, this lens has HORRIFIC quality control issues. 1/2 of them are decentered (exaggerating but might actually be true). I personally bought 3 copies, and 2 were extremely decentered to the degree that you can clearly see 1/2 the image is blurry without even zooming in. others complain that the various rings have quite a bit of play. etc. I wouldn't be surprised if your lens is bad.
I would never buy this lens used unless the seller personally guarantees that I can return it if I find problems.
but yeah, on my 5d, the 3rd copy, the only good copy, is extremely sharp, especially in the center. each copy was purchased at a different store (buydig, adorama, bh) so it wasn't a bad batch or anything--probably a systematic problem.
and as others have said, don't use this lens for crop. there are much better choices out there like canon's 10-18 or 10-22.