Please Help! AF microadjustment will be the end of me...

I'm having a mind numbingly difficult time with my Canon 17-55 2.8 and its autofocus. It's gotten to the point where I'm wondering if I should try sending it in to be serviced. Basically, all of my images look fine in the studio but once I get them into photoshop and zoomed to 100% they're embarrasingly out of focus, even when using manual focus and live view zoomed in to 10x. I spent the afternoon yesterday seeing if the built in microadjustment on my 7D would at least make it bearable, but no luck. Everything is just way way too soft to be acceptable.

My question is, what should I be looking for exactly when evaluating whether or not to send in a lens to be looked at? Let me know what type of image zoomed to what level and I'll post one here for analyzing. Also, I am not the original owner of the lens and I've never sent equipment to Canon before. Has anyone here had experience with that sort of thing or know about how much autofocus services will cost?

Thanks

ssuperdreww wrote in post #16178033
even when using manual focus and live view zoomed in to 10x.

Which means it's got nothing to do with the AF system and that AF microadjustment cannot halp your problem.

You need to post pictures and 100% crops with EXIF intact. As Frank says, if the problem persists with MF, then it is not an AF issue and MA will not help.

MFA scares me. I am a bit on the OCD side of things. My first camera with MFA WILL be my last excursion out of the house for quite some time.

Sounds weird, especially since you tried manual focus with live view.

Have you tried repeating the same experiment with another lens, just as
a sanity check to see if the problem might be related to the body?

Ok, so I took a few quick shots this afternoon with my Sigma 10-20mm and my Canon 17-55 2.8 on the same 7D body. Included below is the EXIF data and corresponding images for each. From left to right, Sigma 10-20 using AF / then manual focus, and Canon 17-55 using AF / then manual focus. Let me know what you think. Is it just me or are these images too soft to be acceptable?

Looks normal to me... But my laptop screen isn't great for judging photos.

i don't see any issues. the 100% crops seem to be what i would expect based on the original image, that's a pretty hefty crop, you are bound not to pick up a super sharp level of detail.

im trying to figure out how MF is a problem, if you zoom into 10x and focus, that's your focus, it wont be off when you press the shutter unless the camera moves.

Sirrith wrote in post #16183494
Looks normal to me... But my laptop screen isn't great for judging photos.

mike_311 wrote in post #16183511
i don't see any issues. the 100% crops seem to be what i would expect based on the original image, that's a pretty hefty crop, you are bound not to pick up a super sharp level of detail.

Ok, i suppose that's the answer I was looking for. Somehow I just started second guessing my image quality and went off on an unnecessary tangent. That's what I get for processing huge batches of images on a weekly basis, tends to get a little mind numbing

Thanks for the reassurance!

mike_311 wrote in post #16183511
im trying to figure out how MF is a problem, if you zoom into 10x and focus, that's your focus, it wont be off when you press the shutter unless the camera moves.

I just included the manually focused images for comparison, just to make sure it wasn't a problem with one of several elements: lens problem, body problem, user error, etc...

Thanks again for the help!

They look pretty close but one thing I do notice, the 17-55 is about 1/2 stop more exposed than the 10-20. Not sure if it's the light changing or the Canon's glass better transmission properties made that big a difference.

ssuperdreww wrote in post #16183628
lens problem, body problem, user error, etc...

And the subject/target problem as well.
For such tests, it is highly recommended to use a high contrast target and avoid the one you have used.
In spite of that the images do not show any issue which is good thing..
So relax and enjoy..

Yeah, I think of 100% like looking into the electron microscope equivalent of an image. Fine for technical questions, but nothing artistically usable at that level. Though it appears the massive 200 2.0 can render clean @ 100%, that's a hell of a price to pay for pixel level perfection!

I wouldn't expect anything sharper than what you are getting with what you have.

Step 1...upgrade glass to L
Step 2...upgrade to full frame, perhaps a 5D3 or 6D.

Either one will give you improved sharpness. Both will do wonders.

I agree that those images are in focus, but, shooting something flat (parallel to the focal plane) at f/7 will not give you any indication of focus issues. There is simply too much DOF, especially at 20 and 55 mm.

Below is a focus test I did last week. Without setting up something like this you don't really know if you have a focus issue. The b/w target is parallel to the focal plane and the ruler is at a 45 degree angle. You are supposed to set the camera on a tripod and at a distance of fifty times the lens' focal length.