50 1.2L + 5D3 Non-Center Point AF = Problem?

Hi,

So, I've been having a love-hate relationship with my 50L since I bought it a few months ago. The pictures are amazing when the focus is on, but the problem is that it's almost never focused right!

I almost never use the center point in actual use since for compositional purposes I almost always want the point of focus somewhere else. I've gotten pretty good with the joystick for choosing AF points, so I pretty much never focus with the middle and recompose, thinking using the right point will be more accurate. And this has usually been good for me with other lenses (like my usual go-to 24-70 ii). But the 50L has been killing me! Driving me nuts!!!

So, when using the center point, my 50L seems to require 0 MA in broad daylight. Great. Then I switch to the outer points and then everything is front-focused (as in the focus is between the subject and my camera right?). It appears that the points adjacent to the center are acceptable, and close in focus to the center point enough to be within error. But all other points clearly front focus. Again, this is in daylight, outside.

Note that I have all non-cross points disabled in my camera.

I've sent this camera + lens to Canon CPS 2x already. The first time they said nothing was wrong with either and did not test them together. The 2nd time they noted that AF was wrong in the body and "made adjustments", and found nothing wrong with the lens again, and tested together they found the combo worked fine after adjustment.

Well, CPS seems to be wrong.

I've searched this subject a little and came across a few threads similar (field curvature issues, 5d3 outer point issues, etc.), but not really specific to the 50L. Anyone know if this is a known issue with this lens? Or should I send it in to CPS again?

On another note, maybe a related issue, maybe not, but I found that in my bedroom at night, using halogen lighting, the MA needs to be -7 for the center point. And while shooting a dance in a conference room the MA needed to be +7 for the center point. WTF??? So frustrating... (anyone have this sort of issue too?) And yeah, in those situations forget about using anything other than center-point.

I feel stupid since I bought this lens for low-light situations, but it performs so much worse in those cases compared to my other lenses. What's the use of having f/1.2 if it can't focus right when it gets a little darker?

I had a lot of problems with with 5D2 but with the mark III all the 21 cross points seemed fine in all conditions. Sure I'd have to +10 MA but that doesn't bother me at all. It has small amount of focus shift between the f1.6-f2.5 at the danger zones but I've learnt to work around it.

I had to send my 85L twice with the body. Maybe a third time's a charm for you. Try talking directly with their customer support before sending it in, if you haven't already done so. Good luck!

sportmode wrote in post #16656164
I had to send my 85L twice with the body. Maybe a third time's a charm for you. Try talking directly with their customer support before sending it in, if you haven't already done so. Good luck!

thanks... were your problems similar?

IMO, the 5D3 outer points are still not as good as center. I've had similar troubles with the 50L and 85L. Granted, it was only 1 body I used, it wasnt having the same accuracy of center point. Single point seemed more accurate than zone, but would love for zone to work due to the simplicity and quickness of moving around the points.

Charlie wrote in post #16656300
IMO, the 5D3 outer points are still not as good as center. I've had similar troubles with the 50L and 85L. Granted, it was only 1 body I used, it wasnt having the same accuracy of center point. Single point seemed more accurate than zone, but would love for zone to work due to the simplicity and quickness of moving around the points.

To be specific in my case the outer points consistently front focused by consistent amounts. So technically they were accurate but invalid. Was this the case for you?

Common issue I think with some lenses. Two copies of the 16-35 share the same behavior with mine, 17-40 and 35L (only one copy though). I think more common with wide lenses and field curvature issues that AF can't correct for. Live view should look perfect though, correct?

I sent my 16-35 in to Canon to try and get fixed but same behavior when it got back. They did adjustments of some sort. No decentering in the lens itself that I see, plus two copies acted the exact same... Rules out lens issues for me. I think something with the 5D AF system.

None of my telephoto lenses have issues though... I checked with dot time with my old 24-105 that the outer AF points on both sides front focused compared to center, until about 35-40mm. Then it went away. You could never really see it in shooting though due to the f/4 and 24mm depth of field.

It's weird the 50L has this issue as well, but it may be an optics thing with that lens. I tend to see it at 35mm and less myself.

Xyclopx wrote in post #16656310
To be specific in my case the outer points consistently front focused by consistent amounts. So technically they were accurate but invalid. Was this the case for you?

didnt take notice if front or back focused, but whatever the amounts, dont think MFA would have help, not sure. I delete all my bad raws, so sorry cant help beyond this.

Xyclopx wrote in post #16656174
thanks... were your problems similar?

I had inconsistent focusing initially, sent it in, and they fixed the center point but outer points were all over the place. After the second trip, everything was pretty spot on. I think this is probably common with really fast glass. Canon supposedly made changes on both body and glass both times.

so i did more testing this morning with my backup camera, the 7d. it appears on that camera most points AF similarly. however, the farthest left and right points do front focus a lot, just like the 5d3.

i know this lens has curvature issues, but should it matter to the AF system?

since the spacing of the 7d's AF points are closer to the center than that of the 5d3 (i think) perhaps that's why most are okay except for the farthest points?

Here's a check with the 85L at 1.2 with center, right, left, and bottom-left cross-AF points.

Center

IMAGE LINK: http://s192.photobucke​t.com …nter_zps7aa2539​8.jpg.html

IMAGE LINK: http://s192.photobucke​t.com …Left_zpsc8c1fe9​2.jpg.html

IMAGE LINK: http://s192.photobucke​t.com …ight_zps5fb1641​7.jpg.html

IMAGE LINK: http://s192.photobucke​t.com …Left_zps182a866​5.jpg.html

I often use focus recompose on the 5D3 because the AF points are a bit unreliable on the outer points particularly for the 24LII and the 50L

I had a 50L that did the same exact thing on my 5D3. Sent it to Canon twice, and while the AF was improved using the non-center cross-type points, they were still front focusing enough to where it wasn't reliable at any MFA setting. I'm really leaning toward it being a 5D3 issue, as it works perfectly fine on my friend's 5D2, and my other 3 lenses work perfectly fine on my 5D3.

I would have sent the body in along with the lens, but they wanted to charge me $280 for an out of warranty "repair", but I didn't want to mess up the flawless performance it has with my other lenses.

Xyclopx wrote in post #16655804
Hi,

So, I've been having a love-hate relationship with my 50L since I bought it a few months ago. The pictures are amazing when the focus is on, but the problem is that it's almost never focused right!

I almost never use the center point in actual use since for compositional purposes I almost always want the point of focus somewhere else. I've gotten pretty good with the joystick for choosing AF points, so I pretty much never focus with the middle and recompose, thinking using the right point will be more accurate. And this has usually been good for me with other lenses (like my usual go-to 24-70 ii). But the 50L has been killing me! Driving me nuts!!!

So, when using the center point, my 50L seems to require 0 MA in broad daylight. Great. Then I switch to the outer points and then everything is front-focused (as in the focus is between the subject and my camera right?). It appears that the points adjacent to the center are acceptable, and close in focus to the center point enough to be within error. But all other points clearly front focus. Again, this is in daylight, outside.

Note that I have all non-cross points disabled in my camera.

I've sent this camera + lens to Canon CPS 2x already. The first time they said nothing was wrong with either and did not test them together. The 2nd time they noted that AF was wrong in the body and "made adjustments", and found nothing wrong with the lens again, and tested together they found the combo worked fine after adjustment.

Well, CPS seems to be wrong.

I've searched this subject a little and came across a few threads similar (field curvature issues, 5d3 outer point issues, etc.), but not really specific to the 50L. Anyone know if this is a known issue with this lens? Or should I send it in to CPS again?

On another note, maybe a related issue, maybe not, but I found that in my bedroom at night, using halogen lighting, the MA needs to be -7 for the center point. And while shooting a dance in a conference room the MA needed to be +7 for the center point. WTF??? So frustrating... (anyone have this sort of issue too?) And yeah, in those situations forget about using anything other than center-point.

I feel stupid since I bought this lens for low-light situations, but it performs so much worse in those cases compared to my other lenses. What's the use of having f/1.2 if it can't focus right when it gets a little darker?

50L is a hot mess of a lens. It renders beautifully, and its bokeh is unmatched by any of Canon's other offerings....but the focus shift is something you have to contend with.

And MFA, calibrating the lens the camera, etc etc isn't a solution. It's inherent in the design of the lens. It lacks a floating element which allows for more uncorrected spherical aberration (creating that beautifully dreamy look), but it wrecks havoc on AF systems.

mystik610 wrote in post #16659222
50L is a hot mess of a lens. It renders beautifully, and its bokeh is unmatched by any of Canon's other offerings....but the focus shift is something you have to contend with.

And MFA, calibrating the lens the camera, etc etc isn't a solution. It's inherent in the design of the lens. It lacks a floating element which allows for more uncorrected spherical aberration (creating that beautifully dreamy look), but it wrecks havoc on AF systems.

Thanks but I wasn't stopping down. All tests were at 1.2 so it's not because of focus shift.