Sort of sounds similar to the problem the 1d3 had with the 35L. Lots of users reported that the combo just didn't play well together. You might just have an unlucky camera however because there are lots of initial reports of other photographers using this exact combo with much better results than you are experiencing. Sorry to hear about your misfortune.

What about manual focus? Sharp, huh? Is the problem simply related to focus point selection and relying on automation?

RogerC11 wrote:
Sort of sounds similar to the problem the 1d3 had with the 35L. Lots of users reported that the combo just didn't play well together. You might just have an unlucky camera however because there are lots of initial reports of other photographers using this exact combo with much better results than you are experiencing. Sorry to hear about your misfortune.

Thanks. It's been super frustrating. I never had issues like this with my 5D or 5D2, and it's taken a lot of the fun out of shooting. I am not usually the kind of person that takes pictures of test charts or focuses too much on my gear. But I've been forced into that because I'm not getting the results I got with my previous cameras and lenses, and I know there's something wrong but can't figure out exactly what.

Case in point: I'm borrowing a friend's OMD E-M4 + Panasonic 25/1.4. I'm getting a far higher keeper rate and sharper pictures than with my 5D3 + L and L-quality lenses, even with moving subjects. That does not seem right.

Paul Mo wrote:
What about manual focus? Sharp, huh? Is the problem simply related to focus point selection and relying on automation?

Not really. As I said before, I didn't have these issues with the 5Dc or 5DII, both of which had worse AF than the 5D3. I am not focusing and recomposing; I use other focus points depending on the position of the subject and where I want them in the frame.

Some of the test shots that seemed to come out okay were with the center focus point. Is it possible that the center focus point could work well but something could be wrong with the other focus points?

badlydrawnboy wrote:
... Is it possible that the center focus point could work well but something could be wrong with the other focus points?

My gut feeling is that yes, it's most assuredly possible. I'd second your idea of sending the camera to Canon.

Case in point: I'm borrowing a friend's OMD E-M4 + Panasonic 25/1.4. I'm getting a far higher keeper rate and sharper pictures than with my 5D3 + L and L-quality lenses, even with moving subjects. That does not seem right.


Sure it does, it's a badass combo (assuming you meant the E-M5 ). In good light, the AF is fast enough to AF-S moving targets, if they're also slow enough.

Try tracking something faster than a rock, though - it'll remind you why you bought that DSLR.

I do feel for your AF woes, though - Been there & done that, though not to this degree, and it can be incredibly aggravating, especially when you start using a CDAF rig that has near 100% accuracy.
I'm certainly not looking forward to this BS when I get back into DSLR territory - fingers crossed.


Also, it's quite possible your AF module could be mis-aligned, which would in fact throw off AF on certain points, while being sharp on others (also, you're assuming your lenses aren't decentered and sharp uniformly across the frame).

Since you mentioned Olympus, that reminds me of yet another awesome feature of their older E-5 DSLR. Not only would it MA zooms and apply varying levels according to zoom level (so Canon's zoom MA is actually way behind the curve), but you could apply different MA values to every AF point. Ridiculously awesome, though I'll be damned if I ever had the patience for that!

Good luck and I really hope you get things situated.

wow, feel bad for you badlydrawnboy, it must be frustrating!

At some point I think it would be wise to bail on the Sigma 35 if the OP cannot get it to work
with confidence. I remember the frustration and wasted hours testing my Sigma 50/1.4
and don't intend to go through all that again.

I am happy for all of you getting good results with your Sigma primes. Enjoy.

badlydrawnboy wrote:
Paul Mo wrote:
What about manual focus? Sharp, huh? Is the problem simply related to focus point selection and relying on automation?

Is it possible that the center focus point could work well but something could be wrong with the other focus points?


Could be; the 5D MK3 AF is complex. There must be some trick to it.

trumpet_guy wrote:
At some point I think it would be wise to bail on the Sigma 35 if the OP cannot get it to work
with confidence. I remember the frustration and wasted hours testing my Sigma 50/1.4
and don't intend to go through all that again.

I am happy for all of you getting good results with your Sigma primes. Enjoy.

Yeah, I had the same experience with both the 50/1.4 and the 85/1.4, though the 3rd time seems to have been the charm with the 85/1.4.

Here's what I'm going to do:
1. Send the 5D3 back to Canon.
2. Test the 35/1.4 again when I get the 5D3 back.
3. If the 35/1.4 is still front-focusing it, return it and re-purchase a 35L.

2a. Test your 85/1.4 again. Just in case....

Happy shooting,
Yakim.

Good point.

Frankly, I haven't tested it thoroughly enough yet to say it's okay. I just took a few quick shots. In the past I've had Sigma lenses that focused accurately at certain distances but not others. That's what's so frustrating about Sigma AF issues; they can't be fixed by MA.

I may have gotten this wrong, but were you doing MA adjustment at 50X focal length and then shooting at much shorter distances? With third party lenses in particular (but even Canon lenses to some degree) the AF algorithms can be different at different focusing distances. Have you tried MA adjustment at the distances you are shooting yet? I see it was mentioned earlier but didn't see the results.

Also (again, forgive me if I missed this but I didn't see it mentioned) you are shooting wide open aperture with all, correct? Because a couple of those lenses do have a bit of focus shift stopped down slightly. Not awful, but certainly might show up at those close shooting distances.

+1 roger (Infact I mentioned of the first page of this thread to do the focal MA at the same distance)

I have had an old sigma 300/4 HSM that had vastly different MA numbers at close and long range. As it was the sort of lens that I would shoot stuff at very similar distances I'd have the MA numbers noted down .


Roger: have you had a chance to play with the USB attactment? The way I understand it is that the USB widget and software should reprogram the lens to avoid the front/back focus differences at different shooting distances (or did I get that bit wrong)

badlydrawnboy wrote:
That's what's so frustrating about Sigma AF issues; they can't be fixed by MA.

The only Sigma lens I owned was the 50/1.4. I had it with the 40D (no MA) so I took it to Sigma's lab in Israel. The technician worked for 3 hours and in the end he could calibrate it only to the central AF point and only to distances 3-5m (he told me that I can choose the distance range, but not to have it calibrated to all distances). Frustrating is such a small word.

Happy shooting,
Yakim.

RCicala wrote:
I may have gotten this wrong, but were you doing MA adjustment at 50X focal length and then shooting at much shorter distances? With third party lenses in particular (but even Canon lenses to some degree) the AF algorithms can be different at different focusing distances. Have you tried MA adjustment at the distances you are shooting yet? I see it was mentioned earlier but didn't see the results.

Also (again, forgive me if I missed this but I didn't see it mentioned) you are shooting wide open aperture with all, correct? Because a couple of those lenses do have a bit of focus shift stopped down slightly. Not awful, but certainly might show up at those close shooting distances.

Thanks, Roger. I think that may have been the case with my first copy of the 35/1.4, i.e. it seemed to focus accurately at 50x the focal length, but it front-focused horribly at close shooting distances (which is what I normally use). When I say "horribly", I mean more than even a +20 AFMA could correct. But even if it could, it's no fun to pay $900 for a lens that can only AF accurately at one focal distance.

The jury is still out on this second copy. It is indeed front-focusing at closer distances, but not nearly as badly as the first one. For example, when I focus on my wife's eye it appears blurry but the eyebrow is in focus. I think that could potentially be corrected by MA, right?

But if it has the same problem where it focuses accurately without MA at 50x but not at 15x, it's going back to B&H and I'll probably just go back to the 35L or I may try one final time. After all, my third copy of the 85/1.4 does indeed seem to be sharp wide open and focusing accurately — though more testing is needed at different focal distances.

Yes, I'm definitely going to send the 5D3 into the Canon Service Center.

Consider sending it with the lens(es) with which you want it calibrated, and ask Canon to make sure they're good.

kevinsullivan wrote:
Yes, I'm definitely going to send the 5D3 into the Canon Service Center.

Consider sending it with the lens(es) with which you want it calibrated, and ask Canon to make sure they're good.



And you belive Canon will calibrate his sigma lens

Okay, did some testing with the second copy of the 35/1.4. I set up the LensAlign target, dialed in ISO 100, f/1.4 and a shutter speed of 1/400. I started very close to the target (i.e. about 1 foot), and then took a big step back for each shot. As you can see, the front-focus becomes progressively worse as I step back from the target. These were all center-point focus.









For comparison, here are 4 shots from my Canon 50/1.4. This lens was sent to Canon for AF problems. They discovered that the focus assembly was broken and fixed it.









Strangely, the third distance with the 50/1.4 is the only one with a slight back focus. I took two shots at each shooting distance, and both were back focused at that distance. But it seemed to be accurate again at the next step back. I'm not sure what to make of this?

I also did the same thing with the Sigma 85/1.4. Unlike the 35/1.4, which is front-focusing badly, this one is back-focusing. However, it looks to me like it's mild enough that it could be corrected with AFMA?

FWIW, I'm having similar issues with the S35 on my 5Dmkiii...although not nearly as pronounced.
With the lensalign/focustune, I was getting +2 at 25x focal length, and 0 at 50x focal lenth.

Doesn't sound like much, but it's noticeable.