Yep, those definitely look like the Mark III AF issues. It's the inconsistency that drives people nuts.

Front focusing by looking at the water near the bottom of that second photo, while the Mark III "thinks" it has a good focus there. It's what I see too.

Good to see you again Jeff always loved you posts and data. I got my issues taken care of too, made the jump to Nikon about a year ago. It would be interesting to see if it would have fixed the issues with my bodies but I was scared they were going to call it fixed and move on to the Mk4 and I am left holding the bag. I still find it interesting people still blame RG or user error even after 3 recalls and firmware upgrades. Good luck with yours!

Tom

Garylv wrote:
Yep, those definitely look like the Mark III AF issues. It's the inconsistency that drives people nuts.

Front focusing by looking at the water near the bottom of that second photo, while the Mark III "thinks" it has a good focus there. It's what I see too.


In my experience my previous 1DIII was always front focusing when it was off. I'd call that consistent

I haven't tried AI Servo with outer AF points on the exchange body Canon finally handed me. I'm happy; my previous body was unpredictable in every situation: Servo, One Shot, good light, low light, center point, outer point - you name it. Did every shot at least three times just to be sure - a habit I'm still trying to get rid off. My new body is better in every way. But AI Servo / outer AF points? Too afraid to ruin my confidence yet again I guess... Haven't tried and not about to start now. Still, I'm finally happy with my 1DIII, for my purposes.

Jeff, good to see you back.

Hey thanks, guys, appreciate it!

Mine also backfocuses at times for no apparent reason, and even shooting center point-only it occasionally hiccups, but tolerably so, at least compared to the first body. I wish I had both the cajones and pesos to switch, but I just couldn't do it on several levels. I do still believe that Canon will make it perform to spec, so hopefully this last trip will be it.

I'm still waiting for Canon to make my D700-equivalent...

Best to all,

Jeff

Jeff wrote:
It's really quite difficult to know without getting a large number of cameras together, and having them tested using identical protocol.

You mean the protocol that RobG has still not developed, right?

Jeff, I wish you all the best. After everything you've gone through, you deserve it.

Happy shooting,
Yakim.

I would also like give a warm welcome back to Jeff and to extend some courtesy towards him as well.

To the guys who are either disputing his claims or thinking of doing so, specially the ones who have joined FM recently, please take into consideration that Jeff is a former Moderator, and he was a very good one at that.
He is a very talented photographer, and highly technicaly knowledgeable individual, at least as far has photography gear is concerned.
So I would advise you guys to think twice (or "thrwice") before disputing his claims. And while doing so (no one is always right after all), please do so in a courteous way, as he deserved all the respect possible from all of us, as an individual, as a photographer and specially for all the hard work, trouble and time spent moderating the forums in the past.

He did have serious issues with the MIII, and they are well documented in forum archives. Do some research.

Thanks Yakim and Luis, but I don't deserve any more respect than anyone else around here, and I'm as likely to be wrong as the next guy/gal. Given the (intentional, unfortunately) tone of one single sentence of my post, perhaps even less. But it does get me when someone jumps on, evangelically telling all how things are based on their own singular experience in the world. Very similar to lots of things that have gone on in the world in the last 9 years, and for the last 900. This 'argument' has become just like all the other age-old arguments that you'll never win by discourse: left vs. right, mac vs. PC, this religion vs. that religion, two wings vs. X number of rotor blades, Peets vs. Starbucks, etc., etc...

Appreciate the warm welcome 'back'...

Cheers and coffee to all for the AM (Peets comes highly recommended by a friend of mine in the know)!

-Jeff

Jeff, I'm assuming you weren't using AF expand on those? That behaviour would be typical of AF expand being enabled, but I generally don't see it with it disabled, unless the subject is very small and low contrast.

Jeff we will apreciate fi you could also give us the CustomFunction settings for the OOF picture that you demonstrate

Isn't the latest fix supposed to address inaccuracies in the outer focusing points?

garyvot wrote:
Isn't the latest fix supposed to address inaccuracies in the outer focusing points?

Yes it is and in my case it not only improved the outer points but the center point as well.

Jeff you would be remissed to not try the latest fix and firmware update on your camera. Having owned (4) MKIIIs, and shooting sports and wildlife with them, the latest fix really improved the performance and reliability of the AF. It definitely performs better than any of the (5) MKIINs I have owned.

What do you have to lose at this stage of the game.

Welcome back!

Jim

My experience with the camera suggests that there was a fairly fundamental issue with the calibration of the AF until the latest fix. I had perhaps 3-4 shots like the OOF swimmer shot above in every burst of 20. Like this shot they couldn't be explained away by too slow a shutter, lack of subject contrast, AF point drifting off the subject etc. Despite these frames I still got at least 15/20 or so usable frames with most subjects and was pretty happy. Now I still get some OOF frames but they happen when I expect them, such as when I have difficulty keeping the point on the subject, marginal shutter speeds or with very fast subjects/wide apertures where I know I'm pushing the system to it's limits.
Interestingly, inexpilcable totally OOF shots in one-shot mode no longer occur following the fix either (the 35L was the real offender here) and MA has a much more predictable and precise effect than it did before. Most lenses are now pretty much bang-on with no adjustment (35L needed -17 on my camera). Previously adjusting by +/-1 in MA, it was very difficult to judge what effect it was having. Since the fix I can plainly see the focus point moving forward/backward by a predictable increment with each adjustment. An element of randomness in the AF process (be it in the AF algorithms or mechanical calibration) seems to have been removed.
I think a little more explanation from Canon is warranted because the latest fix has had far wider ranging effects on my cameras AF than just improved outer point accuracy.
My advice would certainly be to send it in and get the fix.

willis wrote:
My advice would certainly be to send it in and get the fix.

Jeff, that would be my advice too; my new body is so different in so many ways it's like a totally new camera to me. Of course: it is a totally new camera but I'm convinced my first really was a lemon. In fact: my new one is the best camera I've ever had... I'm getting a 5DII for the MP but I'm convinced it won't match the 1DIII for all the rest.

Thanks for all that positive input, I absolutely will send it in soon. I've got a major crit to shoot, as well as the State swimming championships in the next month, so I can't afford to be without it until after those events.

For those of you that commented above and got a successful outcome, did Canon get very specific about what was done? Any hardware changed that you know of? I know Canon can be very inconsistent with the details provided, and what is on the repair sheet may or may not be what was actually done.

Mostly curious if they are still changing out hardware, especially if non-Error 99 related. However, my main shutter button has recently been flaky when on a tripod only, but no Error 99 messages. The vertical shutter button works when the main one appears not to, so I'm guessing the shutter button is going, though I have no idea why it only occurs on a tripod.

Anyway, thanks again for all the input, I do appreciate it very much. I'll certainly report back after it's been in for the recalibration, once I've had a chance to truly test it. Also, the IS is pretty much dead on my 70-200/2.8L, not looking forward to paying for that one...

Best,

Jeff

Jeff I think you will be pleasantly surprised once you get it back and mine was shipped out by UPS on a Thurs and I got it back the next Thurs.

Jeff - I sent mine in and they replaced the entire mirror box even though it was a "blue dot" model (bought Feb 2008).

My personal experience is more anecdotal... I've always sucked at servo/tracking. It started with the 20D and when the 1DIII gave similarly poor results trying to shoot my dog I just chalked it up to poor technique since the camera was much better in every other respect.

After the fix its like a new camera, I no longer totally suck a servo (although not by much).

No hardware fix for me just 'AF calibration'. My turnaround was inside a week (in the UK).

Jeff wrote:
Also, the IS is pretty much dead on my 70-200/2.8L, not looking forward to paying for that one...


Replacing IS on the very same lens cost me close to $500 IIRC...

my MK3 is from december 2007 (quite early model it seems). so far, i could not observe OOF in tracking mode, f.ex. BIF or action scenes.
is this possible?
what is the ultimate test then?
are there users with early models AND without these AF problems??

It could easily be possible, not explainable but possible. I would still recommend sending it in for all updates and fixes as it will be a resale question asked by any buyer. Plus you get a good looking over and cleaning too.

e.aland wrote:
my MK3 is from december 2007 (quite early model it seems). so far, i could not observe OOF in tracking mode, f.ex. BIF or action scenes.
is this possible?
what is the ultimate test then?
are there users with early models AND without these AF problems??

Yes, its possible.

There is no "ultimate" test. If there were Canon wouldn't have ended up in this position in the first place.

Yes, there are (or were) (some/many?) users with very early cameras who never had an AF issue, which further slowed Canon's initial investigation.

Based on everything I've read I would send yours in, just to be sure.

thanks for the replies. the resell-argument is logic. i knew what i was going to buy and i am still very pleased with the speed and the IQ of this camera. its a poem. anyway, the fact that i do not have these OOF phenomena yet, even shooting highspeed bursts on action, makes me hesitate to send it in too early. which guarantee i have what i'll get back? anyway, i called canon and they told me "not every camera makes this problem, most of them are ok, if there are problems, i should document them by photos and send the camera in".

e.aland, the issues are apparently more noticeable with longer telephoto lenses, certain aperture settings and certain subject movement. Much of RG's testing was done with a 300 f2.8, and with subjects coming towards the camera quickly.

http://www.robgalbraith.com/bins/multi_page.asp?cid=7-8740-9068-9357

e.aland wrote:
my MK3 is from december 2007 (quite early model it seems). so far, i could not observe OOF in tracking mode, f.ex. BIF or action scenes.
is this possible?
what is the ultimate test then?
are there users with early models AND without these AF problems??

Sure, it's possible, many people have no apparent problems, but I think that the number of hardware fixes that Canon has performed would leave me to believe it prudent to send the camera in for whatever repairs they deem necessary. I'd not bet a dime to a donut that we have been apprised of the full story in relation to how widespread hardware fixes are, and I know for a fact that repair summaries included with returned cameras do not necessarily reflect what has occurred when a camera is in for repair.

I actually think that the runner coming straight at you is indeed the ultimate test for the issues that the MkIII has. Unfortunately the test takes a bit of skill to be able to conduct in a manner that will give meaningful results (i.e. shooting sports isn't as easy as it might appear), and the subject needs to be wearing an appropriate shirt. A telephoto lens set at f/2.8 I'd say is almost required, set for center-point+expansion, or another off-center AF point. I can't use the 'ring of fire' for any sport or moving subject I've tried, it's simply very inconsistent, so I suppose that's another test you could try. I've also shot kids cycling and scootering and they work as well, but due to flailing body parts running is a pretty consistently inconsistent 'target' to reveal any AF/AI Servo problems that a camera has.

This testing also takes a significant amount of time at the computer using Canon's software (ImageBrowser) afterwards to truly determine the source of any AF errors, so prepare yourself to invest some time.

-Jeff