Thanks for the charts. It's very easy to see the results this way.

Happy shooting,
Yakim.

Hmmm.. Interesting! Except for the 525xxx - 529xxx range; there are more cameras with no AF problems than those with. But it was "reported" that those with 525xxx & above are supposedly those without AF problems but the poll indicates its those 515xxx - 519xxx that are performing as it should!

Cheers!

I bought mine end of July, serienumber: 525xxx
Today it's the first warm day with blue sky, so I went out and for AI-servo testing.
I went to a viaduct and tracked cars with a speed of 80km (it's a controlled traverse)
In every serie of lets say 15 pictures I had several OOF pictures
I found out I have one shot AF problems as well.


Edited by Chris Dees on Aug 12, 2007 at 05:47 PM GMT (Reason: Poor English corrected )

I'm confused by that Chris. Are you saying you do have problems with that unit or you do not? And did you vote?

Nill
~~
www.toulme.net

Nill Toulme wrote:
I'm confused by that Chris. Are you saying you do have problems with that unit or you do not? And did you vote?


I corrected the text, I do have problems, in AI-servo and one-shot.

I voted wrong, but you can't correct it.

It's not my day.

Jeff, thanks for this poll and Nill thanks for the graphics.

I think we don't have enough data to make firm conclusions, but it seems that the problem isn't solved after a certain serial number.

I was planning to buy the mkIII in September or so, but I wait a bit more.

What is with Canon lately? Obviously there is SOMETHING wrong here. At least to a degree. There wouldn't be this much discussion (not to mention the firmware update that acknowledges the issue) about AF problems on the MarkIII if they didn't exist.

It seems that almost every product the company releases lately has some sort of problem. Every day I just wish that they would admit to, and fix the AF problems of the 50mm f/1.2L. Oh how I would love to have that lens (in proper working order, worth it's price tag).

James Cripps wrote:
What is with Canon lately? Obviously there is SOMETHING wrong here. At least to a degree. There wouldn't be this much discussion (not to mention the firmware update that acknowledges the issue) about AF problems on the MarkIII if they didn't exist.
...

I agree completely about the smoke-there's-fire aspect of the discussion of this issue, but as to the firmware, in fairness, I believe there have been AF firmware updates of one sort or another for every 1-series. I know there were at least three for the 1DsMkII because I just bought one and updated it.

They're complex devices with complex programming. We're lucky they *can* tweak them in firmware.

None of which is to belittle the problems with the Mark III, which more and more seem perhaps to be of a different order.

And all of which is OT for this thread, which Jeff has asked us to limit to reporting of actual experience with problems on this particular camera.

Nill
~~
www.toulme.net

Maybe the Canon's "prototype-less core technology" is to be blamed for all these AF problems

http://www.canon.com/technology/canon_tech/explanation/cae.html

I just had a router bite the dust so my site will be down for a few days. That means I won't be able to post/host the update charts again till it's fixed, but here's the current tally in text form:

<#505xxx: YES, it has AF issues 8 10%
#505xxx - 509xxx: YES, it has AF issues 3 4%
#510xxx - 514xxx: YES, it has AF issues 7 8%
#515xxx - 519xxx: YES, it has AF issues 3 4%
#520xxx - 524xxx: YES, it has AF issues 3 4%
#525xxx - 529xxx: YES, it has AF issues * 6 7%
>#530xxx: YES, it has AF issues 1 1%
Subtotal YES 31 37%

<#505xxx: NO, it works as advertised 8 10%
#505xxx - 509xxx: NO, it works as adv 5 6%
#510xxx - 514xxx: NO, it works as adv 6 7%
#515xxx - 519xxx: NO, it works as adv 13 16%
#520xxx - 524xxx: NO, it works as adv 11 13%
#525xxx - 529xxx: NO, it works as adv * 5 6%
>#530xxx: NO, it works as advertised 4 5%
Subtotal NO 52 63%

* (These two numbers adjusted to correct one miscast vote)

TOTAL SAMPLE 83 100%

Nill
~~
www.toulme.net

Canon 10D wrote:
Maybe the Canon's "prototype-less core technology" is to be blamed for all these AF problems

http://www.canon.com/technology/canon_tech/explanation/cae.html


There does seem to be something different about Canon's recent testing methods (apologies if this has already been covered--my brain is in meltdown mode due to reading off too many MKIII threads. ) Quote from Paul Pope, a UK photog, on this with a link to his posting about recent tests he did.

"Canon also needs to look at the (new) way its been testing its cameras. If their testing failed to pick up a serious flaw like this then they need to consider their testing methods a failure. The old way would have picked this up months before production began."

http://forums.dpreview.com/forums/read.asp?forum=1032&message=24329243

Paul B wrote:
Canon 10D wrote:
Maybe the Canon's "prototype-less core technology" is to be blamed for all these AF problems

http://www.canon.com/technology/canon_tech/explanation/cae.html


There does seem to be something different about Canon's recent testing methods (apologies if this has already been covered--my brain is in meltdown mode due to reading off too many MKIII threads. ) Quote from Paul Pope, a UK photog, on this with a link to his posting about recent tests he did.

"Canon also needs to look at the (new) way its been testing its cameras. If their testing failed to pick up a serious flaw like this then they need to consider their testing methods a failure. The old way would have picked this up months before production began."

http://forums.dpreview.com/forums/read.asp?forum=1032&message=24329243

Very minor correction, Paul Pope hails from down under.

Regards,

Joe Kurkjian

Can we add another option "I plan on owning a MkIII if the issues are resolved, show me the results"

stuuke wrote:
Can we add another option "I plan on owning a MkIII if the issues are resolved, show me the results"
X2

I am in the same posititon.. As soon as the AF issue if solved I will order mine.
But not untill I am positive I won't get a lemon.

Have you considered discussing it with your vendor? They should take it back if it was defective. I took the risk and have been very happy that I did. My camera has a little over 2000 images taken with it, and it's been stellar. Yesterday I tested fast motion (water slide) AF Servo in bright sun, temperature was 101, and my bursts were all very sharp. I was very happy with the results. Of course they were not shot with an ultra-tele at F2.8... (F4 on a 24-105L at 100mm)...

I posted this in another thread.. but here is a crop from a series just converted straight from Lightroom. No post processing. Thought it might be helpful. Just my son staying cool down here in Texas.

Here is the original image not cropped. Just wanted to illustrate how much was going on in the frame.. I had the AF set to Servo, Center point, and surrounding assist points, Speed set to 1 step below default (slower). I also had it at ISO 200 with Highlight Priority ON.

pcho wrote:
Mine is as advertised. I have done running shots and numerous birds in flight shots using 45 focus points as well as single centre focus points with great success.Works fine with one shot mode as well. Only thing is, it is winter here and I have not been able to test it in hot days yet. I will have to wait for spring to arrive

One thing too, I have not fiddled with any custom settings nor micro focus adjustments on any of my lenses. I have tested with 500F4 and 300f2.8 lenses. So that also eliminates issues raised about lenses with f2.8 and F4

Not really... it just means yours is working properly for your subject matter. I'm still seeing more inconsistent AI Servo AF behavior with my Mark III and my 400 2.8 IS than with my 600 f/4 IS. But that's my findings. Can't claim it's a general problem. As I stated before in the previous poll thread, it's possible it's also a problem caused by my lens because in the past I have had certain lenses that didn't work as well on certain bodies as they did on other bodies. What I see with the 400 is generally OK focus accuracy but a fair number of shots that are back focused and a few that are front focused, and this is based on evaluation of recent indoor/evening/night sports events. I haven't seen the same level of AF inconsistency when using the 600 at the same kinds of events (mostly football).

I voted yes, there is an AF problem and the camera's serial # is 505xxx. I've only had the firmware updated by Canon to 1.0.9 and by me to 1.1.0, no other servicing yet as I'm still waiting for delivery of a second Mark III. Once I receive it, I'll compare it to the first one and send in the first one to Canon for servicing.

One Shot AF seems to be very good and other than the somewhat flaky AI Servo behavior, am very happy with other aspects of the camera. Even with the AF problems, I feel I'm getting at least as high a percentage of keepers as I did with the Mark IIN cameras, especially at lower light levels where the Mark III's AF seems to be more responsive.

Wednesday tally with 100 units voting:

<#505xxx: YES, it has AF issues 9 9%
#505xxx - 509xxx: YES, it has AF issues 5 5%
#510xxx - 514xxx: YES, it has AF issues 8 8%
#515xxx - 519xxx: YES, it has AF issues 4 4%
#520xxx - 524xxx: YES, it has AF issues 4 4%
#525xxx - 529xxx: YES, it has AF issues * 6 6%
>#530xxx: YES, it has AF issues 1 1%
Subtotal YES 37 37%

<#505xxx: NO, it works as advertised 8 8%
#505xxx - 509xxx: NO, it works as advertised 6 6%
#510xxx - 514xxx: NO, it works as advertised 12 12%
#515xxx - 519xxx: NO, it works as advertised 15 15%
#520xxx - 524xxx: NO, it works as advertised 11 11%
#525xxx - 529xxx: NO, it works as advertised * 7 7%
>#530xxx: NO, it works as advertised 4 4%
Subtotal NO 63 63%
* (These two numbers adjusted to correct one miscast vote)
TOTAL SAMPLE 100 100%

Nill
~~
www.toulme.net

Nill, we got accustomed to the chart….

Happy shooting,
Yakim.

Jeff wrote:
Ron, how is your AI Servo tracking sensitivity set? I'd be curious to see if changing it to one click on the 'fast' side would change anything. I was getting a fair number of backfocused shots with the occasional front-focused shot as well, and moving one click to the fast side seems to have cleaned up that issue a bit.

Jeff,

It's set between slow and normal. I've played around with that setting somewhat and found for the field sports I photograph regularly, setting it to the faster settings has caused problems when a player is momentarily blocked by something closer to the camera. Even at the semi-slow setting, I've found the AF tends to jump rather quickly. If I remember, I'll give your suggestion a shot at my next football game. Over at the naturescapes forum, in the huge Mark III thread, one of the guys (LesZ) recommended BIF be done at the fast setting, which makes sense since usually a BIF is the closest object to the camera. I think I will probably use the fast setting for specific situations where I know the subject will usually be the closest object to the camera.

My CF III settings are: 2-semi-slow, 6-5, 7-2 (I've made adjustments for the 400 2.8, which seemed to have helped somewhat), 8-2, 10-1 and 16: set H to 9 fps, L to 8 fps. I usually shoot at 8 fps and it "seems" to improve results somewhat, though I have not made a controlled direct comparison against 10 fps. I've used 4-1 somewhat but am not convinced it made a difference. 5-1 I've found to be too dangerous for my liking, having been burned using it a couple times.

Way back in early June I commented on the AF problems I had photographing a triathlon with the Mark III. Some of it was due to poor AF point placement, but where I saw the greatest problem was with relatively slow runners in a backlit situation. AF was all over the place! At that time I had CF III-8 at 0. Setting it to 2 (I haven't tried 1) has since seemed to help, as has the aspect of the firmware update that keeps the AF from jumping to higher contrast information in the background.

I'd like to point out that I also had issues with AF accuracy in sports situations with the 1D Mark II and IIN. Backlit was usually a problem. In any case, it always produced more oof frames in a sequence than shooting with full sun on the subject (I like the look of backlit a lot, so endured the poorer results). This was what I immediately noticed with the Mark III at the triathlon in June, but that the initial results were considerably less consistent and focus was much farther off when it was off. Night sports also caused problems for the Mark II series, where often the fps rate would really slow down while the AF tried to track the subject, often with mixed results. Here I find my Mark III does considerably better and even though there are some oof frames, overall it produces a higher percentage of usable images. I feel the poor AF performance I've sometimes experienced with the Mark II series, and now the Mark III, have to some degree been caused by low subject contrast combined with less than optimal AF point placement. CF III 8-2 seems to have helped my situation and so has using off center AF points, which I previously rarely used for action because I found the non-cross type to struggle under fast paced situations. Yet even though these changes to how I set up and use the camera have improved the Mark III's results, it's now flawless, especially with the 400 2.8 for some reason.

A point I'd like to make is the need for each photographer to test the camera with a variety of lenses under a range of situations before coming to a conclusion. If I only ever used the Mark III with my 600 and not the 400, I would feel more confident in its AF performance. Even more so if I rarely shot sports or moving subjects in challenging situations.

I find it curious that Canon Japan, has not had any Mark 3's for sale since June 14th. Does that mean that the limited availability we have seen in other countries indicates that the export pipline was already flowing and is just slow to stop? We always hear that the Japanese consumer is most unforgiving. From reading these forums I get the impression that many westerners are willing to take a chance on the Mark 3 and expect Canon to solve the problem. Canon's silence on the reported AF problem coupled with the unavailability of the camera in Japan, make one wonder if any fix, short of hardware modifictation is possible.
Cheers,- John

Mariner wrote:
I find it curious that Canon Japan, has not had any Mark 3's for sale since June 14th. Does that mean that the limited availability we have seen in other countries indicates that the export pipline was already flowing and is just slow to stop? We always hear that the Japanese consumer is most unforgiving. From reading these forums I get the impression that many westerners are willing to take a chance on the Mark 3 and expect Canon to solve the problem. Canon's silence on the reported AF problem coupled with the unavailability of the camera in Japan, make one wonder if any fix, short of hardware modifictation is possible.
Cheers,- John

Just out of curiousity, how do you know there have been no MKIII sales in Japan since mid-June? It would be kind of interesting. Canon has stumbled very badly in camera sales in Japan. Their sales there for the first half of this year are 10% lower than a comparable period from 2006 (and I believe they've ceded the #1 slot in Japan to Nikon.) Which is not the case in the rest of the world where camera sales increased quite a bit.

My simple example test.
Shot a flower at noon in high contrast while set up on a tripod. Ap5.6, Tv 640, ISO 250, single shot with AF.
I used a 100mm canon2.8 macro.
I only changed Ap to 2.8 and reshot. The second shot was horribly OOF.
I have used my other body, 1Ds MKII, as a comparison using the same settings and didn't suffer any probs.
Same result with other set ups using 70-200 L and DsII but found the AF on the DIII was very intermittent.
.

515xxx

I have AF problems with One Shot, AI Servo. Seems to backfocus badly.

Ryan

Paul B wrote:
Mariner wrote:
I find it curious that Canon Japan, has not had any Mark 3's for sale since June 14th. Does that mean that the limited availability we have seen in other countries indicates that the export pipline was already flowing and is just slow to stop? We always hear that the Japanese consumer is most unforgiving. From reading these forums I get the impression that many westerners are willing to take a chance on the Mark 3 and expect Canon to solve the problem. Canon's silence on the reported AF problem coupled with the unavailability of the camera in Japan, make one wonder if any fix, short of hardware modifictation is possible.
Cheers,- John

Just out of curiousity, how do you know there have been no MKIII sales in Japan since mid-June? It would be kind of interesting. Canon has stumbled very badly in camera sales in Japan. Their sales there for the first half of this year are 10% lower than a comparable period from 2006 (and I believe they've ceded the #1 slot in Japan to Nikon.) Which is not the case in the rest of the world where camera sales increased quite a bit.

Paul,
In all honesty, this information comes from several different sources all posting on "Digital Photography Review". Perhaps Canon is facing issues in there home market, I can't say but it would be interesting to hear from FM'rs who are based in Japan.
Cheers-John