Please do. I'd very much like to see Pekka get his hands on Rob's cameras, and vice versa, and then to sit in on a discussion between them. ;-)

Nill
~~
www.toulme.net

Nill Toulme wrote:
Please do. I'd very much like to see Pekka get his hands on Rob's cameras, and vice versa, and then to sit in on a discussion between them. ;-)

Nill
~~
www.toulme.net

Me too!

For the record, while there are a few throwing barbs at one another there, but for the most part, the discussion is rational and is similar to this one. Several camera owners, some with problems, some without. And a lot of us bystanders waiting for some resolution before making any kind of buying decision.

I'm watching both forums (forii?).

Well there are at least a few barbs flying on every forum addressing this issue because the level of frustration is high. It's extremely tempting for those who are not experiencing problems to believe that those who are are doing something wrong, and equally tempting for those with problems to think that those who say their cameras are fine are shooting under different conditions or worse yet don't know a sharp shot when they see one. And when either of those sentiments gets even slightly hinted at, people understandably get their backs up.

Reminds me of the way I more and more frequently find myself feeling on the interstate — everybody going slower than me is a moron, and everybody going faster is a maniac. ;-)

Personally I'm still holding out the hope that there really is some significant sample variation at work here — truly some good units and some bad units. That's the only thing I can think of that would readily explain the very different results that people are seeing, and if it's the case, it means it can indeed be fixed, and probably very easily.

Nill
~~
www.toulme.net

I think some of the difference may be attributable to experience and expectations. For example, as part of my testing I shot a worker in a local park on a tractor moving at a decent clip on a curved path. I focused on his chest as an easy target and got good results except when I slipped off slightly and then the camera focused on the more brightly colored and contrasty fire extinghiser that was a couple of feet behind him on the rollbar and partially visible in the open area between is his chest and is arm.

My expectation is that if I move off the target that the camera is going to find something else to focus on. To me the camera is working as I would expect but my experience in shooting moving objects is somewhat limited so perhaps I am wrong in accepting it moving off the acquired target so quickly.

Hrow,

That is actually a function of tracking sensitivity. If it is set to slow, the AF will not jump to another point as quickly. If set to fast, I will move very quickly. Canon describes the AF sensitivity in terms of something moving into the active focus point, but the reverse is also true. If you move the AF focus point off the intended target the result is the same. Here is a simple test of what I am saying.

1) Use a fast focusing lens; 300f2.8, 400f2.8, 70-200f2.8, etc. Some primes do not focus as fast as these. At any rate use your fastest focusing lens and IS off.

2) Set the camera for center point focus only, no expansion and AI Servo focus mode.

3) Set up a scene with a fairly large object, a person, a sign, something easy to hold the center AF point about 30 to 50 feet in front of you, and a house, treeline, or bigger sign about 30 to 50 feet behind the first object.

4) Set the AF tracking sensetivity to Slow.

5) Focus on the closer target using the center AF point only. The focus should be steady with no hunting.

6) Move the camera quickly to the the distant target without releasing the half shutter or AF Start button. Note the time delay before the camera refocuses on the distant subject.

7) Repeat with the AF tracking set to Fast, and note the difference.

I have a MK IIN, but my MK III has not yet arrived so I only have a feel for the II. I am curious to see the difference.


Richard K.

Thank you Richard, I do need to play more with the settings and that is a good way of seeing the differences. Currently I am set to Fast as this was the recommended setting by a very talented wildlife shooter who was having good success with the M3 while others were having focusing problems. It seems to work as long as my skill level is up to par, which it isn't most of the time. Lots of practicing to do and hopefully I'll get a chance this weekend with some polo ponies.

Nill Toulme wrote:
Well there are at least a few barbs flying on every forum addressing this issue because the level of frustration is high. It's extremely tempting for those who are not experiencing problems to believe that those who are are doing something wrong, and equally tempting for those with problems to think that those who say their cameras are fine are shooting under different conditions or worse yet don't know a sharp shot when they see one. And when either of those sentiments gets even slightly hinted at, people understandably get their backs up.

Nill
~~
www.toulme.net

You've hit the nail on the head here Nill. I would like to believe that I know a sharp photo and an accurately focusing camera when I see one. My MKIII is a stud! But, I also have great respect for some of the photogs who are experiencing great frustrations with their MKIII's. My theory is that the bad cams came from a single source and that they will be eventually culled out.

My big fear is that Canon will have a blanket firmware fix and my MKIII will start screwing up after installing it.....

Bob

Hrow wrote:
I think some of the difference may be attributable to experience and expectations. ...


OK, so where am I? I can use some advice. As I alluded to earlier in this thread, I have amassed a respectable collection of boring pictures of my wife trying not to run me down with the family car, and indeed I can see many of the poor&jumpy-focus issues that are similar to RG's reviews.

Three possibilities I consider:

(1) Coming from a 10D to the MkIII (504xxx), I am clueless how to set the camera up to make it work to its potential.

(2) The camera itself is not operating within Canon's expected envelope.

(3) What I am seeing is as good as it gets, and my expectations are too high.

How realistic is the car-coming-at-you scenario anyway? Well, it is as close as I can currently manufacture to shooting soccer next fall, where I often want action that's coming directly toward me at wide open apertures. So my test are with a 135F2 at F2.0....

Now it's possible the problem is (1) or (2) or (3), or even worse, maybe some combination of (1) AND (2) AND (3).

What can I do to figure it out?

So far, the best answers I come up with are (a) wait it out and see what everyone figures out, and (b) go out and shoot something that really matters under realistic conditions.

If anyone here has a better plan, what are you doing?

Thanks,
Bob Walker

Nill Toulme wrote:
Reminds me of the way I more and more frequently find myself feeling on the interstate — everybody going slower than me is a moron, and everybody going faster is a maniac. ;-)



You and I must drive the same speed.

I'm starting to wonder if the problem is in the lense AF circuit chip, and cannot be fixed in camera firmware, or even AF sensor redesign/replacement....

Wouldn't that be a bummer.....

Really, you'd think Canon would be all over this and already have issued the firmware fix, if it were possible.

Someone posted here about using a hot mirror filter fixes the problem. I know that RG said that UV filters don't help but a hot mirror makes a considerable difference to the wavelength of the light hitting the AF sensor and kills all the IR present in the warmer light which hot and sunny weather brings. I know that Pentax people found that this cured a lot of their focus problems, it might be worth seeing if more people could report back on whether it helps.

What is a hot mirror filter?

Nill
~~
www.toulme.net

Nill,
I believe it is a filter that will remove IR light, it is also called IR cut fillter. I am using those filters for lenses on my Leica M8 to get rid of the ugly purple shift.

BTW, did you guys notice on the poll that the ratio of having AF problems to not having AF problems has shifted? For the first time, the number of people that did not experience any problems is higher now. Since you can vote only once, I started to believe that the early copies had quite a few of bad apples or that the newer bodies were further tweaked by Canon. I voted "not sure" before but now I am ready to jump to the "not noticing any AF problem" camp but I can't....

Does it reduce light transmission, like a polarizer, or not, like a UV?

Nill
~~
www.toulme.net

Nill Toulme wrote:
What is a hot mirror filter?

Nill
~~
www.toulme.net

Nill,
Just did a search on B&H. Tiffen makes a Hot Mirror filter. 77mm is $179. God does it ever end. If this is the fix and you have a few different sized lenses the price seems to stay at around $179.
Rob Eckert

Nill Toulme wrote:
What is a hot mirror filter?

Nill
~~
www.toulme.net

its a sharp cut IR blocking dichroic filter.
http://optical-components.globalspec.com/LearnMore/Optics_Optical_Components/Optical_Components/Hot_Cold_Mirrors

Edited by sjms on Jul 04, 2007 at 08:53 AM GMT

Hmmm... wonder what a 52mm drop-in for the BWL's runs? :-(

Nill
~~
www.toulme.net

a tiffen screw in 52mm is about $58

Isn't this funny. A few years ago people complained that the Hot Filter glass in front of the sensor was creating sharpness issues. Companies responded by reducing the strength or eliminating the hot filter altogether. Color shift issues have become a problem. Highlight ghosting on some cameras. Now a possible focus issue with the MkIII. That is the reason I just hold on to my MkII. Great camera and I know its quirks. I can work around them. I hate how many companies today practically use the public as "Beta Testers". Camera companies are not the only ones doing this.
Rob Eckert

sjms wrote:
a tiffen screw in 52mm is about $58

I have had problems fitting some screw on filters into the slot of my 300 IS when mounted onto Canon's holder. I used to use only "thin filters". Problem is many filters are not available in the thin version. If I recall correctly Tiffen is already on the thinner side. My drop in filter holder originally came with this thick Canon filter that did not even fit into my 300 IS slot. Go figure.
Rob Eckert

Basically the poll is a statistical 50/50 split. Does not seem to be a difinitive result yet.
Rob

Nill Toulme wrote:
Hmmm... wonder what a 52mm drop-in for the BWL's runs? :-(

Nill
~~
www.toulme.net

Nill,

I got the Canon holder and B+W IR filter for my 400/2.8 from B&H.

See this thread http://www.fredmiranda.com/forum/topic/549801 with baseball pics shot with the filter and MKIII.

Here is the link to the IR filter:

http://www.bhphotovideo.com/c/product/154800-REG/B_W_65014688_52_mm_486_Digital.html

And you will also need the screw-in type filter holder if you don't have it already. The 400/2.8 comes with the drop-in type.

http://www.bhphotovideo.com/c/product/186092-REG/Canon_2612A002_Drop_In_Filter_Holder_for.html

-John

Very nice John, and thanks for the links. I'd really like to get your impressions with vs. without the IR filter, when you get a chance.

Nill
~~
www.toulme.net

This was posted to a different site, but I thought folks here might find my experience useful. This is the first time I have posted here--it's the first time I have had anything worthwhile to add to this excellent forum. Are there bad units in the wild? Yes. My 1D Mark III just shipped out to Canon CPS per their request following my description of my Mark III's AF problems.
I normally assume I am at fault when my results are sub par. That assumption is what made me keep this body beyond the time limit for returning to the dealer. First, I noticed many unsharp results amidst some stunningly detailed frames. Although not limited to telephoto shots, the soft results dominate my telephoto frames (70-200 F4 L IS & 300 F4 L IS).
Trying to sort this out, I was unable to get any sharp frames of oncoming traffic--but speeding trucks really slow down fast when you point a camera at them. Needing a more repeatable test, I have done a ridiculous # of test series of my wife riding toward me at 10-15 mph on her bike (in hot and cool, sunny and cloudy weather). The results? Nine gigabytes of distressingly out of focus boring pictures.
Had I suddenly lost all my photographic ability? Did I not know the magic CF combination?
OK, I needed a more controllable & objective test. Out came the metal yardstick and focussing chart. Crude, but very enlightening.
With the MKIII on a huge old Gitzo 'pod and Arca Swiss monoball head & focused on the 8 in the #18 on the angled yardstick at 45 degrees, I ripped off a burst at 10 FPS. Viewed on the monitor I watched focus shift, frame to frame, from back focus to front focus in the burst series. The series was shot in AI Servo, Center point with no focusing point expansion on. Focus shifted from almost #16 to #19(camera approx. 10-15 feet from ruler). These results were repeatable with two different 70-200 L lenses(yes, IS was off) and the 300 lens. The same lenses on my wife's 20D in the same test yielded (repeatedly) virtually no variation at all.
Next a focussing chart was used, again tilted 45 degrees to the camera's focal plane. The results with each of the three lenses showed focus shifting from front focussing to back focussing by as much as 2-3 inches. This occurred in single shot, non-Servo mode as well as 10 FPS bursts in AI Servo mode. In some series, focus would shift from pronounced front focus to pronounced back focus in just two frames. The focus shifting was very repeatable except the degree and direction varied from series to series. The new focus microadjust feature can't cope with variable focus shifts. Neither can I. Next week I should know if Canon CPS can resolve this problem.
For me, this proves that some MK IIIs are indeed "bad". I don't question for a moment that many others have perfectly wonderful MK IIIs. But for those of you who are unable to come even close to results with other bodies, you too may have a MK III that keeps guessing where to focus.

John, that sounds like a very good testing method.
I have to try that one!

Thank you for sharing.