Mark III AI-Servo test on static subject

[Scott Larson]

These pictures aren't exactly high contrast. The background appears
to be grainy brown, so I'm wondering if you're using a high ISO and
the weakest possible color. Do you get the same results with black
and white? Different lighting?

My subjects in real life aren't exactly high contrast either. Almost of them have grainy brown skin. The test was to show that my Mark III is not focusing as well as my Mark IIn in a real life situation.

I'm sure it would do better if I gave it more light and a target that a $100 P&S camera could focus on but unfortunately that's not what I shoot.

 

[Scott Larson]

That's good...now I won't feel bad sending my camera in for the
repair. I had not seen any focus issues, but I hadn't done much to
stress the focus in servo mode. Perhaps this is the easiest test to
determine if the focus problem exists.

I think we're both seeing the exact same problem with our unfixed Mark III's. Let's hope Canon schedules our cameras to get fixed sometime soon.

 

[blackhawk13]

See "I did that test" message above.
--
Leon
http://homepage.mac.com/leonwittwer/landscapes.htm

--What works real world is all that matters. Instead of trying to find faults with this camera, why not harp on it's strengths?

Dollar for dollar the 5D has been more of a disappointment for me. If the Mk-3 cost $800 these same people would be drooling all over it like a rabid toddler.

These threads were old a half year ago, and even then my MK-3 had a much higher keeper rate then my 5D, or 20D for that matter.

At this point with the new firmware and mirror fix in... your probably just beating the dead horse into a liquid.
My MK-3 is not for sale...

-nothing beats a fast lense, except a fast girl-

 

[Tomas Hellstrm]

Hi folks,

I thought I let you know. My fixed Mark III does not shift when targeting static objects. I did the same test like scott and the focus locked and stayed locked. This is also my impression in real life shooting where I more and more use AI-Servo when shooting static objects as well.

--
Tomas

 

[Curator]

The submirror issue is a mechanical problem. The issue described in
this thread is not a mechanical one related to the mirror. If so, it
would manifest only when the mirror is being activated.

I'm not sure how you came to this conclusion based on the images posted in this thread, but I'm also confused as to what you are trying to say with that last sentence.

I believe it to be a calibration issue

Possibly, and from your previous post regarding calibration, it prompted me to try this static test. I tried various micro-focus adjustment settings, and none of the settings made the jittery behavior better or worse.

or possibly an issue with loose/dirty contacts.

I have doubts about this. I would think we would see a communication error, or Err 99, rather than strange servo mode behavior.

 

[Victor Engel]

The submirror issue is a mechanical problem. The issue described in
this thread is not a mechanical one related to the mirror. If so, it
would manifest only when the mirror is being activated.

I'm not sure how you came to this conclusion based on the images
posted in this thread, but I'm also confused as to what you are
trying to say with that last sentence.

Not just from the images. The OP also stated that when he holds down the focus button you can observe the lens shifting back and forth. Obviously, in that situation, there is nothing mechanical happening with the camera -- only with the lens. That is how I eliminated the submirror being the problem.

By the last sentence, what I mean is this. The problem with focus caused by the submirror involves the mirror moving and then coming to rest in a different position after a shot than it was in before the shot, as I understand it. A shutter actuation is required to move the mirror. In the case of the sample photos, obviously that is happening. However, in the case of the lens moving around without the shutter actuation that is not happening. Hence the source of the problem is not the submirror (although that may be an additional problem).

or possibly an issue with loose/dirty contacts.

I have doubts about this. I would think we would see a communication
error, or Err 99, rather than strange servo mode behavior.

All I know is that I've read posts stating that this sort of problem went away after contacts were cleaned. Hearsay, I know, but that's all I have.

--
http://www.pbase.com/victorengel/

 

[Bartek]

Not just from the images. The OP also stated that when he holds down
the focus button you can observe the lens shifting back and forth.
Obviously, in that situation, there is nothing mechanical happening
with the camera -- only with the lens. That is how I eliminated the
submirror being the problem.

... with an exception of the case when the sub mirror stopper that has improper physical parameters got "used" to early as it was to soft and now after some time of use (some time after the camera was initially calibrated in factory with new stopper in place) the position of the sub mirror is no more the same as it used to be at the time it was calibrated (slight difference might be enough to foul the AF system).

It's just my speculation but wondering what might be wrong with the rubber part I come to conclusion that one of features of faulty rubber element might be the slight but permanent change in it's shape, softness or it gets used way to fast etc.

--
Bartek

 

[Jay Bean]

My Mark II and my original 1D do the same thing. AI is for moving objects not static objects.

If you put your blue dot/submirror-replaced Mark III on a tripod, set
the AF to AI-Servo and continuously focus on a static object (a paint
can for example), does your focus shift around like mine does?



Is this even a valid test? I'm trying to qualify my AI-Servo problems
some more before I send it to Canon again.

 

[Scott Larson]

My Mark II and my original 1D do the same thing. AI is for moving
objects not static objects.

You might consider reading the entire thread before posting.

 

[Curator]

My Mark II and my original 1D do the same thing. AI is for moving
objects not static objects.

"Even if AI Servo AF is used to focus on a still subject, stable control is enabled so that the lens drive isn’t constantly fine-tuning. If the subject starts to move, the focusing responds immediately to detect the subject constantly".

 

[Victor Engel]

I can't comment, because I can't parse out what you said into anything meaningful.
--
http://www.pbase.com/victorengel/

 

[ihbond]

i spoke with a canon tech rep and they said anything after 546561 had the blue dot. But you never know
--
If Stupid had a major you would probably have a P.H.D.

 

[blackhawk13]

My Mark II and my original 1D do the same thing. AI is for moving
objects not static objects.

"Even if AI Servo AF is used to focus on a still subject, stable
control is enabled so that the lens drive isn’t constantly
fine-tuning. If the subject starts to move, the focusing responds
immediately to detect the subject constantly".

--That's a nice quotation, but not if it's wrong, or missing an important detial. Even the white papers have errors... what else is new?

-nothing beats a fast lense, except a fast girl-

 

[Victor Engel]

You're right. AI Servo is designed for moving targets. However, let's apply the algorithm it uses to a static subject.

First, assess the original scene, find a target, and determine how far out of focus it is. Command the lens to move to the position that would bring the subject in focus. Now assess the scene again and see how far the subject is out of focus. Compare the position the lens was supposed to go the first time with where it needs to go the second time. That gives an estimation of the direction and speed of the subject. Extrapolate from that to get the future position of the lens.

Now for a static subject, the first command should move the lens to the correct focus point. The second assessment of the subject should then result in no adjustment by the lens. Extrapolating no movement out to the future results in no movement.

But, no matter. You can do the same experiment with single shot focus. Just repeatedly press the focus button and see what happens. If the situation is like it was when I had my 10D 50mm combo that exhibited the problem, the lens refocuses slightly each time the focus button is pressed. After the cleaning/calibration, this no longer occurred.

Speaking of cleaning, I would try the experiment with different focus points to see if the effect is the same. I'd also check if the focus sensors (or the light path to the focus sensors) needed cleaning. I think dust in the light path to the focus sensors can result in this sort of unpredictable behavior.
--
http://www.pbase.com/victorengel/

 

[Bartek]

I meant that however the submirror is resting during the focusing it still does affect the process itself and if the resting position is different from the original one (due to faulty stopper) the one shot focusing is that way affected by mechanical failure.

--
Bartek

 

[Derek Hawkins]

Well calibrated samples have no AF issues with AI Servo whether the subject is moving or stationary. In fact this is one way to separate the "sheep" from the "goats".

Picture this...You're in AI Servo tracking a hummingbird or dragonfly and the critter suddenly stops in midair as they often do....Click! Well, you get the picture. No doubt there are many other scenarios that would render AI Servo somewhat worthless if it was unfaithful by design in handling static subjects. Use AI Focus instead you say....LOL!

 

[Scott Larson]

--That's a nice quotation, but not if it's wrong, or missing an
important detial. Even the white papers have errors... what else
is new?

Hmm, a white paper from manufacturer that reflects most user experiences of the product, or a quote from disbelieving user on the Internet with no evidence to the contrary.

Which to believe?

 

[Victor Engel]

I guess I'm still missing the point.
--
http://www.pbase.com/victorengel/

 

[Jay Bean]

I dont' believe that to be the case, both from personal experience with multiple generations of bodies, & from some of the features offerred on the cameras and lenses. Cases in point:

C.Fn-04 (looking at 1D Mark II) mode 3 let's you set the " " button for "subjects which keep moving and stopping" If AI Servo AF locked when stopped, then there'd be no need for this mode.

C.Fn-19 mode 4 let's you set the lens AF stop button (for lenes that have it, usually the long sports / events lenses) let's you swith AF mode while the button is depressed. This mode is "convenient when you need to keep switching between One-Shot AF and AI Servo AF for a subject which keeps moving and stopping"

Finally, appling Occams Razor, let me pose the following question -- If AI Servo AF will lock onto a static subject just as good as One Shot mode does, then why is there a need for One Shot Mode? Why not just have a single AF mode and be done with it?

Q.E.D.

Well calibrated samples have no AF issues with AI Servo whether the
subject is moving or stationary. In fact this is one way to separate
the "sheep" from the "goats".

Picture this...You're in AI Servo tracking a hummingbird or dragonfly
and the critter suddenly stops in midair as they often do....Click!
Well, you get the picture. No doubt there are many other scenarios
that would render AI Servo somewhat worthless if it was unfaithful by
design in handling static subjects. Use AI Focus instead you
say....LOL!