PDA

View Full Version : Who has tested their 1DIII in hot Servo mode?


Asher Kelman
September 11th, 2007, 02:20 AM
Now we should be getting many more reports.

What your the experience in using the brand new 1DIII in sports, bird photography or anything that moves and needs tracking?

Asher

Paul Bestwick
September 11th, 2007, 04:04 AM
Asher that is the only time I had an issue. The subject was walking...... not even a fast action motion.
Out of 70 shots there were very few in focus. A1 servo.

For stationary subjects..... no prob.

I remain a strong Canon supporter though I am a little disturbed with the current situation. I notice that the shop where I picked mine up is unable to get stock at the moment. Is Canon holding back ?
Will there be a global recall ?

Time will tell & I am confident of a positive solution.

John_Nevill
September 11th, 2007, 07:34 AM
I was shooting Duxford Airshow at the weekend with a colleague, who used a 1DIII and EF300 f4. It was bright, sunny, quite humid and very hazy for most of the day.

I lost count how many times he cursed the dSLR throughout the day, needless to say he had more than a fair number of AF lock ups, the AF would just freeze.

I believe he's now going to drop it into Canon UK, as he wasn't a happy bunny.

Ironically, we were shooting at WHF the day before in less humid and sunny conditions and it was fine.

I'll hopefully have a chat with him later in the week to see what the effects were on output.

Asher Kelman
September 11th, 2007, 10:30 AM
Asher that is the only time I had an issue. The subject was walking...... not even a fast action motion.
Out of 70 shots there were very few in focus. A1 servo.

For stationary subjects..... no prob.

I remain a strong Canon supporter though I am a little disturbed with the current situation. I notice that the shop where I picked mine up is unable to get stock at the moment. Is Canon holding back ?
Will there be a global recall ?

Time will tell & I am confident of a positive solution.
There are guys who have OOF shots every time in AI Servo (in hot weather I call it "Hot Servo", for fun) and others that are spot on. It seems that at least an early batch was bad. Someone pointed out that Canon replaced all the early 24-105 f4 L IS as they has flare problems. If they find the issue, I guess there will be a recall. At the moment, if you fuss, they'll swop you a new one.

Now, who knows they may just hope the new one is O.K. or else they have quietly made changes.

What's the serial number of your camera?

Asher

Paul Bestwick
September 11th, 2007, 06:36 PM
well I don't seem to be able to get a new one. It has been suggested to me to try the 300 2.8 on my 1DS2 to ascertain if it is a lens issue. I will try it.
This image was shot as the subject was walking toward the camera. F5.6 @ 1/400 sec

http://www.studio58.com.au/OPF/example.jpg

Steve Saunders
September 12th, 2007, 11:53 AM
That whole shot looks soft, maybe camera shake? It's hard to tell as there asren't any nearby front or rear items to look at and see if they are sharper than the guy in the picture.

Asher Kelman
September 12th, 2007, 02:49 PM
Paul can you show the whole frame?

Asher

Paul Bestwick
September 12th, 2007, 05:08 PM
camera shake @ 1/400 sec with IS..... ?? I thought the focus point indicated what is in focus..... nothing is in focus in this shot.

http://www.studio58.com.au/OPF/ghoul%20a.jpg

Ivan Garcia
September 12th, 2007, 07:37 PM
I think I might have a similar problem not sure...
This is the first image in which I am sure the focus failed to lock in the right place.
I have taken a few oof images, but I have always though it was camera shake.
As you can see, everything is in focus, except the chosen point.


http://www.escueladebaile.co.uk/ivan/OOF_1.jpg
1DIII
100-400L @ 400mm,
Av F5.6 Tv 1/640
ISO 800

What do you guys make of this?

Bart_van_der_Wolf
September 13th, 2007, 01:59 AM
As you can see, everything is in focus, except the chosen point.

I'm not sure how much the actual focus sensitive area extends beyond the lit-up indicator. The red box, on it's lower right, is bordering on a small area of high contrast background, which may have taken precedence when the bird shook its feathers. It may be worthwhile to test how far the sensitive area really reaches beyond the indicator (I'd presume almost halfway towards the bordering focus sensel).

Bart

Ray West
September 13th, 2007, 03:24 AM
A few thoughts, which may be of use.

For Ivan's image, I think Bart could well be correct. It looks as if the leaves above the bird, are more in focus, and I think they may be behind the bird. afaik, the af works on a high contrast line on the object. The high contrast line is where? the edge of the stick? so does it focus on the stick end, or the bird? AF can be confused quite easily. So, perhaps the camera focusses on the stick edge, then possibly moves some distance, detecting bird feather movement and keeps going since it gets the higher contrast area, (mentioned by Bart) could be one rough explanation. If Ivan has only a few oof issues, there may be a consistency within them, and the cause may be determined by careful testing. It may be that in certain conditions, the behaviour is slightly different than may be expected.

For Paul's image, and I guess many are similarly blurred, the hair looks as if it may be more in focus, but that maybe the nature of hair. I don't know the depth of field for that lens/distance, but I guess if focussed properly, everything would be sharp (except the background). Is the lip area high enough contrast, is curved lines a problem? It may be it is sort of retaining the focus from where it was yesterday, so to speak, it's caught out, looking for sharper movement/contrast and then Paul presses the button. What to do? say 'I'm not ready, no image yet', or 'you know best, is this good enough? I think the camera and lens may need calibrating together, since the problem seems to be consistent. The way this stuff works, the speed at which it works, is quite amazing. To get the speed, there has to be tolerances in the components. Sometimes these tolerances add up, sometimes they subtract. Trying the lens on another camera proves the lens on that camera, it does not prove this lens/camera combination. (If the lens used to work fine on the other camera, but now it doesn't, that does not prove that the lens is faulty, either.) It may be worth trying with is on and then off - more user testing, I guess. I can see people walking around, carrying trays of spice jars ;-)

Best wishes,

Ray

Steve Saunders
September 13th, 2007, 10:37 AM
Ivan's bird shot is probably more useful. I wonder why the AF point didn't lock onto the clearly defined line between the end of the branch and the birds feathers. To my eye the end of the branch really should have been sharp.

Will_Perlis
September 13th, 2007, 11:49 AM
how far the sensitive area really reaches beyond the indicator

Bart,

IMX with the 1V and the 1DMk2 the AF starts to lock on to a small high-contrast area against a blank background about half a box width beyond the indicator box. Someone will need to check out the Mk3 to see if it's similar.

Ivan Garcia
September 13th, 2007, 07:05 PM
Thank you all for your answers.
I am going to meet up with Ray tomorrow to assess how this camera focuses and whether or not I have a focusing issue.
Will keep you posted on our findings.

Jack Joseph Jr
September 14th, 2007, 08:08 AM
how far the sensitive area really reaches beyond the indicator

Bart,

IMX with the 1V and the 1DMk2 the AF starts to lock on to a small high-contrast area against a blank background about half a box width beyond the indicator box. Someone will need to check out the Mk3 to see if it's similar.

Will, I think that you are correct in that the relationship between a focus box in the viewfinder and the part of the scene that the focus sensor sees is not precise.

Last year at a track meet I was playing around with focus between events. I focused my 1D2N on a large "E" at the end of the word "COLLEGE" printed evenly spaced between the top and bottom of a hurdle. If, on servo, I moved the bottom of the focus rectangle up to the middle horizontal of the "E" the focus switched to the track behind the hurdle.

If I moved the top of the focus rectangle down to the middle horizontal and even a bit further down the camera's focus remained on the hurdle. The conclusion is that the center focus sensor on my camera sees a little bit higher than the red box in the viewfinder indicates. Now when I want to focus on a face at a distance I aim for the chin.

Will_Perlis
September 14th, 2007, 08:46 AM
Jack,

Yes, there's both the "bigger" factor to consider AND a possible "off-center" factor that might complicate focussing on a small but critical part of a scene. My 1v is off to the right a bit, the 1dMk2 is centered.

It's easy enough to test for, all one needs is a blank wall with one small and sharply defined contrasting spot. as a target. It's harder to deal with during fast shooting if the target is small.

Ivan Garcia
September 14th, 2007, 07:14 PM
After a productive day with Ray, all the problems I encountered with my MKIII have been resolved successfully.
Ray, thank you for your hospitality, your kind generous input, and patience.
It sure makes a huge difference.

Asher Kelman
September 14th, 2007, 07:39 PM
Hi Ivan,

So impressed that you two U.K. OPFers got together and with such good results. We want a full report on Ray, LOL! Is he short, tall? Does he have a micrometer for one hand and a multipurpse tool for the other? To say the focus problem is fixed, is just like hearing the British foreign Office saying the "Empire is still strong!" What did Ray do? We want the secrets!!

Asher

Ivan Garcia
September 14th, 2007, 08:30 PM
Hi Asher
Ray is tall, handsome, and extremely generous with his knowledge.
I did not notice any bionic body parts ;-).
As for the problems we resolved, they were mostly my lack of understanding of how modern autofocus systems work. Ray was very patient with me in his teachings, and kindly made his explanations very simple and easy to assimilate.
Also, my MKIII and 100-400mm combination, amplified the built in error tolerance margins on both pieces of equipment, resulting in serious front focusing issues.
The 100-400L worked fine on his camera, discarding the lens as the source of the problem. We then proceeded to attach his 70-200mm on the MKIII and that lens performed faultlessly. So the issue was restricted to this particular lens- body combination. After some micro focus adjustments (Thank you canon for including this very useful software with the new camera), and applying the new focusing technique suggested by Ray, the problem was resolved.
Again my most sincere thank you to Ray for his generous help.

Asher Kelman
September 14th, 2007, 09:00 PM
Ivan,

Glad to hear how well Ray looked after you! He's generous in meeting you and getting so involved in your camera's maladies!

How fiddly was the focus adjustment? Why could you not also do it for your 100-400? Also, what are the exact steps in Ray's technic that gets you into 1DIII Servo AI focus Nirvana now? It may have been given above, but I don't see it!

Congrats Ray on being such a sterling help. This is above and beyond the call of duty and very instructive. Now perhaps you can fly in to help Rob Galbraith, LOL!

I wonder how many cases can be fixed by shooting technic and lens adjustment?

Asher

KrisCarnmarker
September 15th, 2007, 12:49 AM
I'd be very interested in what Ray had to show regarding focusing technique. Any chance you could describe it here Ray?

As for the focus micro-adjustments, I have spent considerable time with that. It is a real PITA, but what a wonderful feature to provide. I used the methods and chart described here (http://www.canon-dslr.com/Canon_Jan05/Canon_SLR_Focus_Test.htm), as I belive the other's found around the web, especially the too simplistic one found on photo.net, simply is not good enough. The problem with the micro-adjustments is that the focusing error is not consistent. I usually take 10 images at any given FL, and try to come up with an average. But it is still quite hard, as sometimes I would get five front-focused and 5 back-focused. All in all, I've probably spend a good 10 hours adjusting for the two lenses 70-200 f2.8 and 100-400, and while it is better now than in the beginning, I am still not happy with it. Out of ten shots, I easily get five OOF, either front or back focused.

Any ideas to help stabilize this would be very welcome :)

Ivan Garcia
September 15th, 2007, 09:14 AM
Hi Kris.
These are zoom lenses and therefore difficult to adjust.
The way I did it (although I did not use a chart) was to position the lens at the halve way focal length (200mm in my case) and at a distance of 5-6 meters; this worked fine for me and focus is within tolerance at all focal positions now.
Ray observed my focusing technique and came to the conclusion I was not using the focusing point correctly.
I was not very clued up when it came to the limitations of focusing sensors. So, after some simple explanations of the inner workings of the system (I won´t bore you with it, I am sure you will like a more technical disclosure and I do not have the required knowledge to write about it), He suggested focusing in a more defined contrast area of the image. It works for me, but your millage may vary.

Alan T. Price
September 27th, 2007, 07:29 PM
Paul, maybe your camera was trying to protect you from a horrible sight :)

More seriously, I'm certain that any 1D-series body should have latched onto that subject without any trouble at all. The dark mouth against the pale face offers all the contrast needed for the AF to operate.

- Alan

Paul Bestwick
September 28th, 2007, 05:40 AM
Alan to be honest, I would be prepared to admit it may have even been user caused (if possible)

I find the 300 2.8 difficult to hold steady to critically focus so maybe that had an influence ?

Bev Sampson
September 28th, 2007, 08:28 AM
Here is a link to the most recent comments from Rob Galbraith. You will recall that he was probably the first to call attention to the AF problems in AI Servo in hot, bright conditions.

http://www.robgalbraith.com/bins/multi_page.asp?cid=7-8740-9068-9093