Mark III AI-Servo test on static subject

[Leon Wittwer]

--
Leon
http://homepage.mac.com/leonwittwer/landscapes.htm

 

[Christopher Glaeser]

The speculation is, and this is purely speculation, that the sensor,
which is designed to vibrate during cleaning, may possibly vibrate
during capture.

And why would it do this?

Mirror slap?

Best,
Christopher

 

[photopro300]

Some months back I called Canon for this exact question. Why does the Mark III have such issues with the focus being so jittery when attempting to focus on a static object in AI servo. The Canon tech explained it was completely normal and to be expected, as the Camera was expecting to see motion. He said that the Mark III and the Mark II are not to be compared as they are so different. Different software platform, different sensitivity.

If Canon (IF) designed the camera to be this way, than they can program it to not be this way. So many comments about how after the fix the problem is gone. I really wonder what they are doing that they aren't telling us. Mine should be back in a week and I sent it in with complaints in both one shot and AI servo. I will find out soon...

In the mean time, I think this was a valid test, a real complaint and certainly a test in which the Mark III should have passed but has been "improved" to the point of being as I have read a camera that acts as if it has A.D.D.

 

[Andy Wai]

Blue dot, at tail end of affected range but on exception list. 1.1.3 firmware.

But things are more complicated than that. Lately, I've been playing around with the AF micro adjustment target mentioned here:

http://www.openphotographyforums.com/forums/showthread.php?t=4708

Download the GIF file and display it at 100% on screen. Then AF, stop, and get into LiveView to check. Works great. I've had tons of trouble adjusting the 85/1.2 before and this target helps me lock it in just fine.

Anyway, after adjusting the 135/2, I remembered this thread and did an AI burst test using the target. The focus locked on solidly and stayed in place for all 9 frames. At f/2, every single frame was nailed . Then I did another burst with an Adobe Bridge screen examining the first burst. This time, focus was jumping a little and about half the frames are slightly out.

If you look at the target, it's made up of black and white dots. No intermediate tones at all. AI servo locks solid there. The copy on the Bridge screen is much less sharp. Looks to me like AI Servo isn't 100% sure so it decides to wonder a little just to cover the basis.

This is only one test at 10'. Not sure how significant it is in real use. Will likely do another test at closer distance when I have time. Closer is probably where it will affect real life use the most. And if I have more time, will probably try another lens and perhaps even firmware 1.1.0.

We'll see...

Andy

 

[Leon Wittwer]

I did several ten second AI servo sequences at 10 fps on my blue dot MKIII from a tripod and from hand held. I also did a bunch of single shots from a tripod and hand held. I must have looked at almost a thousand shots. All were in good focus.

It is unfortunate, but the Canon "technicians" that we typically have access to are not necessarily all that knowledgeable. They "know" what Canon has told them or trained them to know but that seems to be not all that accurate.
In the worst case, some seem to tell us what makes us happy and go away.
--
Leon
http://homepage.mac.com/leonwittwer/landscapes.htm

 

[pikeface]

doing the test with the button half down there is a small ammount of hunting, however if I look at the focus scale while shooting 10fps it does not move and pics look in focus.
--
Have a look and leave a message
http://www.pbase.com/jeremy_broome_smith

 

[J A K]

early production pre AF fix, early production with AF fix, or production blue dot? Thanks in advance for your reply.

Regads,

Joe Kurkjian, Pbase Supporter



SEARCHING FOR A BETTER SELF PORTRAIT

 

[pikeface]

SN 539xxx
blue dot
came with 113

I assume it was checked/fixed by canon

--
Have a look and leave a message
http://www.pbase.com/jeremy_broome_smith

 

[Hawkman]

Or any Canon body for that matter - focusing with a shallow DOF on a curved surface like that from my experience.

Gene

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.

--
Gene (aka hawkman) - Walk softly and carry a big lens

Please visit my wildlife galleries at:
http://www.pbase.com/gaocus
http://hawkman.smugmug.com/gallery/1414279

 

[J A K]

"NT" means no text.

Joe Kurkjian, Pbase Supporter



SEARCHING FOR A BETTER SELF PORTRAIT

 

[Scott Larson]

Or any Canon body for that matter - focusing with a shallow DOF on a
curved surface like that from my experience.

I assume when you posted this, you hadn't seen my nearly successful test with my Mark IIn on the same target. It gave me just one OOF frame and the rest from that point were perfect.



I'm really encouraged by the posters who are saying they're getting in-focus shots from the Mark III's with this test. If the submirror fix solves my problem, I'll have to thank every one of you and I'll have to think twice about any advice I get from Canon from now on.

 

[Sennaista]

Interestingly enough my 350D just did a 10 frame burst with the focus being spot on in all shots.

I suppose its true when they say money can't buy you happiness.

 

[Hawkman]

No I hadn't seen it.

I had based this on my experience shooting a small bottle with more curvature (and actually with the 135/2 @ f2 as well) with my mark IIN and noticed some more in focus than others. But I was testing something else at the time (banding), so I could easily have been user error or the greater curvature of the bottle covered by the AF sensor or both.

Some days I feel bad that I cancelled my 1d Mark III order and got a 1D Mark IIN instead, and somedays I feel good. Knowing my luck...

In any case I hope you get it resolved.

• Gene

--
Gene (aka hawkman) - Walk softly and carry a big lens

Please visit my wildlife galleries at:
http://www.pbase.com/gaocus
http://hawkman.smugmug.com/gallery/1414279

 

[Scott Larson]

Some days I feel bad that I cancelled my 1d Mark III order and got a
1D Mark IIN instead, and somedays I feel good. Knowing my luck...

I bought this one the weekend before Canon told the stores to pull them off the shelves and send them back. So there's MY luck for ya!

 

[rrcphoto]

Some months back I called Canon for this exact question. Why does the
Mark III have such issues with the focus being so jittery when
attempting to focus on a static object in AI servo. The Canon tech
explained it was completely normal and to be expected, as the Camera
was expecting to see motion. He said that the Mark III and the Mark
II are not to be compared as they are so different. Different
software platform, different sensitivity.

canon actually confirmed to you that it was a different software platform?

I started to suspect that canon moved from vxworks to dryos in it's DSLR's this year .. you gave some credance to that suspicion.

if that's the case, no wonder for the longest time they thought it was a firmware adjustment / issue. that would have been basically a re-write of the firmware to talk to a new base OS. having tackled problems like that before I'm surprised they found the problem that quickly to be a manufacturing defect. especially considering the rather inconspicious nature of the actual "fix".

 

[joseph papa]

yeah, Canon told me it was a different software platform in the Mark III and that it was significantly more complex. I bet we see at least 3 more firmware versions in 08.

 

[Bartek]

my 1D3 is blue dot 541xxx (within the general range however from the exempt list of unaffected bodies) with 1.1.3, I was using 70-200 2.8 non IS at 200 from some 1.5m distance to the inkjet printed text target at 45deg. to better see potential shifts.

Target at the room temp (22C deg.) was lit with 2 300W modeling lights from studio strobes at full power and it gave 1/640 at f2.8 at ISO 400.

Camera in AI servo (center point no assists) and high speed drive mounted on the studio column stand (very stable over 80kg) and cable release used. No mirror lockup as normally there is also no mirror lockup during typical AI servo operations.

The achieved focus during first frame didn’t move even by 1mm on the target text during the whole batch of 12 shots repeated twice. I was surprised as I was expecting some shifts in at least #2 and #6 frames however nothing like that happened. When the sequence is being watched on quick scroll of frames on my monitor (similar to motion picture) the point of focus simply doesn't move at all throughout the entire batch.

Then I repeated with 100 f2.8 macro and the same (even better sharpness as the macro is sharper at f2.8 by itself).

--
Bartek

 

[Honor]

isn't just the low pass filter on the sensor vibrating not the sensor itself at least that is what i think i read. I would not think that is possible it would be too gross a design overlooking. I believe Canon engineers are much above that.

 

[Honor]

I wish you had put all the required info in your first message too

So you had your affected 1dmk3 replaced by another 1dmk3 also without blue dot and when ?

 

[Honor]

I read every single message in this thread. I am pretty happy with the number of people with blue dots cameras who have dead on focus in ai servo. We already know that but feels good to hear it from actual owners. Also realized that Canon technicians words are to take with a grain of salt (lots of it) when an issue like this surges it is understandable that they will be a lot of misinformation among them first because they are not engineers and can only repeat what they have been told to the extend that they understand it and second their priority tends to be to drown the fish to appease customers.

• There seems to be a misconception by many people that ai-servo can not work in static shooting. IT CAN and WILL if the camera is a good one.

• It is logical to assume the servo motor will stop if both the camera and subject are static. Nervousness in the early 1dmk3 seems to be the result of the faulty mirror assembly and the software. The seems to be a pb of the past now.

-Canon service seems quite good.