Ive had similar thing happen to me. I actually attributed it to user error. But shooting a 1DIII with center focus. I focused on the face and of course when the image came out. everything was just blurry. Like wtf blurry. Of course it was an indoor shot with not as much lighting, but my thinking was, why in blazes would the body autofocus confirm if it indeed could not autofocus?
Oh well. In other instances my 1DIII performs like a charm. I just hope it won't screw me when I need it most one of these days.
nle57 wrote:
Ive had similar thing happen to me. I actually attributed it to user error. But shooting a 1DIII with center focus. I focused on the face and of course when the image came out. everything was just blurry. Like wtf blurry. Of course it was an indoor shot with not as much lighting, but my thinking was, why in blazes would the body autofocus confirm if it indeed could not autofocus?
Oh well. In other instances my 1DIII performs like a charm. I just hope it won't screw me when I need it most one of these days.
Yeah... Sooner or later it just might. It has ruined a few good moments for me, but nothing that would cost me a job or anything. The upside is still there though and I am getting some results that are quite good...So it's a love/hate thing with me right now.
This thread is slipping into the same pattern as all the previous MIII AF threads.
Yes, the MIII misses even though it looks good in the viewfinder.
Yes, the selected AF point shown in ZoomBrowser does not correspond to the actual focus.
This is off the original thread topic of how AI Servo functions when the subject moves off the center AF point.
In the future, if you are taking a picture of a non-moving subject, don't use AI servo.
Apdieb - Any chance you had the shutter set to "silent single shooting?" I've had terrrible results with this mode with oof images as there seems to be a delay between depressing the shutter release and taking of the image which causes me to move the camera instinctively before the shot is actually taken. Sometimes I mistakenly set it in this mode and the delay completely throws me for a loop until I figure out what I'vd done.
For what it's worth Iv'e been struggling with the exact same problem so has my best friend.
We both have had our camera's back to Canon 5 times, just recently took my lenses and camera to Ivine to have the lens calibrated to the Camera, no change.
Funny thing is after I got the camera back from New Jersey and sent it to Irvine, Irvine said on the work order that AI servo was adjusted, wonder why Canon New Jersey didn't see that, doesn't matter it still doesn't focus.
I'm in the process of speaking with an attorny about Canon's attitude towards this issue on my camera, after 1 year of constant frustration and inconsistancies I give up, my attorny can handle it from here on out.
apdieb wrote
Here is an example. Only thing REMOTELY in focus was the grass in the bottom right corner.
Shot at ISO 100, F2.8, 1/160th, at 23mm
Boy does that example look familiar!
No surprise you were shooting at f/2.8, nor wide. I still say this new AF module has quirks when shooting at f/2.8 and below, and those quirks are somehow exacerbated when shooting wide.
Daan B wrote:
The point is, the mkIII AF requires a different approach under certain conditions. The sooner you are aware of that, the better results you are going to get
It'd be fine if that 'required' approach were based at all on some sane, rational approach to autofocus.
When the camera requires you to focus on a point 3 to 4" behind the eye in order to get the eye in focus, something is very, very wrong. A camera should not require that kind of adaptation (which will negatively transfer to other cameras, I might add) in order to get the result that you want.
Learning and adapting to minor idiosyncrasies of equipment is one thing, but Andrew should have been able to nail that 'simple' shot using a fairly normal approach.
nathanlake wrote:
This thread is slipping into the same pattern as all the previous MIII AF threads.
Hmm... Gee, I wonder why that would be?
nathanlake wrote:
This is off the original thread topic of how AI Servo functions when the subject moves off the center AF point.
Although it doesn't directly address the OP's question, it does address the fact that the AI Servo 'feature' in question may not work as advertised.
Years before I bought my 1Ds I'd envisioned the 45-point AF system practically being able to track a specific 'subject' in the frame, as it moves around within the AF array. It sure sounded cool! Even though generally described that way in Canon marketing materials, it clearly does not work in that manner. I believe there are video cameras that attempt to do that (and are relatively successful), but Canon's AI Servo AF is not (yet?) that sophisticated. Suffice it to say that if the user-selected AF point loses the original subject, the AF system may select an AF point on the highest-contrast part subject, not that part of the subject that was originally used to obtain focus.
For the OP, in practice it's not quite as magical as Canon makes it sound, however good it may be.
If your subject is clearly separated from the background, such as a bird against sky, the AF system can do a pretty good job of following it from the center point to any of the other points. In most other situations, though, it really only works if (1) the subject is high contrast and (2) the subject is large enough in the frame to start covering a neighboring sensor before it leaves the first one.
Jeff wrote:
It'd be fine if that 'required' approach were based at all on some sane, rational approach to autofocus.
When the camera requires you to focus on a point 3 to 4" behind the eye in order to get the eye in focus, something is very, very wrong. A camera should not require that kind of adaptation (which will negatively transfer to other cameras, I might add) in order to get the result that you want.
Learning and adapting to minor idiosyncrasies of equipment is one thing, but Andrew should have been able to nail that 'simple' shot using a fairly normal approach.
Not a fair response at all Jeff... I initially responded with the quoted statement assuming that his camera functions properly. However, after seeing his shot with the highlighted AF points I am not so sure it does. Because the AF point is placed for maximum contrast and not covering solely the entire face (as Andrew first stated) - which would have meant a low contrast situation. Unfortunately I have no clue as to what has happened here. It looks like it seems to be random as well.
The bottom line, Canon has challanges with their autofocus systems regardless of settings I use when using professional/non-professional cameras, regardles of lenses used. I just shoot a lot and hope that the law of averages rewards me with some great images in focus.
when I look at the grass, I'm not sure but what the 1/160th is not a bigger factor than the focus.......that said, I do miss more focus w/ my 16-35 lens than with other. I'm also shooting 1DmkIII and my body is spot on focus, so I'm not so sure everybody should assume it's the mkIII. So imho, it's more about the shutter speed. ....also, fight the urge to put the focus on the face when busy background, just go center point, not above the center point (especially w/ sports), and let'r rip.
elfanucchi wrote:
These MKIII focus errors are at least entertaing to read ...
It's why I decided to buy a classical 5d for vacation and wedding stuff,
but then I use Mamiya Phase One 645 MFDSLR for high $$ stuff.
Perhaps you should try manual focus ...
EL
Edited by elfanucchi on Aug 29, 2008 at 01:57 AM GMT
Well I decided to check back in on this thread.
Hey El... Do you actually brag about your MF camera on every forum you vist? Just curious. How many sporting events you shooting with that Mamiya these days?
As far as shutter speed being part of the problem. The day I cannot hold steady a 23mm shot at 1/160 when shooting 3 different frames separately and with flash is the day I need to just quit being a photographer. I have seen exactly the same results on some frames at 1/1000th+ as well. I have shot many dSLRs in my life (both Nikon and Canon...and Proline from both) and shoot an average of 2000 frames or so a week professionally. I know for a fact that this is not working right and I refuse to accept that it's user error. But I guess showing proof is still not enough for some people.
I'll just deal with Canon in my own way..(exactly what Canon is hoping for...stop inquiring and posting evidence and switch to apathy). I am going to really have to rethink my long-term future as a Canon user. Sad that it's coming to this.
I appreciate everyone's time here. And please forgive me for taking the MKIII focus discussion away from OP.
There's a nice little write up at dpreview by a self-confessed Canon fanboy that's been using Canon since the 70's and used a Nikon for the first time, specifically a D3 for indoor volleyball. D3 does extremely well straight out of the box.
Pixel Perfect wrote:
There's a nice little write up at dpreview by a self-confessed Canon fanboy that's been using Canon since the 70's and used a Nikon for the first time, specifically a D3 for indoor volleyball. D3 does extremely well straight out of the box.
