Canon 1D MKIII Autofocus- Interactive Review

Canon 1D MKIII Autofocus Problems- Interactive Review

Click here to read the entire article

I just ran into a buddy of mine who's been using the new MKIII for while now. He's done a few weddings, and some newspaper shifts. He says that he figures that he's getting 10% less photos in focus, but the extra image quality more than makes up for the difference. He couldn't say enough about it.

David Buzzard's Technical Blog

David,

That's exactly right. And is also why I haven't yet made any edits to my original recommendation, yet. Because I am still working on finding out the limits of the problem.

For a lot of people this issue is almost a total "non-issue." But, I can see how for others it would be huge.

Believe me, I didn't really want to have to go back and basically say, "Hey we missed something."

But, I have to report what I'm seeing. Just like I did the first time when shooting at a larger aperture, I did not see this issue. But Rob's report says nothing about wide apertures. In fact, he recommends to test at high shutter speeds and larger aperture numbers.

I originally set up this test to just look at the heat and light factor. I must admit, I probably had a good bit of experimenter bias. I expected it to be an open and shut case in favor of the autofocus system. I put in the f/5.6 tests at the last minute just to "cover my bases" so to speak. I have done a huge number of test trials at f/2.8 with different lenses, different lighting, different settings of c. fn 3, different runners, different clothing and none of them, I mean none (except for one outlier) produces acceptable results across the range of the run. Whereas, everything at f/5.6 is great regardless of variable. I would do a formal statistical analysis but the results are so glaringly obvious, p is definitley less than .05 if not .01 in this case.

Do tests that exercise the predictive AI servo focus in the Mark III have anything to do with the camera's one shot focusing? It'd be downright silly to set a Mark III to AI servo for most wedding shots. It's not as if anyone is running down the aisle (though a bride or groom running off at the last minute might make a good news shot).

Is 10% fewer in-focus shots with a Mark III still more in-focus shots with a 5D or 20D or...? That we don't know.

Drew, my time is getting tight as I'm prepping for a road trip. I may not be able to get to do an analysis of your further test images very soon.

-Noel

Both very good questions, Noel. No problem about the graphs. Have a good trip.

Yes, Noel, enjoy your trip. We potential Mk III owners appreciate all you and Drew do. (Drew do, Drew do, uh, there must be a better way to say it.)

Being from the Nikon camp, I can't comment directly, but I can say that the original D1x's had almost the same back focus problems, and it was fixed with a firmware update.

If you're a wedding photographer, it's going to be less of an issue, but still problematic, and agravating to have a great photo slightly out of focus. If you make your living shooting sports in arenas with a 400 2.8, then it's a serious issue.

You can also bet that some engineers at Canon are working overtime to come up with a solution, so it's probably just a matter of time.

David Buzzard's Technical Blog

Day 5 added.