Latest report on the 1D MkIII AF saga

[Malcolm_Stewart]

This article has appeared overnight.
Rob Galbraith's latest report on the revised 1D MkIII

I've not used my body and lens on football, but I took my 1D MkIII (latest firmware, both of Canon's fixes) and EF300 f2.8L IS to an athletics meeting a few weeks ago with good results. I'll have to read Galbraith's article carefully. I do know from other tests that subject variation has a huge effect on AF tracking, but have to admit that the real world for sports shooting is probably closer to Galbraith's tests than those I've done.

[Malcolm_Stewart]

Oh dear!

I'd planned to get to the first of this weekend's two "Race for Life" events in Milton Keynes yesterday morning. Having seen the gist of Rob Galbraith's latest update, I left my EF 200 f2.8L at home* and took a monopod and my EF 300 f2.8L IS instead, and by chance had my 1D MkIII CFns set up with, to quote from R G's article:

" C. Fn III-8 [AF Expansion with Selected Point] is enabled and there's more than one subject in the frame, but it's not something we've emphasized. That was probably a mistake, because this setting is a problem. Switching on the camera's Assist AF points can lead to more out of focus pictures in some situations, including those that this option is meant to help."

As I said to a friend, if I was a working PJ, I would have been sacked after yesterday's results.

By far my worst results yet, with many really out of focus shots, and lots of "iffy" shots. (I used Av and f4 so that depth of field was short enough to show up problems.) Fortunately the weather seems to have held, and I'm off to cover this morning's event with the AF assist points turned off and with most CFns set to their default settings.


*BUT, The EF200 f2.8L and 1D MkIII, same settings as yesterday, did really well at the Tour Series cycle race in Milton Keynes 10 days ago. The conditions then were cooler, and with evening sun, not as bright. Trials with my EF300 f2.8L IS at an athletics meeting a few weeks ago were fine. Weird...