1D III focusing issue

[ed_baumeister]

Photo District News, a U.S. publication, has information on the 1D III focusing

problem at http://www.pdnpulse.com/2007/07/canon-mark-iii-.html

[yakim_peled1]

http://www.robgalbraith.com/bins/content_page.asp?cid=7-8740-9006#july20

 

 

:-(

 

Happy shooting,

Yakim.

[ali_al_mohannadi]

Except for one comparatively minor problem, no

 

That it's mean that the Mark III still haveing the same problem

[colinsouthern]

I'm wondering if it's simply a case of lenses not being able to focus fast enough, and people expecting too much?

 

I was stressing out over this to the point of phoning up to cancel my order - got talked in to keeping the camera - and just love it to bits.

 

I'd be very interested to see what tests might reveal if it done at speeds slower than 10 FPS.

 

Cheers,

 

Colin

[rcox2]

I've decided to keep my order and take delivery next week. From what I've read, the issue is related to certain shooting conditions that may not apply to me very often. Besides, the autofocus on my current camera body has been broken for months and I somehow manage to focus manually! My main interest is the low noise at high ISO.

[steve_dunn2]

<cite>I'm wondering if it's simply a case of lenses not being able to focus fast enough, and people expecting too much?</cite>

 

<p>In some cases that could be possible, but it's not the case at least for Galbraith. He cites two of the lenses he's used for this testing: the 300/2.8L IS USM and the 70-200/2.8L IS USM. Both are quick to focus, and if the 300 isn't the quickest-focusing lens in the Canon lineup, it's gotta be pretty darn close. These two lenses have worked perfectly well on every other EOS body so far, including the previous 1D-family bodies and the 1V and 3 film bodies with their highly-regarded AF systems.</p>

[beauh44]

It's interesting that, in the link provided by Ed, they say: "In our own testing of the camera, we found no autofocus issues similar to Galbraith's results. You can read our full review in the upcoming August issue of PDN."

 

I must confess that I have not downloaded RG's test shots and scrutinized them at 100 percent in PS CS3. However when this first came to light some of the shots he presented did not appear "out of focus" to me.

 

It's true that some shots I saw may not have been *optimally* focused, however. (Perhaps a little back/front focus... which this camera can supposedly correct for; I wonder if RG has tried that?)

 

I guess we just don't know all the facts quite yet - but this has to be hurting sales of this camera so it would seem to me that it would behoove Canon to either put out a statement that there is in fact some problem that's not addressed by the latest firmware upgrade (i.e RG is right) or demonstrate that he's wrong.

[colinsouthern]

Steve Dunn wrote:

 

"In some cases that could be possible, but it's not the case at least for Galbraith. He cites two of the lenses he's used for this testing: the 300/2.8L IS USM and the 70-200/2.8L IS USM. Both are quick to focus, and if the 300 isn't the quickest-focusing lens in the Canon lineup, it's gotta be pretty darn close. These two lenses have worked perfectly well on every other EOS body so far, including the previous 1D-family bodies and the 1V and 3 film bodies with their highly-regarded AF systems."

 

I hear what you're saying Steve, but at the end of the day 10 frames-per-second is only 100 milliseconds - in that time it has to:

 

(a) Wait for the previous exposure to finish

(b) Compute where the new position needs to be

© Command the lens to move there

(d) Wait for it to move there

 

And if it moves to ever so slightly the wrong position then a soft focus is the result. Keep in mind that - compared to cameras shooting at 8 frames per second - the exposure time still comes out of the total time available - the computation phase may be faster or slower (don't know), and commanding the lens to move will take the same time (it's a serial protocol - not the fasted thing in the world).

 

Taking all of this into account the poor lens probably has something like 20 or 30% LESS time to move on a 1D3 (compared to the 8 FPS camera - I've got no idea what FPS the film cameras shoot at, but shurely it can't be 10?).

 

That's why I'd dearly love to see tests repeated at slower rates - the camera is programmable for this, and may just give punters an acceptable work-around.

 

Cheers,

 

Colin

[mark u]

A 1VHS film body shoots 10 fps. Perhaps someone may care to compare AF performance against the 1D MK III?

[beauh44]

Good point, Mark! (Although isn't a 1VHS 9 FPS? Splitting hairs :) I sold my 1VHS and don't own a 1D Mark III but if there were no focus problems with the 1VHS, that 9-10 FPS speed theoretically shouldn't be a problem with the 1D III, either.

 

The 1V is getting a bit long in the tooth by comparison. I would think Canon's AF sensors and algorithms would be much improved now over a camera that was designed the better part of a decade ago.

 

But again, this underscores that the potential target market for this camera is probably very confused about whether it's a good idea to buy now or wait. It would be nice if Canon could definitively figure out if there is a problem and if there is one fix it.

 

If they can't reproduce the problem, say so and tell everyone what's up with RG's results, if possible. Once this is behind them I'm sure the camera will sell like crazy.

 

Does anyone know if someone *other* than Rob Galbraith has experienced the same problem in a reproduceable manner?

[colinsouthern]

It's a pity that those doing the testing can't test the camera more objectively - I don't mean that to be offensive, but it seems to me like there are just too many variables being introduced at once (different camera - different FPS - different subjects - different interpretations of the results etc).

 

I had a close look through the images on the RG site - and frankly, many of the ones that are flagged as being OOF I couldn't see a major issue with - perhaps a touch more sharpening at best might have been required.

 

How all of this translates into "real world" photography is another question - if joe bloggs is shooting from 75 years away with an F5.6 lens then I suspect that DOF is going to totally mask the problem (in which case it's not a problem in that circumstance).

 

I'm sure that Rob Galbraith sets the highest of standards - and I'm sure that if Canon can make him happy then they'll make everyone happy - but having said that, I'm wondering just how big a problem it is for the vast majority of users? Only time will tell I guess. If Canon do find something that makes it a lot better then I'd probably get mine "fixed" too (for the sake of completeness) - but in the meatime I'm watching closely for developments on one hand, whilst remaining happy with mine on the other.

 

Cheers,

 

Colin

[beauh44]

<I>frankly, many of the ones that are flagged as being OOF I couldn't see a major issue with</i><P>

 

I felt the exact same way.

[ali_al_mohannadi]

http://www.the-digital-picture.com/Reviews/Canon-EOS-1D-Mark-III-Digital-SLR-Camera-Review.aspx

 

A New review and same problem