Why can't Canon make a 1D focus right?

George

I appreciate the central zone is lit up but did you have all 45 zones in operation or just the central zone, or another AF set ( 8 or 11 zones) selected?

Steve


As you can see focus was set on lizard body but 1Ds focus much
under focus point. I already hear you need change, calibrate lens,
calibrate camera, change setup etc. no the Canon 1Ds does not
always focus where you wont.

george
I use two 1D for supercross and motocross. I switched from Nikon
about 15 months ago.....

But, Canon can't seem to make these things focus correctly and with
any reliability.

I wrote this up on my own site:
http://www.stevebruhn.com/TheBiz/Camera4.htm

........ they simply can
not take a sharp photo using servo mode. All you get is focused
way off the bikes or backfocused bikes.
Your posts, and George’s post are worrying. I also came from Nikon
(F5s) to the Canon digital. I tried two D10s, one of which was
“repaired” (they were all back focusing, even with stationary
subjects) and none worked, so I sent them back and bought a 1Ds.
The image quality is incredible, but - for the al servo with
certain subjects, it misses virtually 100% of the time (the actual
plane of focus is behind the place focused upon - the face in this
case - while the F5 was always spot on. Specifically, if you are
curious, the issue for me is with the very fast moving dogs coming
toward me that I photograph for magazines and calendars. The
highlighted red “in focus” squares are right on the face, yet the
shoulders are where the camera actually focuses.

Now, what worries me is that Canon told me that while the AF sensor
is the same, the predictive focusing algorithms are different
between the 1D and my 1Ds, and the “old” 1D is better at
successfully catching what Canon terms “erratic” movement (their
term to describe the dog photography I do) than my 1Ds. They also
said the (old) 1D predictive focus was better than my F5, and the
new 1DM2 was yet better again. So, all I have to do to solve my
problem is buy the 8MP 1DM2 and I’m covered! The problems you both
have make me worry this isn’t true, and I’ll be back using two
systems and scanning again. By the way, I have not had these
problems with the 1Ds for human sports or race car photography. I
sure hope there will be feedback from photographers using the new
1DM2 for the kind of work giving you problems with the 1D. Do you
have one on order? I certainly hope you will share your findings,
or at least email me directly with the result.
 
Steve I used 11 points but this doesn’t matter camera suppose to focus on any number of points or do not turn on LED in focus indicator.

george
I appreciate the central zone is lit up but did you have all 45
zones in operation or just the central zone, or another AF set ( 8
or 11 zones) selected?

Steve


As you can see focus was set on lizard body but 1Ds focus much
under focus point. I already hear you need change, calibrate lens,
calibrate camera, change setup etc. no the Canon 1Ds does not
always focus where you wont.

george
I use two 1D for supercross and motocross. I switched from Nikon
about 15 months ago.....

But, Canon can't seem to make these things focus correctly and with
any reliability.

I wrote this up on my own site:

........ they simply can
not take a sharp photo using servo mode. All you get is focused
way off the bikes or backfocused bikes.
Your posts, and George’s post are worrying. I also came from Nikon
(F5s) to the Canon digital. I tried two D10s, one of which was
“repaired” (they were all back focusing, even with stationary
subjects) and none worked, so I sent them back and bought a 1Ds.
The image quality is incredible, but - for the al servo with
certain subjects, it misses virtually 100% of the time (the actual
plane of focus is behind the place focused upon - the face in this
case - while the F5 was always spot on. Specifically, if you are
curious, the issue for me is with the very fast moving dogs coming
toward me that I photograph for magazines and calendars. The
highlighted red “in focus” squares are right on the face, yet the
shoulders are where the camera actually focuses.

Now, what worries me is that Canon told me that while the AF sensor
is the same, the predictive focusing algorithms are different
between the 1D and my 1Ds, and the “old” 1D is better at
successfully catching what Canon terms “erratic” movement (their
term to describe the dog photography I do) than my 1Ds. They also
said the (old) 1D predictive focus was better than my F5, and the
new 1DM2 was yet better again. So, all I have to do to solve my
problem is buy the 8MP 1DM2 and I’m covered! The problems you both
have make me worry this isn’t true, and I’ll be back using two
systems and scanning again. By the way, I have not had these
problems with the 1Ds for human sports or race car photography. I
sure hope there will be feedback from photographers using the new
1DM2 for the kind of work giving you problems with the 1D. Do you
have one on order? I certainly hope you will share your findings,
or at least email me directly with the result.
 
George

In my experience AF innacuracies often result from using more than one focus zone as the camera is left to its own devices to locate the subject you wish to focus on. The AF system has no idea there is a lizard present as it relies on contrast to identify its target.

Whilst the lizard contains contrast, through its bold stripe, the tree bark is presenting competing contrast. Even if the camera is set to try and locate the nearest object if the contrast fools the AF system you will get an innacurate result.

I will guarantee if you set up your camera to use the central zone only placed on the lizard, then re-composed if necessary, it would have been tack sharp or at least much sharper.

I have tested AF systems both, a Canon user and Nikon user approx 6 years a piece, as a full time professional sports photographer. I have raised issues with Nikon and Canon and worked with them to rectify the problems. I have also worked on projects with technicians in Japan.

It is simply too much to ask an AF system to know what subject matter you are trying to nail within the picture area unless the subject fills most of the frame. Some subjects will fair well and others not, as you will appreciate.

Most of the guys I know personally and work alongside, who are world class photographers working for Getty or SI (Canon users), select only the central zone for AF. Very occasionaly they use 8 or 11 points. None of them rely on all 45 zones.

I am half through way changing from Nikon back to Canon but have to say Canons AF system although fast and accurate if calibrated correctly is not quite as flexible as Nikons system on the F5 and 2DH.

Both systems have their strengths and weaknesses and I am not saying one is superior to the other but Canon would do well to expand the cross type sensors horizontally across the frame instead of vertically in no-mans-land, something I am going to dislike but must get used to.

The 8 or 11 zones works for certian subjects but I would not use multi AF type zones for the type of subject you have cited but rather use a single point and recompose. The Nikon F5, D1X and 2DH offer greater flexibility in this respect allowing a cross tyoe sensor to be selected off centre for imporved compostion.

AF systems are no more reliable than auto exposure matrix metering when used under testing and creative lighting matrix metering is unreliable, a hit and miss affair I am not prepared to trust in the heat of the moment. However a single spot reading, incident light reading and single AF selection with user knowledge to over ride the system will stand an better chance of accuarcy.

Too many photographers rely on auto for every situation including, sadly to say, some professional I have come across. Two students I recently employed with degrees in photography did not have a clue about the basics and were reliant on auto everything. I am just glad I learnt the hard way.

Geroge I am not suggesting you do not know your craft but you have already identified a better success rate with an F5 and yet get some good results with the EOS 1DS. If it works some of the time but not others you must adapt to the best method the EOS 1DS prefers for the type of subject in hand, hence the reason fellow pros usually use one AF sensor on Canon cameras as they find multi AF unreliable.

I liken Canons' 45 multi point AF system to a patriot missile send out enough at the target (200 projectiles) and you might get a hit. In the latest desert campaigns, Gulf War and Iraq the US miltary identified only a 45% and 65% success rate on destroying scud missiles; blaming failures on the poor quality of the incoming scud missile! Maybe the next time Tiger Woods fools my AF system I should have a word with Tiger and blame him for moving too erraticaly!

Steve.
 
The camera’s focus is perfect except in this situation, which
unfortunately is one I need fairly frequently. I would be loath to
sent it back and have anyone mess with it because i have no idea
what the result would be. Furthermore, Chuck Westfall acknowledged
this issue with the 1Ds in an exchange of emails we had on the
subject, and this makes it seem even less likely an “adjustment”
could rectify the problem. I was resigned to this situation, but
delighted to know the new 1DM2 would be a fix for me. However, he
also said the 1D was better than the Nikon F5 in this situation,
and this was based upon feedback from ex-Nikon F5 users, and, I
suppose, Canon’s own tests.
As a former F5 user I can verify that Canon's focus is a LOT better. I have used the 10d which is slow but on both the 1d and the 1ds they both ate up the F5 on focus speed and accuracy

I have shot macro with both cameras as well with focus correct.

Just my experiences.
 
Good reply;

Hey, one of the 1D cams that refused to focus in AI servo would work if I picked the focus point in the center, but when I used a side point and focused on a motorcycle chasing another, every single time the shot was way, way off and unusable. This was a camera right out of Canon, Irvine service center that I put in for focus problems.

What the heck could be wrong that a cam can't pick up something using a side point?

When I used a different 1D the very next event, same lighting, same racing, same settings, everything was fine.

I understand there is technique involved. I make my living doing this and I get by just fine with cams that work right. I just don't understand why some are far, far off, why so many AI users have have similar problems, and why can't Canon find these problems on the bench and correct them?

Reliability and QC at Canon are simply unacceptable.
 
As far as I know, all the focus issues we are talking about right here involve Canon's AI Servo, not things sitting still.

Canon focus on things sitting still is fine with me, but it's pretty boring shooting stuff that just sits there.
 
Steve,

I wasn't suggesting you were wrong. I agree there are problems with the accuracy of AF systems and lenses IMHO as a result of poor production line quality control standards.

However I felt the lizard shot demonstrated too much of a reliance of multi AF points and multi point AF was not the best technique for the circumstances.

I have rejected 300mm and 400mm F2.8 AFS lenses and asked for replacements. I have also rejected a Nikon F5 and 2DH for AF innacuracies.

Just before I changed from Nikon to Canon I returned a 2DH which was front focusing and could not be resoved at the time by Nikon. A suggestion that the AF might be slightly out but still within tolerance was definately not acceptable by me and the camera returned.

If I can turn a lens manually and get a sharper image than the AF system it is simply not accurate and acceptable, especiially a tripod mounted test using one shot, single focus point, non servo, on a fixed high contrast target. If it can't get this simple test right I am not going to trust it.

I don't see why a £6,500 lens, capable of producing razor sharp imges, should have its optical abilities degraded by an innacurate AF system. I bet the lens designer wouldn't be too impressed either!

Both Nikon and Canon suffer problems. But from all of the first hand evidence I have gained Canon do seem to have had more than their fare share of AF problems since introducing their digital range.

Lets hope they get their act together from now on!

I sympathise with your predicament. The problem with camera companies is they make wild claims about there equipment. Nikon used to quote an example fo their lens manufacturing accuracy as, ' accurate to within a grain of sand in the middle of a football pitch' and 'fool proof matrix metering'. I have never found either statement to be true even though they try hard to achieve such standards during their R&D work.

Steve
Good reply;

Hey, one of the 1D cams that refused to focus in AI servo would
work if I picked the focus point in the center, but when I used a
side point and focused on a motorcycle chasing another, every
single time the shot was way, way off and unusable. This was a
camera right out of Canon, Irvine service center that I put in for
focus problems.

What the heck could be wrong that a cam can't pick up something
using a side point?

When I used a different 1D the very next event, same lighting, same
racing, same settings, everything was fine.

I understand there is technique involved. I make my living doing
this and I get by just fine with cams that work right. I just
don't understand why some are far, far off, why so many AI users
have have similar problems, and why can't Canon find these problems
on the bench and correct them?

Reliability and QC at Canon are simply unacceptable.
 
Steve

My point was that in my experience if a cameras AF system cant hit a static subject it has always turned out to be innacurate using servo!

Steve
As far as I know, all the focus issues we are talking about right
here involve Canon's AI Servo, not things sitting still.

Canon focus on things sitting still is fine with me, but it's
pretty boring shooting stuff that just sits there.
 
As far as I know, all the focus issues we are talking about right
here involve Canon's AI Servo, not things sitting still.
Oh dear me Steve get your head out of the clouds! Please do not assume that because somebody has shot 'macro as well' means they have not shot anything moving on ai servo.

Also don't assume you are the only person to ever shoot anything moving.

I was referring to moving subjects in my post - the macro as well was an add on in reply to people's posted back focused (sitting relatively still) shots of lizards etc.

Heck even that notoriously bad camera the F5 could focus relatively accurately on a sitting still subject.
Canon focus on things sitting still is fine with me, but it's
pretty boring shooting stuff that just sits there.
--
~ there are those that do ~ and those that talk about it ~
 
touchdown catch, he gestures at the screen. "Eleven guys. Eleven
versions out of focus."
I apologise if it sounds as if I'm getting defensive, but working pros don't necessarily know the ins-and-outs of Canon's AF system. E.g. How many here realise that the first shot in AI Servo mode is very likely OOF? And how many of those eleven realise that? (I know, it's a well known fact, but...)

This thread has evolved into bitching rather than attempting to solve what could be a user problem. No-one has bothered asking about CF settings (which can affect AF performance in a profound way). I understand the need to vent one's frustration, but is this really the right place to do that? (i.e. if the original poster wants to add some more suggestions to the list of things to try, then this is the right place for that -- just venting brings us no-where real fast)

--
Rune, http://runesbike.com/
 
Rune

This is not a bitchy reply.

I work alongside full time many professional sports photographers, including myself, who do know how their AF system works in order to get the best out of it. Maybe not to geek level but enough to understand what the systems strengths and weakness are.

Some of us do like to know how the tools of our trade function. The AF system at the end of the day simply cannot be expected to be 100% accurate but there are ways to set up the system to get the best percentage rates for different types of subject as you will fully appreciate.

This will come through some user knowledge gained from tips which are not printed in the manufacturers manuals for example the one you have highlighted.

Regards

Steve
touchdown catch, he gestures at the screen. "Eleven guys. Eleven
versions out of focus."
I apologise if it sounds as if I'm getting defensive, but working
pros don't necessarily know the ins-and-outs of Canon's AF system.
E.g. How many here realise that the first shot in AI Servo mode is
very likely OOF? And how many of those eleven realise that? (I
know, it's a well known fact, but...)

This thread has evolved into bitching rather than attempting to
solve what could be a user problem. No-one has bothered asking
about CF settings (which can affect AF performance in a profound
way). I understand the need to vent one's frustration, but is this
really the right place to do that? (i.e. if the original poster
wants to add some more suggestions to the list of things to try,
then this is the right place for that -- just venting brings us
no-where real fast)

--
Rune, http://runesbike.com/
 
I use two 1D for supercross and motocross...

But, Canon can't seem to make these things focus correctly and with
any reliability.
Was shooting motox trials yesterday (normally do enduro) with my 1D. Looking through the results, I had a whole bunch that just weren't sharp. Found I'd accidentally set the lens (a brand new 17-40/4) to MF!

I rarely have trouble with the 1D shooting enduro. I've used C.Fn-17 to expand the AF area and select the centre spot. I sometimes stop down a little with the longer lenses, but no more than f4.

Chris.
--
Freelance sports photography
http://www.hockeyphotos.com/
 
Bitching on a message board on da internet ? IMPOSSIBLE.

SMTL :)
touchdown catch, he gestures at the screen. "Eleven guys. Eleven
versions out of focus."
I apologise if it sounds as if I'm getting defensive, but working
pros don't necessarily know the ins-and-outs of Canon's AF system.
E.g. How many here realise that the first shot in AI Servo mode is
very likely OOF? And how many of those eleven realise that? (I
know, it's a well known fact, but...)

This thread has evolved into bitching rather than attempting to
solve what could be a user problem. No-one has bothered asking
about CF settings (which can affect AF performance in a profound
way). I understand the need to vent one's frustration, but is this
really the right place to do that? (i.e. if the original poster
wants to add some more suggestions to the list of things to try,
then this is the right place for that -- just venting brings us
no-where real fast)

--
Rune, http://runesbike.com/
 
Since I started this thread I hope I get to comment.

I don't think its going to help if we know all the ins and outs of exactly how the AI works in minute detail. What I am talking about is some cameras that are simply unacceptable no matter what you do, compared to one that works fine for me.

I don't care to read technical articles on how the AI is designed.

Canon can't make the bad ones into good ones, no matter what settings you use. That in itself is a problem.

I don't think I should have to buy a third body as a floater or become a camera design engineer to get my shooting done. That is what I am buying professional equipment for! (And paying a professional price.)
 

Keyboard shortcuts

Back
Top