Thinking back over what I wrote yesterday regarding operator error I realize that I worded it in a less than loving way. My apologies to those offering user experiences and other opinions.
Initially, I did not really think there was an issue either. Even after reading Rob's report I thought there probably wasn't really a problem. My experience initially reviewing the camera did not show any real issues. However, he is a respectable voice in the industry. So, I decided to investigate the issue further. Taking his running test and further refining it showed that indeed there is a difference between the two autofocus systems under certain conditions. These conditions are not very relevant for many photographers, but for others it amounts to a large problem.
While user experience reports can be highly valuable. They do not fall in the same class as a carefully controlled study. They amount to anecdotal evidence. At this point we have lots of anecdotal evidence and very little empirical evidence to suggest a problem. This running test and other carefully controlled studies offer a better chance to drive down any issues. This study design is obviously not the only one that may yield useful information.
This is a difficult issue to try and troubleshoot with a small chance of success from the end user standpoint. However, if enough people do enough controlled studies we might glean some useful information.
In this spirit we will be offering study awards. I realize very few, if any, will take us up on this offer. But, for those who sincerely wish to get to the bottom of this issue here's a little help.
Canon 1D MKIII Autofocus Study Awards:
$100 for the first person to replicate the running test using the methodology outlined on this page. You must use the same lens/ camera combination and the same settings. Reset all camera settings and Custom Functions prior to making your settings. (Canon 300mm f/2.8L Lens, AV f/2.8/ AI Servo/ Auto Select Focus Point/Shoot in Large Jpeg/Ensure Focus Lock/Tripod Mount etc.) You must do at least 10 runs under the f/2.8 conditions and 3 at the f/5.6 condition. You also agree to provide all your results and settings for publication here for all to see and share. This includes all unprocessed jpegs.
$150 for the first person to replicate the running test and find a significantly different result. Reset all camera settings and Custom Functions prior to making your settings. (Canon 300mm f/2.8L Lens, AV f/2.8/ AI Servo/ Auto Select Focus Point/Shoot in Large Jpeg/Ensure Focus Lock/Tripod Mount etc.) You must do at least 10 runs under the f/2.8 conditions and 3 at the f/5.6 condition. You also agree to provide all your results and settings for publication here for all to see and share. This includes all unprocessed jpegs.
$50 to the first two people who do a quality study of any of the conditions in the Ideas for Further Study page. The study must must be rigorous and easily repeatable. If you have questions about the validity of your approach please feel free to email me through the contact us link on the website.
$50 to the first two people who perform a study that pursues an entirely different avenue regarding this issue. The study must must be rigorous and easily repeatable. If you have questions about the validity of your approach please feel free to email me through the contact us link on the website.
Everyone welcome! You do not need to be a member of Pro Photo Home to pursue one of these studies. In fact, if you are not already a full member, you will receive a free Lifetime Membership as well.
To navigate this review use the arrows at the bottom of each page and/or the content box at the top right of this article.
So far so good. Reinforces my theory that a problem such as Rob saw must be a manufacturing defect. Let's be realistic: It would actually be difficult to design an AF system that would be negatively affected by temperature.
The fact that Rob tested two different bodies and found virtually the same defect in both may mean it's a component problem with a particular batch, which was a different batch than the one from which the camera you tested, Drew, was made.
It's a shame there had to be a controversy over this new camera, even to the point where folks canceled orders. Even the best home run hitters in baseball don't hit home runs every time at bat, and even a whole game devoid of homers doesn't mean their capacity to hit them is diminished.
-Noel
P.S., your review's second page is a bit wide as viewed in IE7.
Last edited by Noel_Carboni; 06-27-2007 at 09:28 AM..
Even the best home run hitters in baseball don't hit home runs every time at bat, and even a whole game devoid of homers doesn't mean their capacity to hit them is diminished.
You HAVE got to be kidding. The difference between us and the masses is that we ARE expected to come back with 'the' shot EVERY time. I can hear it now..... Lets just go into the church's basement for your portraits, it's too hot and bright out here. MY batting average HAS to be better than .500, it darn well be better than .750 or I can trade all my gear in for a mini van and start shuttling kids off to soccer practice.
Good thing the Super Bowl is held in February and in the evening.
Drew,
I appreciate your effort to evaluate the AF quality in the Canon 1D MkIII. As one who has taken delivery of an early model, I am most interested in how the AF is working. The football season is still a few months away, but concerns about the AF system's effectiveness are weighing on my mind right now.
I would have liked to see some comparisons of the "runner" standing still at a few distances from your camera to see how the AF works on a non-moving object. In addition, there seems to be some motion created blur which indicates that a faster shutter speed could have resulted in less overall blur.
Thanks for contributing to our understanding of the issue. My hope, of course, is that Canon is listening and will make appropriate adjustments to both currently manufactured and future Mk III bodies.
Nope. Perhaps you misunderstood me. I meant to imply that Canon has quite likely designed a great camera, and that several may have gone out the door with a particular problem that is not seen in every one. Drew keeps proving the camera he has his hands on does NOT exhibit the AF problem seen by Rob Galbraith.
The issue here is whether the Mark III has an overall DESIGN flaw or Rob's cameras have a common MANUFACTURING flaw. The sample size is WAY too small to prove it DOES have a design flaw, but just one camera shown to be working properly can go a long way toward proving it DOES NOT.
I got in late last night from testing the camera at a local Atlanta professional soccer club with Ron Holt. I just wanted to get something up.
I am planning on doing more running tests today with additional settings and a higher ISO to yield a faster shutter throughout the run. I spent at least two hours yesterday in the heat lining up and running at the camera. While I was generally satisfied that all was working well. There was too much motion blur due to the low shutter speeds to produce consistent and definitive positive results. This was especially true of the ones where I was wearing all black to try and more closely simulate the runner on Rob's site. I must say I am not really looking forward to running at a camera over and over again today. It is almost as hot as yesterday.
The difference between us and the masses is that we ARE expected to come back with 'the' shot EVERY time. I can hear it now..... Lets just go into the church's basement for your portraits, it's too hot and bright out here. MY batting average HAS to be better than .500, it darn well be better than .750 or I can trade all my gear in for a mini van and start shuttling kids off to soccer practice.
Linear Mode
