Day 14: August 4, 2007
Summary of Findings to Date
So, let's review again. Our many tests of the autofocus system allow us to reasonably draw the following conclusions.
1) The autofocus problems are not caused by sun or heat.
2) They are not caused by operator error. Although, many, many reports out there are obviously operator error.
3) It is not endemic to the new autofocus system design
4) The autofocus of the Canon MKIII is not "Broken."
5) It is not a function of bad lenses.
6) The New Firmware does not improve performance of the "bad" autofocus unit.
It is most likely caused by manufacturing defects.
Does this mean that all of the bad bodies are only in the early batches. Unfortunately, not necessarily. It also doesn't mean that even a majority of the bodies in the early batches are bad. It also does not mean that all of the newer bodies are "good." However, if I were looking to buy I would probably try to make sure I at least had a serial number as high as the "good" body listed in this review. Then I would make certain to run a rigorous test of the af system as outlined here to make sure my numbers were somewhere in the same ballpark. We will cover an approach to do this in a later entry. It isn't a super simple test. But, it's not that hard to do either.
Even the poorly functioning body only performs really poorly when you stress the af system in AI Servo mode. Many photographers can and are getting on fine with the poorly functioning body. Consider that the "bad" body in this review has been in constant rental over the last 2 to three months and very few, if any, have complained of any focusing errors. However, someone did complain about the AI Servo on the "good" body. Probably because, by this time, they had read the many reports of a problem and went looking for one. Again, this just goes to show, a controlled test is required to test for this issue. I have spent way too much time on this issue. I am actually getting a bit tired of seeing the MKIII at this point. But, in the normal portrait type shooting I do, I still cannot tell the difference between the two bodies without looking at the serial number on the bottom, or running a test such as this.
We are a professional forum and I for one don't feel much good can come from bashing vendors (or website operators for that matter 
). However, I find it incredible Canon still has not directly addressed this issue. It is getting about time that something should be said or done. They have the carefully controlled results from here to analyze and they have had time to interview Rob and others. While I find no evidence of heat or sun being the culprits, I have little trouble believing that Rob Galbraith or Edmond Terakopian have received a "bad" body. I know the folks at PPR are frustrated because they have no way of having their "bad" body repaired at this point. The Canon techs do not acknowledge any known issues or repair procedures for this defect when they call.
Coming up, we plan to provide a basic tutorial, including a how-to video on testing your camera. We also ran some more static focus tests with somewhat surprising results. Lastly we plan to look at the performance of the 300mm f/4 IS performance. Stay tuned...
(Update to Summary for Those Who Have Not Read the Entire Article)
I guess I may need to clarify my view on this once again. I have said repeatedly in this article and in the full review that this is one of the finest cameras ever produced. We originally wrote that it is "The best all- around DLSR on the planet." This overall assessment of the camera still stands in the review, and in my opinion. A good MKIII is a great camera without a doubt in my mind. If I continue to cover sports for the school this fall I will most likely be trading up from the 5D. I have no interest in the MKII after shooting both cameras side by side. Rob has another view. But, you can bet if I do purchase a MKIII, I will immediately put it to the test within the return period. I would never intentionally hold onto a body that performed well below spec.
Everyone knows that Chuck is a highly respectable active member of this forum. I have intentionally not tried to bring him into this discussion due to my respect for his position at Canon. He has continued to visit and answer other questions during this time. In short, he has no doubt seen the article. If he wishes to comment he may do so at any time. As I said, this is not a place to bash vendors. For the most part everyone here has respected that policy.
I have also stated repeatedly how difficult this problem is to replicate without a specific test to do so.
However, when someone does have a bad body that is past the 7 to 14 days that most retailers allow for returns, they seemingly have nowhere to turn at this point. The fact that there are bad bodies in circulation is really beyond dispute at this point. Unless I have somehow missed it, there has been NO admission that there are bad bodies in circulation.
The occasional comments on unofficial channels such as those cited above is commendable and I have noted this in the article. I do believe they are looking into the issues. However,
the gaping hole at this point is something official on the Canon website and sent to the repair techs allowing for return and adequate insepction of bad bodies. Camera owners have no desire to send off a camera to simply have new firmware updated, or nothing really done to correct the issue. A trip to the service center at this point will probably reap nothing but frustration and lost time for the pro.
Rob says there is a "Roadmap" for investigating these issues that has been agreed to by he and Canon. If this is so, it should be published through an official channel as well as unofficial ones such as this forum, or his site. They have known about these concerns for well over two months and made no official comment on their website whatsoever.
Now, as a business owner, I can certainly understand the risks associated with commenting on the issue until they can define the parameters of the problem and prepare a fix. Our tendency is to want to believe the worst. So, they would receive a deluge of cameras that for the most part have no problems. So, what to do.
Suck it up and test your camera yourself. The right way. The empirical way.
You will spend at most two hours testing your camera according to the simple protocol we will soon outline here. You've probably already spent much more than two hours reading forums about this problem by now. So why not find out definitively whether your camera is among the afflicted or not. It most likely is not. Almost all of the other random "tests" on the internet are useless as there is no control group for their "test." By following this simple protocol exactly you will know exactly where your camera falls on the performance spectrum. We are going to provide the control groups for you. You do not need to have a MKII or another MKIII to test against.
You will find out whether your camera is as good as the "good" MKIII in our tests, at least as good as the old MKII, or just plain bad. Then you can rest easy, or send your camera in for repair armed with specific data that you can reference. This should help Canon get to the bottom of the issue as they will only be receiving bodies that are truly affected with an autofocus issue.
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.
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
