New carping about high-end Canon SLR focus

Photographer and consultant Rob Galbraith isn't alone in griping that the autofocus feature in the new $4,500 Canon EOS-1D Mark III is a step down from that of its predecessor, the 1D Mark II N.

Seattle Times sports photographer Rod Mar posted two blog entries last week criticizing the camera. He complained of fuzzy artifacts in his two brand-new models after photographing several games.

Canon's EOS-1D Mark III

(Credit: Canon)

"Believe it or not, Canon's new flagship camera...has trouble focusing. And it isn't a rare, isolated problem," he said in the first blog entry. "The camera has trouble tracking subjects in its autofocus. Further, the problem is worse in bright, sunny conditions with low ISOs."

In the second post, he complains of the fuzzy-focus problem, which he believes is different from ordinary out-of-focus softness. And he adds, "Instead of choosing frames by judging light, composition and moment, I'm choosing them solely by the accuracy of the focus...Editing the Pittsburgh game was tough because of this."

It's not all bad news. Mar also praised some camera features, including its "amazing ability to shoot in very low light." And he also gave high marks for the camera's results when shooting at high-sensitivity settings that, in lesser cameras, produce multicolored speckles of image noise: "The Mark III shot at ISO 3,200 looks like the older cameras' (shot) at ISO 1,000."

[GGGlen]

1912

Titanic strikes an iceberg, sinks to bottom of ocean with huge loss
of life.
It's not all bad news. The decks are clean and shiny.

[dandear99]

2 Canon Digital Rebels XT's BOTH HAVE FOCUSING PROBLEMS

Both of my Canon Digital Rebels XT's have focusing problems. One was almost brand new and the other was about one year old.

Horizontal photos were out of focus and vertical photos were sharp. Took me about one month and 500 photos took figure out what was wrong.

Called Canon Customer Service Dept. and lady acknowledged that there is a flaw across their COMPLETE LINE OF DSLR's. Since it was out of warranty, they charged me $200.00. I believe this should be a recall issue, but since it's really hard to detect, Canon prefers to keep it quiet.

My friend called Canon a couple of weeks later and the Customer Service Rep. claimed he's never heard of such a problem. Both of them REFUSE to allow me to talk to a technician.

Therefore, as a long time Canon camera fan, I am switching over to Nikon. Shame on Canon for hiding a serious problem.

Any questions, please e-mail me at [e-mail]

[R Jueneman]

Carping is not helpful -- facts are needed.

Mere carping about the problems with the Canon 1D Mark III is not particualrly helpful. Without FACTS, it is difficult to distinguish between operator error, a momentary problem with servo tracking, and a persistant design flaw.

I don't shoot sports, but I believe that the problem is that the camera is tracking focus on the NEAREST object it sees, rather than the original point of focus. I also believe that the problem may be exacerbated by using the auxiliary focus-assist points, and almost certainly by using the automatically-selection option.

In the picture of the football player Rod Mar complained about, the football is in sharp focus, but the players face is blurred. Now, the depth of field of a 400 mm lens in the Mark III at f/2.8 is only 14 inches at 100 feet, and a scant 3 inches at 50 feet, so it is perfectly possible for the football to be sharp but the player's face to be out of focus. So the first fact that needs to be recorded is the distance of the subject.

The second question is whether the particular lens/camera combination has been accurately calibrated for focus displacement. In my tests, several long lenses required adjustment of anywhere from 12 to 19 points, or as much as twice the single-side depth of field -- a whopping error. Rob Galbraith stated that he had done such a calibration, but I don't know if Rod Mar did or not.

In addition, although I don't expect this in a news article, it would be very helpful to document the AF settings that were in use. In addition, was the subject shot in RAW or JPEG, and what kind of post-processing software was used?

But perhaps the most important fact would be where the photographer initally focused, and how much tracking was involved. Was the face the point of focus, or the ball? Or was it just somewhere in the the action, with the primary intent to capture the moment with a reasonable composition? I suspect the latter, given fast moving action and a high-speed camera.

A final question, given the admittedly outstanding performance of the camera at higher ISO speeds, why was he shooting at f/2.8 at ISO 200, at 1/1250 of a second? Wouldn't f/4 at ISO 400 been more reasonable for an outdoor event in bright sunlight, in order to gain a little more depth of field?

Please note that I am not defending the camera. I also believe there is a problem, but that it is caused by the design decision to focus on the nearest point seen by the sensors that have been selected. That works for bird photography, where you want the beak in focus, but not in sports, where the face is more important visually than the ball.

To me, it doesn't make any sense to have 45 auto-focus points but only use one of them. I would like to be able to select those subset of those points and have the camera focus on the point where the largest number of points are in focus, rather than selecting the closest point. As I understand it, that was the algorithm used in the 1-Ds Mark II.