7D Focusing Anomalies With 24-105mm IS Lens

[digiquisitive]

Expert advice will be appreciated, please. My Canon EOS 7D (not MarkII), coupled with my Canon 24-105mm f/4 L IS lens has started misbehaving. The lens is 8-9 years old, the 7D has possibly 50K shutter actuations. Shooting in a relatively dark forest recently, on tripod, using self-timer for release, ISO200, matrix metering, shutter down to ½ second or slower, some use of internal flash, fully-charged battery, using AF Point Expansion for focusing, some of my images (totally hit and miss, random, unpredictable) are not focused properly. About 15-20% of the images show a focus considerably different than where DPP shows (red square) the camera was directed to focus. For example, in many of the incorrect images, DPP shows the red square clearly on a large tree trunk (which agrees with my placement), but the trunk is out of focus and a small maple vine 2-3’closer to the camera than the trunk is sharply focused. This seems to have occurred with IS turned on and off, but I cannot be sure of that (EXIF info does not tell).

To subsequently analyze the problem, I took dozens of images,handheld, with the same camera and lens, but using an appropriate ISO for handholding, and in a brighter outdoor setting (again, matrix metering, AF Point Expansion, IS on, no flash). I found the same situation – hit and miss out of focus conditions, with DPP showing correct AF point placement but the camera/lens again focusing closer than the AF-indicated subject. In some cases, images were totally out of focus. In others, the AF point (red square) was on a subject dozens or hundreds of feet away, and the entire image was out of focus except for some shrubbery 20’away and completely over at the left of the frame (this happened with more than one image). Turning IS off did not really make a difference that I could detect in this handheld test. Changing from AF Point Expansion to Manual Selection (single AF point) resulted in fewer problem images, percentagewise. Overall, the test was producing about the same percentage of incorrectly-focused images as the previous shoot (about 25%).

Changing to an old 35mm camera kit lens (Canon 28-90mm, AF, no IS) resulted in no image problems, whether using Manual Selection or AF Point Expansion. The problem may thus (??) have been narrowed to the use of IS and/or AF Point Expansion and the 24-105 L IS lens, or a combination of those. I was quite pleased and surprised with the quality of the images coming from the non-IS 28-90 kit lens, and its focusing accuracy!

One other factor. During the real shoot, using the 7D on a tripod, shooting with self-timer release (10 second), atg one time and for one exposure the timer stopped beeping and its light extinguished, and the camera just froze up. Powering it down and back up seemed to clear the problem. I do not know whether the lens IS was turned on or off.

Advice would be appreciated. Does this sound like a lens (IS) problem, or a problem with the dual Digic processors in the 7D, a combination of those, or what? Images shot with the 24-105 during this test that are properly focused are very sharp, as they always have been. What further testing would you suggest?  This is a "first" for this camera and lens, both of which have been bulletproof until now. Thanks

[ScottyP]

Hi.

Perhaps your camera has the opposite problem to the old saying that someone can't see the forest for all the trees (it can't see trees for all the forest being in the way).

You say it works ok if you set a single AF point. That is exactly what I would have been doing in the first place, personally. Particularly in the shot of a woody vertical vine in the foreground, a woody vertical tree as the subject and woody vertical trees all around and in the background. That sounds pretty challenging for the AF system to figure out.

Did you ever take a similar picture before NOT using single point AF and have the camera do a perfect job of guessing which tree was your intended subject, or is this your first time shooting this exact scene?

I assume the camera and lens work fine for normal shots of individual subjects?

[digiquisitive · Accepted solution]

Thanks, ScottyP.  Your response, and that of ebiggs1 (thank you, also) prompted me to do some additional testing with this lens, specifically testing AF Point Expansion versus single AF point (EXIF shows as Manual Selection). ebiggs1 is correct -- I do not think there is anything wrong with this lens or camera (whew!). And, yes, ScottyP,  I did take this exact picture before with this lens, using AF Point Expansion and placing the AF point in the same place -- and that image hangs next to my desk as a gorgeous framed 20x30 enlargement.

My testing (fairly extensive, very carefully planned and executed), shows the same (but very infrequent) anomalies as the original images that started this thread, whether using AF Point Expansion or Manual Selection -- both failed on occasion during my testing. But, I have learned from the testing, and that will be valuable in future shooting (still learning after 60+ years of shooting, mostly with Canon). The summary is that this is not a problem to be concerned with -- as I hope is explained below.

This testing appears to validate several things -- all of which are already well known, but provide a good refresher for at least me.  First, that "electronic focusing" is excellent but not foolproof. One alternative is to use LiveView and manual focusing when I feel the camera could be fooled by the subject to be focused on (scenics, no movement, tripod). Second, consistently correct focus appears easier to achieve when the subject has sharp, well-defined lines, little detail (not complex or intricate), with adequate contrast and adequate brightness (for example, small Fall leaves, all the same yellow color, and at substantial distance, did not seem to produce the most consistent focusing results compared to larger, simpler, closer subjects).  Third, the single AF point may be (believe it or not) too large to achieve the desired result (distant objects, in particular, may be so small within the frame of the single AF point that unwanted subjects may end up being focused on instead of the desired subject) -- the remedy may be to use the even-smaller single-point Spot AF.  Fourth, there are a lot of variables in play at once in the making of any one image, and those may or may not be obvious -- the remedy for this, when possible, may be to shoot many images of the same subject, varying whatever technical attributes of the camera or lens are within the photographer's control.  For now, and based on this testing, I will give more thought, when the situation allows, to which AF area to use, and how to use it (placement, size, appropriate subject, etc.).  Of course, not much of this applies to my normal subjects -- waterfowl exploding off of a body of water.  Thanks again for the ideas.

[ScottyP]

If you have been shooting 60+ years you no doubt know some things few of us know about cameras.

I hesitate to give advice under the circumstances but in general you only want to let the camera choose AF points from a field of possible points if you have a moving subject, like a sporting event. You may need to give the camera that leeway in a fast sporting event.


If you are shooting something unmoving, like a tree trunk, it is better to select a single AF point and put that point on your subject tree. Same thing if you have a wide aperture prime lens and you are shooting a person;, you definitely want to pick a single AF point and you place that point directly on the subject's nearest eye.

[ebiggs1]

"... "electronic focusing" is excellent but not foolproof."

+1    A smart man you are!

[ebiggs1]

I also think this was not the best place for AF expansion.  I would use center point and put it where you want it.

I don't think there is anything wrong with camera or lens.