5D: Focus Accuracy within spec?

[robert_ades1]

<p>I just received a refurb 5D from Adorama today. In testing it out, I'm observing some apparent front focusing issues. Am I expecting too much, or is there a problem.<br /> <br /> This cropped shot was taken at ISO 200, 50mm Canon f/1.4, at f/1.4, tripod mounted, in Raw, converted in LR 2.3. Firmware is 1.1.1. It was taken to test the accuracy of the AF detection system.<br /> <br /> The number "24" filled the center the viewfinder focus square, and should have been the point of accurate focus. The arrow is pointing to the sharpest point in the image, sits outside the CF square, but inside the spot metering circle.<br /> <br /> My question: Is this a typical miscalibration issue or am I expecting too much?<br /> --<br /> ----------------<br /> Robert A<br /> Canon 20D</p><div></div>

[zml]

<p>And what happens when you are photographing something different than a ruler at f/1.4 at close range? Are the pictures sharp when shooting at more reasonable f/stops and focusing distances..?</p>

[robert_ades1]

<p>Sure, at f/5.6 and 20 feet, focus should be dead on. But, I may also be shooting someone's eyes, or a strand of hair, where imprecise focus can ruin a photograph.<br>

This camera is expensive, and returnable. I simply want to know if it is functioning within normal operating parameters</p>

[bobatkins]

<p>Hard to say from one image. Since the AF zone may have covered parts of the target at different distances from the camera, it's difficult to say where it <em>should</em> have focused. Could also be a lens issue of course.</p>

<p>Why not try "shooting someone's eyes, or a strand of hair" and see how that looks?</p>

[PuppyDigs]

<p>My 5D is generally pretty accurate even at close focus. However the EF 50 1.4 USM was a poor performer on my 5D and was never any good in low light or close focus. On the other hand, my EF 50 1.2L fares far better. You should try several other lenses to isolate the problem. It's been my experience back- and front-focus issues more often lie in the lens, not the camera. </p>

[mercedes1]

<p>You might have 'AF point activation area' set to 'expanded' in the custom settings. When shooting obliquely like you are here this is the normal side effect of it. The expanded points surround the center focus point and are there to help it when tracking moving subjects and the such. The expanded points at the bottom of the circle are picking up the closest area that can be focused on, exactly like they are supposed to do.\<br>

My 5d does the same thing if I forget to turn the expanded points off and shoot obliquely to the subject.</p>

<p>Hope this helps.</p>

[btmuir]

<p>Expanded af points only activate during AI Servo mode.</p>

<p>I'm assuming you were using one shot AF, correct?</p>

[frank kennedy]

<p>Have you tried the lens on a different camera? I had a 300 mm f4L on my 5D and it focused like this photo. EXIF attached. The center focus is on the unpainted mark in the center.</p><div></div>

[frank kennedy]

<p>I tried it on my 30 D and the photo looks like this. Focus is still on the unpainted mark. By the way. I mounted my 24-70 on the 5D and the focus point was in the middle of the depth of field.</p><div></div>

[robert_ades1]

<p>All good answers, thank you. Yes, I do have the camera set to "expanded", however I was using one-shot AI focus. My sense is not there there's front or back focus, but that the focus sensor is misaligned with the mark in the viewfinder. I'll have to test this some more, as was suggested.</p>

<p>What I cannot understand is what a lens has to do with it. If the camera body is driving the lens focus motor until satisfactory focus is observed inside the camera, how can a particular lens mislead the camera?</p>

[robert_ades1]

<p>Okay, I did a little follow up work. I disabled the expanded setting and retried the testing. This time, I used the 50mm at f/1.4, f/2.0 and f/2.8. Then, I used the 24-105L, at 105mm, at f/4.0 and f/5.6.<br>

The tripod was steady, but the results are a little mixed:<br>

With the 50mm at f/1.4 and f/2.0, the images still appear a bit front focused, but at f/2.8 where the DOF expands a bit, the point of sharpest focus seems to coincide with the focus sensor.<br>

With the 24-105, at both f/stops, the point of sharpest focus seems to be the focus sensor, although the DOF is a little broader, making it harder to tell.<br>

So, this begs an interesting question: Why does the lens make a difference when the focus sensor is in the body? Or rather, how do you calibrate a lens?</p>

[robert_ades1]

<p>Here's the 50mm shot at f/2.0</p><div></div>

[neill_farmer2]

<p>Off focus can be due to either the lens or the camera. The contol method used is an open loop so there is no way for the camera to know if it achieves focus or not and no closed loop to measure that error, calculate a correction and try again. I think their is a reason for building the system this way and it has to do with speed of focus. Errors of 1% are considered to be pretty good for open loop systems, but 1% is too much if the lens is at 1.8 and the photo is a portrait. Canon does seem to be able to calibrate cameras and lenses so the error is essentially zero. If the small error you are having is a problem for your style of photography it might be worthwhile talking to Canon and getting your equipment calibrated.<br>

Neill</p>