Consistently blurry photos with L lens on a 10D?

[marissa_c._boucher]

Hi everyone,

 

I'm getting really frustrated with blurry photos I'm getting back

from Wedding shoots on one of my backup cameras. It's a Canon 10D

with a 24-70mm 2.8L EF lens.

 

I primarily shoot on a 20D with a 70-200mm 2.8L IS lens and all my

shots are usually sharp as can be.

 

I'm certain that the exact same auto focusing mode and points are

selected between the two cameras. I have ONE SHOT selected for the

autofocus mode and the center focusing point selected. I often find

that the background is in focus rather than the subject. Or if I

focus on the eyes and then re-frame the photo the center is in focus

rather than the eyes themselves.

 

Now I'm completely aware of the solid green dot in the in-camera

readings display when the focus point is found. I've also paid close

attention to the red rectangle's initial focus point, then I also am

holding down the shutter halfway to hold the initial focus point in

place until completetion of the shot.

 

I'm holding the camera in my hand now and testing shots on things

around the room and the problem isn't occuring at the moment, of

course.

 

I dunno, this lens just seems slower, not as sharp, and overall

lower quality in comparison to my 70-200mm 2.8L IS. Anyone have

these two lenses and notice a quality difference? Consistent clarity

is just so difficult with this lens, if that is indeed the source of

the problem. I don't know why Canon didn't make this an IS lens.

Keep in mind that low light isn't the issue here though, I've

checked the properties on several blurry photos or incorrect focus

points and the exposures and iso settings are perfect for the

situation or available light.

 

Any ideas?

[marymac]

When you mount the 24-70 on your 20D do you have the same problems you do on the 10D? I haven't used the 24-70 so I don't have any insight into how that particular lens performs but since it has the L designation it should be a great lens. Having said that, though, I returned a 70-200 f/4 L for back-focusing problems just those like you describe. Could be the lens or the body - you'll have to try out differnt combos to determine the problem. Also, I think that some lenses just work better with certain bodies; I have a 28mm 1.8 that is super sharp on my 350D and backfocuses terribly on my 20D. Both are at Canon now to see if their service dept can help them get along...

 

Good luck - Mary

[peter_white2]

If you focus on the eyes and don't reframe the shot, are the eyes in focus?

 

Remember, you're focusing on a plane. In theory, everything on a parallel plane to the film is in focus. If you place the center of the frame in focus, and then move the frame, what was at the center of the frame now moves ahead of the plane of focus.

 

Take an example. You have a wide angle lens, about 17mm. You're standing 10 meters from railroad tracks. You point your camera down the tracks to a railroad car 20 meters away. You focus on the railroad car. Your lens is focused at 20 meters. Now, you turn and frame your image with the railroad tracks perpendicular to your lens, or parallel to the film plane. You keep the lens focused at 20 meters. A plane, 10 meters behind the railroad tracks is now in focus. You take the picture. The railroad car is off to the side in your image. It's 20 meters away. But it's 10 meters in front of the plane of focus, so it's now out of focus. It was in focus when you pointed your camera at it. But now it's out of focus, because you have moved the plane of focus. To get the railroad car into focus, you would need to set the lens to focus at 10 meters, even though the car is 20 meters away. You need to focus at 10 meters because the railroad tracks are ten meters in front of the camera, and the railroad car is on the railroad tracks. And the railroad tracks are on a plane, 10 meters in front of the camera.

 

Everything on that plane, 10 meters in front of the camera, will be in focus. And the car is on the tracks, in that plane. It's 20 meters away, but the plane is 10 meters away, parallel to your film plane.

[digitmstr]

The 24-70L is an excellent, compared to ANY thing out there. Remember, it is better than the 28-70L which it replaced, a legendary lens itself.

 

Sometimes, bodies and lenses need to be calibrated. But, you also speak of wishing this lens had IS, which of course has NOTHING to do with focus accuracy.

 

First of all, before you assume the lens as the root of the problem you must identify WHAT the problem really is: is it motion blur? Or is it focus inaccuracy? The first is easy to fix with faster shutter speeds. The latter is caused by one of two main issues:

 

1) user error

 

2) lens/body miscalibration

 

Of course, there is a 3rd possibility of a broken lens but, it seems not be the case here, from what you have said.

 

My suggestion would be to move the AF function to the * button (AE lock) via Cf 4 (if I remember correctly).

[KenPapai]

This focus lock/recompose issue is covered and diagramed in the current June 2005 EOS magazine. Like described above you are shifting your focus plane and what you might have been point at is now at a different distance from the original focus point. It's a pretty easy mistake you make until you realize the source of the problem. I wouldn't blame the 24-70L lens. If you have enough light switch to manual focus and dial in the correct focus when you frame your shot. I know it is sort of a game to manual focus on these 1.6 DSLRs, but I suggest you try it.

[pturton]

Marissa,

 

I have found that the focus rectangle on my 10D is actually considerably smaller than the actual area used to AF. My 85 f/1.8 lens often back-focuses on a higher contrast object even though the focus rectangle is completely inside the subject's head that I'm shooting at 7 to 10 meters. This happens mostly when the background is better lit and has more contrast than my subject.

 

Possibly your 10D has a similar AF weakness.

[sergey_oboguev]

Assuming you did not accidentally set camera into AI Servo, check for back-focusing.

[jim_larson1]

The 10D AF sensors cover about 2.5 times the area shown on the engravings.

 

One of the active threads in this forum actually has an image showing this relationship.

 

I have often wondered if these large sensors would be acceptable for critical focus applications (such as a 85/1.8 or 85/1.2). Although, I imagine most people who have 85/1.2's would be mounting them on 1' series bodies.

 

I seem to do ok with my 10D body, F4 zooms and my 50/1.8.

[anna yu]

Marissa, my 24-70 never worked well with my 10D but is stellar with my 1DMk2. So probably

it's the lens/camera combination that's the problem.

[fernando lopez]

Check past threads; the 10 D seems to be famous for backfocussing problems. I also had the problem when I used it with my 17-40 f/4 USM L.

[ron c sunshine coast,qld,a]

I have heard many times of back/front focusing problems and experienced it myself with several lenses.

<BR>Only one of those lenses was bad enough to really need recalibrating-a 75-300 IS that consistantly underperformed.At first i couldn't put my finger on what was wrong.It just wasn't very sharp.Even some quick manual focus comparisons didn't show anything seriously amiss.

<P>Only after doing an acurate test with a proper focus test chart did the problem show itself.

<br>Canon recalibrated it as part of a general service (it wasn't under warrantee) and now it is focuses with *perfect* accuracy! Better than it really needs to be i must admit :-)

<P>I remember reading a thread from another forum about a guy who had his not-quite-right 70-200/2.8 recalibrated.Same story there- it's now incredibly accurate and consistantly SHARP

[paul -]

Here's a diagram showing what Jim mentioned about AF sensors being larger than the little rectangles on the focus screen, which can sometimes explain why focus appears to lock onto the wrong thing.<div></div>

[digitmstr]

Of course, if Marissa doesn't post any samples it will be hard to see what may causing her focus/blur issues...

[fernando lopez]

Come on Marissa! We're curious! Samples is what we need! :-)

[guy_howard]

Although the image characteristics are different, and focusing behavior seems better with the 70-200, they both make wonderful pictures on my 20D.

 

One idea I had for accounting for the different focusing behavior on your 10D is that the focus target may generally cover a smaller perecntage of the focus sensor area with the 24-70 than with the 70-200, simply due to the shorter focal lengths when using the 24-70. This would result in the sensor locking to an unintended location when the focus sensor area is much larger than the target. I found this to be something I had to be aware of with my 20D, and would be even more of an issue with the 10D.

[marissa_c._boucher]

Here's an example of the backfocus problem I mentioned. This was shot on a 10D with the 24-70mm lens I described. Shutter was 1/250, F2.8, ISO-200, no Flash, AV mode. Zoom in on the background if possible and you'll notice how the background is definitely where the focus is sharpest. At first glance it may look like he's reasonably in focus but this was shot at 2.8 and it's extremely noticeable when viewed as an enlargement. I do have worse examples but I think this explains enough. Just look at the brick work behind him to see how sharp it is. That area would never been in focus at 2.8 if the autofocus would've properly targeted his eyes or body.

 

This photo is still in raw form so please just comment on the issue at hand. :)

 

I'm going to send the camera and lens into Canon ASAP. I'm going on vacation so I figure why not, they're still under warranty anyway.<div></div>

[paul -]

Many 10D bodies are admittedly prone to backfocus (I thought mine was one until I understood how big the AF sensor area really is, and adjusted my technique accordingly), but which AF point did you use?<p>Overlaying your shot with the AF point diagram from above shows that if you used the AF point on his face, it may have focused on the back wall/door. If you used any of the other points, there is definitely some backfocusing going on.<div></div>

[jim_larson1]

At this range. . . .at F2.8. . .focus recompose could easily be a problem.

 

Read this: http://visual-vacations.com/Photography/focus-recompose_sucks.htm

[axel-cordes]

Paul: You are absolut right - this is what I encountered already with my EOD 30 and now with my 10D. Finally I do not trust AutoFocus anymore and use CF 4 set to one, can use AF by pressing a button, but mainly maually focusing. Maybe an other hint is welcome here: For comfort I was using my left eye to view & focus - but I changed to the right now as it looks like that I can see and focus better with it, eventhough I would say they are both equal good/bad...

Digital makes it easy to test all through. Regards Axel

[phast1]

I had the exact same problem. I have a 20D and I've been using a few lenses with no problems at all (ef-s 10-22 f/3.5, ef 17-40 f/4, ef 70-200 f/2.8 IS USM ). I bought the 24-70 f/2.8 to fill the gap (40-70mm), and have a quality walk around lens. Noticed out of focus after first usage. Did some test with tripod, higher ISO, faster shutter... nothing worked! I then did some tests with tripod shooting text on paper and aimed for a specific word. Back focus was obvious. Tried the test with the 17-40 at same focal length, no problems. I've read on other forums there are 'sometimes' problems with this lens. Some think it's poor quality control, some believe it's the lens/body relationship. For the price of this lens I expected much better, and even if I was willing to settle for a more finicky lens this was completely unacceptable. Brought it back (unwilling to send it to Canon) and bought the ef-s 17-85. Though it's not L series, it's damn good and the IS is a bonus given that it's only an f/4 wide open.