Slighly out-of-focus action shots with Canon XTi

nobody

I have a Canon XTi and a 70 - 200 mm f/4L, non-IS lense that I use for most of my action shots. I shoot sometimes in RAW and sometimes in jpg, but always AI servo and usually single point focus when doing action shots. When shooting races like this, I often get 4 or 5 shots of an individual (the reason I like jpg over RAW sometimes) and throw out those I don't like. The thing that bothers me the most is the large number that are slightly out-of-focus photos. It's not that bad, but, just enough to be noticeable when inspecting the full size file. The problem is the subject just doesn't stand out against the background as much as I'd like. Attached is an example, but it's not full size, you may still notice it to a lesser extent.

Would using a higher end camera like a 5d solve this problem? Any other suggestions? THanks.oof.jpeg

ziggy53

The combination of single point AF and AI Servo is appropriate for this shot, but make sure you choose the correct AF point with respect to the subject.

The AF section of the Canon 5D is very similar to the Canon 20D/30D and may not be any more appropriate for sports than those cameras. (The Canon XTi is likewise very similar in AF.) The Canon 40D uses a different AF section and the jury is still out as to whether it is improved or not over the previous cameras. (Most tests indicate that it is improved.)

The ideal Canon camera for sports is the
1D MK II/MK IIN/MK III series. The AF section on these cameras is far better than the previous cameras in both accuracy and speed, especially using AI Servo.

nobody

Thanks for the info. My interest in the 5d was due to the fact that it's so much less expensive than the 1D and didn't appear to have the action-related autofocus problem the 1D has. But if it's only marginally better than the 40D, then it's not worth the extra money. I wonder how the new 450D (XSi) comes up in terms of AI servo autofocus. THe best option may be just to wait for the next entry-level Canon DSLR that follows the XSi, since my XTI is only a year old at this point.

ziggy53

nobody wrote:

Thanks for the info. My interest in the 5d was due to the fact that it's so much less expensive than the 1D and didn't appear to have the action-related autofocus problem the 1D has. But if it's only marginally better than the 40D, then it's not worth the extra money. ...

My suggestion is that the 40D has (potentially) improved AF versus the 20D/30D/5D/XTi. Responsiveness is much improved versus any of the Rebel line.

The 40D is probably the best bet in Canon for sports unless you go to the 1D line. Since I have both the 1D MKII and the 40D I can say that with some confidence.

I also have both the Canon EF 70-200mm, f2.8L USM and the EF 70-200mm, f4L IS USM and I do believe that the f2.8 version can find find focus a bit faster than the f4. I believe this is partly because the center focus point of the 40D is optimized for lenses of f2.8 or larger. (The optimization is in the form of extra sensitivity which is valuable in both low light and with subjects which have less than ideal focus targets, like human bodies.)

evoryware

I would not shoot AI Servo and Single point. I made the mistake of trying that at the New York Marathon and was not happy until I turned on all points. If I have center point selected, that's the only one that will grab focus. If the runner moves out of that center point, I will try to get whatever is behind the runner in focus and it will take the picture.

nobody

evoryware wrote:

I would not shoot AI Servo and Single point. I made the mistake of trying that at the New York Marathon and was not happy until I turned on all points. If I have center point selected, that's the only one that will grab focus. If the runner moves out of that center point, I will try to get whatever is behind the runner in focus and it will take the picture.

True, but each approach has its advantages. Multi-point is good when you have a pack of runners and you want to focus on those closest to you, but the disadvantage is that some signpost or weed in the foreground can ruin your shot if it gets picked as the focus point. When you have an open area and mostly single-file runners, it doestn't seem to matter very much either way.

Ziggy: that is potentially good news about the 40D, given that it's even less expensive than the 5d. That said, my next investment is probably going to be in glass -- most likely the 24 - 70 mm f/2.8L lense. It will be interesting to see how that high-end lense works with the XTi. If I see the same kinds of problems, then it will be pretty clear that I need a better body than the XTi.

evoryware

nobody wrote:

True, but each approach has its advantages. Multi-point is good when you have a pack of runners and you want to focus on those closest to you, but the disadvantage is that some signpost or weed in the foreground can ruin your shot if it gets picked as the focus point. When you have an open area and mostly single-file runners, it doestn't seem to matter very much either way.

Ziggy: that is potentially good news about the 40D, given that it's even less expensive than the 5d. That said, my next investment is probably going to be in glass -- most likely the 24 - 70 mm f/2.8L lense. It will be interesting to see how that high-end lense works with the XTi. If I see the same kinds of problems, then it will be pretty clear that I need a better body than the XTi.

I haven't had that happen on my 400 or 40, but I know for sure when I have one runner and the single point goes between his moving arms as I try to keep it pinned on his body or I slightly slip off of him as I'm holding the shutter for continous shots, I end up with improperly focused shots.

Awais Yaqub

Once i got opertunity to take action shots following technique gave me better results

Activate that sensor which is near to the face of the runner and keep it on the face while moving the cam. Like at above shot i would activate the top sensor out of 9 in vertical orientation of cam, while in AI Servo mode. I used nifty fifty and 400D got what i am saying ?

pilotdave

Awais Yaqub wrote:

Once i got opertunity to take action shots following technique gave me better results

Activate that sensor which is near to the face of the runner and keep it on the face while moving the cam. Like at above shot i would activate the top sensor out of 9 in vertical orientation of cam, while in AI Servo mode. I used nifty fifty and 400D got what i am saying ?

I have always believed that only the center focus point can be used with AI servo on a 400D. Is that not true?

Dave

Awais Yaqub

pilotdave wrote:

I have always believed that only the center focus point can be used with AI servo on a 400D. Is that not true?

Dave

You can use any of them

nobody

Awais Yaqub wrote:

You can use any of them

Thanks for the info. I had not thought of it before, but when shooting F/4 and zooming to fit a human body into the frame, the depth of field is not very much -- only 10 feet or so for the 70 - 100 mm lense. There is also a possibility that the arms moving at different speeds than the runner are causing the AI servo to miscalculate speed. I know it will be a bit harder to hold the high focus point on a person's face, but if it prevents the interference from the unsteady motion of the arms, it will be well worth it and I'll wind up with a lot more shots I can keep.

Awais Yaqub

nobody wrote:

Thanks for the info. I had not thought of it before, but when shooting F/4 and zooming to fit a human body into the frame, the depth of field is not very much -- only 10 feet or so for the 70 - 100 mm lense. There is also a possibility that the arms moving at different speeds than the runner are causing the AI servo to miscalculate speed. I know it will be a bit harder to hold the high focus point on a person's face, but if it prevents the interference from the unsteady motion of the arms, it will be well worth it and I'll wind up with a lot more shots I can keep.

Try it with that lens an let me know, i never shoot beyond 55mm
Contribute in the inovation and be part of history

evoryware

Awais Yaqub wrote:

You can use any of them

That's what I was saying.

pilotdave

Awais Yaqub wrote:

You can use any of them

Ok, I'm totally amazed. I use AI servo mode probably 99% of the time and for some reason I always thought that only the center point was active. I don't know how it's possible that I haven't tried changing autofocus points in the last 2 years in that mode, but I did it. Anyway, tried changing it and of course it works fine. Now I'm scared to find out what else I don't know...

Dave

nobody

Awais Yaqub wrote:

Try it with that lens an let me know, i never shoot beyond 55mm
Contribute in the inovation and be part of history

I did some experimentation with the different focal points today. I ran a 10k race, and then shot some of the runners still in it and those doing some "cool down" slow easy running. Using the Digital Rebel XTi and 70 - 200 mm f/4L lense, I shot one set with the head as the focal point (top point in vertical orientation) and another set with everything else the same (ISO 100, aperture f/5.6) except that all points were activated. The first thing I did was count the total images in each set. Then, I inspected every photo in full size and deleted any in which I thought the face was too badly out-of-focus to be usable. In all cases, I counted up the deleted files. For the single-point focus on the runner's head, my deletion rate was 15.4% and for the all-points focus, it was 25.4%. These galleries are here:

http://www.enturner.com/Experiments


From looking at my past galleries for sale where I used the middle of the body focus point, I can see that I deleted 10% of the initial photos and did not use another 30% (either due to poor focus or redundancy). Based on all this data, I believe there is a pretty solid case for utilizing the technique you described (single point focus on the head).

Probably, the f/5.6 aperture (instead of the f/4 that I typically use) helped somewhat as well. I don't know of a mathematical equation that describes it, but it looks like depth of field is roughly proportional to aperture. Another thing I have noticed is that when one photographs standing adult human beings with the intent of roughly filling the frame with them, then the depth of field does not change a great deal with focal length. For example, if I'm at 70 mm and shooting a runner 35 feet away, then it looks roughly the same as if I'm 100 feet away and using 200 mm focal length. In both cases, the depth of field is about 12 feet if I'm on f/4. Narrowing the aperture f/5.6 would increase it to about 17 feet. This is from:

http://www.dofmaster.com/dofjs.html

The only disadvantage I can see to the off-center focal point is the need to quickly switch it when there is a pack of competing runners approaching the finish line and I need to shoot in horizontal (landscape) orientation. If the leader just happens to be in the right position, I don't change it, but otherwise, it has to be done in order to prevent the leader from being out of focus. If this technique consistently reduces my deletion rate from 25% to 15%, then it's a small price to pay.

Also, there is a gallery I did using my new Tamron 28 - 75 mm f/2.8 lense. In my opinion, that lense is usable for the job, and produced plenty of usable photos, but I do not think the focal range is really appropriate for the job. With the 70 - 200, I have about 2.5 times as long to compose the shot. Both ends of the focal range are useful. With the shorter lense, the long end of the focal range is useful, but the short end is not because someone 10 feet away from me running a 5 or 6 or even 9 minute per mile pace is simply not going to be visible for very long.

pathfinder

Despite what folks think, depth of field is not altered by changing the focal length IF the subject is kept the same size at the image plane.

Whether shooting a 28mm lens or a 200mm lens, IF THE SUBJECT IS CONSISTENTLY THE SAME SIZE IN THE IMAGE, the dof will be the same with both lenses.

Obviously, a subject has to be a lot closer to the lens at 28mm than with a 200mm lens for the same image size on the sensor.

Tee Why

Just wondering if increasing ISO to 200 and doubling the shutter speed makes any difference. I notice that the shutter speeds were about 1/400 or faster so I don't think that's the issue but you never know.

nobody

Tee Why wrote:

Just wondering if increasing ISO to 200 and doubling the shutter speed makes any difference. I notice that the shutter speeds were about 1/400 or faster so I don't think that's the issue but you never know.

I have seen the same problem occur at faster shutter speeds (1/1000 or 1/2000 sec) with aperture f/4 and ISO 100. Thus, I think in most cases it's focus problem and not a problem with the lense or subject moving too fast.