Focusing accuracy USM vs. regular

[torwinen]

I just bought 10D a couple of weeks ago and a couple of lenses: 50mm f1.8 and 35mm f2 just to mention. I have mainly photographed my children indoors with my new camera using Metz 54 flash. I have had several problems with focusing accuracy: Getting a good focus (even with stationary objects) is pure lottery if f is 2.8 or wider with the given lenses (a lot of problems even with f4 too). Camera gives indication that the correct focus has been achieved but afterwards looking at the pictures it shows that sometimes the subject is within the DOF and sometimes out of DOF (front of back focus, more often front than back).

Now, my question: Is it because 10Ds focusing sensor is not accurate enough or is it that my lenses (no USM) are not accurate enough? Maybe it is the combination of lighting+lens+body.
Can I get better results with USM lenses?

Marko

[scottbergerphoto]

I wouldn't think your focusing problem has anything to do with the kind of motor your lens uses. You are using very wide apertures, which give a very small depth of field. Since depth of field is 1/3 in front and 2/3 back, it isn't surprising that objects look a little back focused. That said:
Try your shots again using a tripod and a shutter speed of at least 1/60th of a second for the 50mm lens. Autofocus should be set to One Shot. Make sure your subject is completely stationary. Try manual focus and auto focus. If your manual focus is better then your auto focus do a focus test using one of the test pages available on the net. Do a search for "Focus
Test" and you will find one.
You also mentioned that you are using a Metz flash. The camera may be having trouble with autofocus in dim light. Does your flash have an autofocus assist light? Have you selected the center autofocus point? Many 3rd party flashes can only use their assist light if the center focus point is chosen.
Let us know what happens.
Scott

[defordphoto]

It's probably that you just need to get used to the camera. Take your time. Feel free to post samples here so we can see what's going on.

[CyberDyneSystems]

One of the things I HAVE noticed ONLY with my 50 f/1.8 is that sometimes my 10D will say it is ready to shoot (ei that the lens is focused) when in fact it isn't quite right yet.. if I continue to hold the shutter.. it will often audibly "tweek" the focus for a few seconds more.. intermitantly.. I assume of I did not allow it to finish (there are pauses so it would be easy to do so) that the image would not be in focus.

I only seem to have noticed this with my 5- f/1.8 which is my only Canon non USM lens.

p.s. this happens indoors in low light.

[maderito]

I have a 50mm/1.4 lens and frequently use it for available light shots. The wide apertures are seductive, but the resulting shots, in my experience, are often quite soft. I can never figure if it’s the lens characteristics or mis-focus or both. I actually LIKE the soft effect – but I am curious about the cause.

Working the with a depth of field calculator (http://dfleming.ameranet.com/dofjs.html) I come up with the following:

50 mm lens on the 10D - shooting at 4 feet
Aperture, depth of field in inches:
f1.4, 1.2"
f2.0, 1.68"
f2.8, 2.4"
f4.0, 3.48"
f5.6, 4.8"

My point is that the wide apertures have VERY small depth of field, especially when working in close. The lens is, of course, inherently softer at these apertures. Moreover, who knows what part of the subject the camera is actually focusing on. Even if you target a subject’s eye, it is not a flat target. Finally, you typically are working at relatively slow shutter speeds (1/30, 1/60). Add all these together (before considering front/back focusing issues) and you still have an image that may appear soft or unfocused.

Even though you are using a flash (which may “freeze” the subject), the shallow DOF and lens softness at wide apertures may be more significant factors than you realize.

[robertwgross]

Autofocus mechanisms can get easily fooled if there is a very high contrast object close to the primary subject and focus point, especially if the primary subject has low contrast.

I was shooting an informal family portrait of six adults. Depending on the autofocus mode, the camera kept going after the man with the brightly striped shirt, and it kept avoiding the woman with the white dress. Finally, I had to "fix it" on the one correct focus point.

I don't see that the type of focus motor in the lens has too much to do with the accuracy.

---Bob Gross---

[scottbergerphoto]

maderito wrote:
I have a 50mm/1.4 lens and frequently use it for available light shots. The wide apertures are seductive, but the resulting shots, in my experience, are often quite soft. I can never figure if it’s the lens characteristics or mis-focus or both. I actually LIKE the soft effect – but I am curious about the cause.

Working the with a depth of field calculator (http://dfleming.ameranet.com/dofjs.html) I come up with the following:

50 mm lens on the 10D - shooting at 4 feet
Aperture, depth of field in inches:
f1.4, 1.2"
f2.0, 1.68"
f2.8, 2.4"
f4.0, 3.48"
f5.6, 4.8"

My point is that the wide apertures have VERY small depth of field, especially when working in close. The lens is, of course, inherently softer at these apertures. Moreover, who knows what part of the subject the camera is actually focusing on. Even if you target a subject’s eye, it is not a flat target. Finally, you typically are working at relatively slow shutter speeds (1/30, 1/60). Add all these together (before considering front/back focusing issues) and you still have an image that may appear soft or unfocused.

Even though you are using a flash (which may “freeze” the subject), the shallow DOF and lens softness at wide apertures may be more significant factors than you realize.
Excellent post. The calculated DOF numbers really make the point
Thanks,
Scott

[torwinen]

scottbergerphoto wrote:
I wouldn't think your focusing problem has anything to do with the kind of motor your lens uses. You are using very wide apertures, which give a very small depth of field. Since depth of field is 1/3 in front and 2/3 back, it isn't surprising that objects look a little back focused. That said:
Try your shots again using a tripod and a shutter speed of at least 1/60th of a second for the 50mm lens. Autofocus should be set to One Shot. Make sure your subject is completely stationary. Try manual focus and auto focus. If your manual focus is better then your auto focus do a focus test using one of the test pages available on the net. Do a search for "Focus
Test" and you will find one.
You also mentioned that you are using a Metz flash. The camera may be having trouble with autofocus in dim light. Does your flash have an autofocus assist light? Have you selected the center autofocus point? Many 3rd party flashes can only use their assist light if the center focus point is chosen.
Let us know what happens.
Scott


I use assist light, center point. Metzs longest exposure time is 1/200. With my previous camera (Oly E-20) I managed to get very sharp pictures (from stationary objects) even with 0.5s exposure times :-) With 50mm I tend to use 1/60 f2.8.

I tried "the ruler test" yesterday with 50mm 1.8 indoors using my Metz. I dont seem to get 1/3 2/3 DOF at all; The shots that are in focus are something like 3/4 front 1/4 back. Most of the shots come out so that the DOF is actually ending before focus point. My 35mm f2 hehaves better: The portion of in focus shots is bigger.

I think I will have my camera checked, just in case there is something wrong with it.

Marko

[torwinen]

CyberDyneSystems wrote:
One of the things I HAVE noticed ONLY with my 50 f/1.8 is that sometimes my 10D will say it is ready to shoot (ei that the lens is focused) when in fact it isn't quite right yet.. if I continue to hold the shutter.. it will often audibly "tweek" the focus for a few seconds more.. intermitantly.. I assume of I did not allow it to finish (there are pauses so it would be easy to do so) that the image would not be in focus.

I only seem to have noticed this with my 5- f/1.8 which is my only Canon non USM lens.

p.s. this happens indoors in low light.

If that is true, the lens is unuseable in low light unless you are photographing non living objects :-)

Marko

[torwinen]

maderito wrote:
I have a 50mm/1.4 lens and frequently use it for available light shots. The wide apertures are seductive, but the resulting shots, in my experience, are often quite soft. I can never figure if it’s the lens characteristics or mis-focus or both. I actually LIKE the soft effect – but I am curious about the cause.

Working the with a depth of field calculator (http://dfleming.ameranet.com/dofjs.html) I come up with the following:

50 mm lens on the 10D - shooting at 4 feet
Aperture, depth of field in inches:
f1.4, 1.2"
f2.0, 1.68"
f2.8, 2.4"
f4.0, 3.48"
f5.6, 4.8"

My point is that the wide apertures have VERY small depth of field, especially when working in close. The lens is, of course, inherently softer at these apertures. Moreover, who knows what part of the subject the camera is actually focusing on. Even if you target a subject’s eye, it is not a flat target. Finally, you typically are working at relatively slow shutter speeds (1/30, 1/60). Add all these together (before considering front/back focusing issues) and you still have an image that may appear soft or unfocused.

Even though you are using a flash (which may “freeze” the subject), the shallow DOF and lens softness at wide apertures may be more significant factors than you realize.

I know when my shot is blurred because of the lens softness or subject movement. I just get shots where I have focused to eye and the result is that the eye is not sharp but the DOF ends before the focus point. In some cases the reason might be subject movement, camera movement etc. but as get similar results very often, It makes me wonder..

Marko

[PPi-]

I don't know if this helps any but I recently saw interesting discussion about 'fix' for focusing problem. The original link is here: http://forums.dpreview.com/forums/read.asp?forum=1031&message=6522286

It have also been shortly discussed in digicamera.net but few people have reported that everything have began work smoothly after reseting the camera. Least it doesn't hurt to try?

[robertwgross]

I would run the ruler test on your camera. That should allow you to eliminate the possibility of the body having a front-focus or rear-focus problem (fixable by Canon).

---Bob Gross---

[torwinen]

ppi- wrote:
I don't know if this helps any but I recently saw interesting discussion about 'fix' for focusing problem. The original link is here: http://forums.dpreview.com/forums/read.asp?forum=1031&message=6522286

It have also been shortly discussed in digicamera.net but few people have reported that everything have began work smoothly after reseting the camera. Least it doesn't hurt to try?

I will try that :-)

I'll perform a proper ruler test propably tomorrow. We shall see what will be the results.

M.

[slin100]

CyberDyneSystems wrote:
One of the things I HAVE noticed ONLY with my 50 f/1.8 is that sometimes my 10D will say it is ready to shoot (ei that the lens is focused) when in fact it isn't quite right yet.. if I continue to hold the shutter.. it will often audibly "tweek" the focus for a few seconds more.. intermitantly.
You must have using AI Focus. The behavior you described can't happen with One Shot AF.

torwinen wrote:
I'll perform a proper ruler test propably tomorrow.
Be sure to do the test properly. Do not focus on the ruler itself because it will be impossible to know which line on the ruler the camera is focusing on.
The ruler should be next to a target of high-contrast. I believe Canon recommends a magazine cover perpendicular to the camera, not tilted.

[RichardtheSane]

slin100 wrote:

You must have using AI Focus. The behavior you described can't happen with One Shot AF.


Oh but it can, and does with my 50mm 1.8 lens too.
You shouldn't be so quick to dismiss the facts posted because it shouldn't happen. Note I uses shouldn't happen instead of can't

[slin100]

RichardtheSane wrote:
[Oh but it can, and does with my 50mm 1.8 lens too.
You shouldn't be so quick to dismiss the facts posted because it shouldn't happen. Note I uses shouldn't happen instead of can't
Just to make things clear, you say you've seen your 50mm lens continue to focus after the camera has indicated focus lock (i.e. focus lock indicator in viewfinder and, if enabled, the audible focus lock beep) and the AF mode was set to One Shot?

I don't what to say except that that's exactly what I might expect and I've seen it happen when the AF mode was AI Focus, but not One Shot.

[torwinen]

I performed yesterday some tests. I draw a thick line above the ruler so that camera is not distracted by the ruler itself. Tried two lenses: 35mm f2 and 50mm f1.8. Camera in one shot mode, center focus point. Tripod used in both cases. Camera to target distance, 1m (3 feet).

Here are the SAD results:

Room lighting (50mm@1.8, 1/80, ISO1600)
50mm@2.8
in focus / not in focus
3 / 7
35mm@2.0
in focus / not in focus
6 / 4

Better lighting (50mm@2.0, 1/800, ISO100)
50mm@2.0
in focus / not in focus
3 / 7
35mm@2.0
in focus / not in focus
6 / 4

In general, both lenses are front focusing. I dont have any camera shake because I can always see where the real DOF is.

Marko

[robertwgross]

I'm not sure how to interpret these numbers, since I think you did the test differently from what I expected.

I once checked my own camera for front- or back-focus using a 30-degree tilted yardstick, and letting the center focus square work on one exact mark on the stick. All of my images came out with the exact mark perfectly in focus, so mine didn't have any problem with normal lighting.

---Bob Gross---

[Daytripper]

CDS/RTS,

When you say the lens continues to focus after the 'ready to shoot' indication, are you referring to the red focus point box lighting up or the green focus confirmation LED (and beep)?

The red focus point boxes will flash before focus confirmation if you have a single focus point selected....

Cheers,
Pete