Focus Test Photo Please Look

I have been trying to figure out if my 1000D actually has a back focus issue or if it is just me. Are these focus tests actually proper? I usually shoot RAW with shutter priority. I almost always use center single point AF. Sometimes my photos are plagued with the ears or earring of a person in a portrait will come out tack sharp and the face is slightly out of focus. It seems to me is that this occurs when the ears, earrings or back ground are more contrasting than that of the face. Generally I do not experience these problems at a higher focal length when for example the face comprises 75% of the picture, but when there is a mix of 50% subject to background ratio I do.

My camera is in warranty through October. I was thinking of sending it in to tell them to calibrate with my lens, which by the way is the Canon 28-105mm f/3.5-4.5 II USM. I don't want to say I have a back focus issue if I really don't, and they simply bring the focus more forward. Here is the test image, let me know what you think. I used Aperture Priority set it wide open, and I believe the focal length was at 80mm. Thanks

Looks ok to me. The 2mm behind the focus point is more clear than the 2mm in front however... so maybe a 1mm backfocus?

I'm no expert, but it looks ok.

dobeonguard said:

Are these focus tests actually proper?

Click to expand...

If your tests were conducted according to the instructions given with the target, and the shot you have shown is typical of a large number of shots taken under the test conditions, then yes.

dobeonguard said:

My camera is in warranty through October. I was thinking of sending it in to tell them to calibrate with my lens, which by the way is the Canon 28-105mm f/3.5-4.5 II USM. I don't want to say I have a back focus issue if I really don't, and they simply bring the focus more forward.

Click to expand...

They don't just "bring it forward", they use a similar target to fine-tune the focus. You might find this interesting - http://forums.dpreview.com/forums/readflat.asp?forum=1031&message=31555078

dobeonguard said:

Here is the test image, let me know what you think.

Click to expand...

It definitely shows a back focus. See how the front "10mm" is about as blurred as the back "20mm", or the front "2mm" and the back "6mm", on both sides.

But the image you provided doesn't show that the subject is outside the DOF. Canon says their phase detect (PD) AF system should put the subject "within the DOF" for lenses like yours. There is no guarantee that the subject will be bang in the centre of the DOF. So your system is performing according to specification under the test conditions.

Can you set up a test situation like the ear-focussed portraits, but using a tripod and a static subject? It's vital to eliminate the possibility that it is just you.

This looks OK to me. The DOF is always greater beyond the focus point than it is in front of it.

George

gscotten said:

The DOF is always greater beyond the focus point than it is in front of it.

Click to expand...

No, up close, DOF is symmetrical. It gets asymmetrical at "long" camera to subject distances (many times the focal length).

You might try repeating the test with Live view and see if it is the lens or the AF system.

However as pointed out it looks as if it is possibly within acceptable limits and not a problem depending on what you are shooting in the real world.

Here is an informative post in a thread where I was doing the same test:

http://forums.dpreview.com/forums/read.asp?forum=1031&message=31396170

I repeated the test, this time with the kit lens 18-55mm IS. I shot this at 18mm 3.5 aperture.

Please respond with your opinion. Also, do I need to contact Canon first before sending it in or do I just send it? And how long is the turn around usually? Thanks

Your first photos the focus looks fine.

The 2nd photo does not look anything close to the 45' that is required for a proper test..

Any angle greater or less than the 45' can throw off the results making it front or back focus.

Go setup your camera on a tripod and shoot wide open at different focal length. Place a box in front and horizontal to the camera and focus on that box. Also place another box to the right and left of the center box but equally spaced from the center box.

The center box should be in focus with each box equally in blur.

This is how I test my lens, I also use the focus chart also as a reference but double check it with the box method. I do have a inclinometer that I use for satellite installs that I use to check my camera angle.

Good luck.
--
Canon 5D II, 7D, 50D, & Fuji 31fd

AnthonyL said:

You might try repeating the test with Live view and see if it is the lens or the AF system.

Click to expand...

How could that tell you that the lens is at fault?

dobeonguard said:

I repeated the test, this time with the kit lens 18-55mm IS. I shot this at 18mm 3.5 aperture.

Please respond with your opinion.

Click to expand...

Yes, it's back-focussed, but no, I can't say for sure that Canon would judge it to be out of specification.

dobeonguard said:

Also, do I need to contact Canon first before sending it in or do I just send it?

Click to expand...

Call them and get a job number (or whatever they call it), before you send anything.

dobeonguard said:

And how long is the turn around usually?

Click to expand...

Two to four weeks seems common.

Yaamon said:

The 2nd photo does not look anything close to the 45' that is required for a proper test..

Any angle greater or less than the 45' can throw off the results making it front or back focus.

Click to expand...

How does that work?

WilbaW said:

Yaamon said:

The 2nd photo does not look anything close to the 45' that is required for a proper test..

Any angle greater or less than the 45' can throw off the results making it front or back focus.

Click to expand...

How does that work?

Click to expand...

WilbaW if you setup your camera on a tripod and then use an inclinometer and measure the camera angle to the focus chart at 45' and take the picture you can compare the results if you increase and decrease the angle from 45' and it will show a slight front or back focus compared to shotting dead on at 45'.

When I sent in my tamron lens to get calibrated with my body, the tech left the test photos on the CF card.

What I saw from his photos is that he shoot the camera straight on but he has a focus type chart on a 45' elevation to the camera.

--
Canon 5D II, 7D, 50D, & Fuji 31fd

Yaamon said:

WilbaW said:

Yaamon said:

The 2nd photo does not look anything close to the 45' that is required for a proper test..

Any angle greater or less than the 45' can throw off the results making it front or back focus.

Click to expand...

How does that work?

Click to expand...

WilbaW if you setup your camera on a tripod and then use an inclinometer and measure the camera angle to the focus chart at 45' and take the picture you can compare the results if you increase and decrease the angle from 45' and it will show a slight front or back focus compared to shotting dead on at 45'.

When I sent in my tamron lens to get calibrated with my body, the tech left the test photos on the CF card.

What I saw from his photos is that he shoot the camera straight on but he has a focus type chart on a 45' elevation to the camera.

Click to expand...

Yes, the best focus test rigs have a perpendicular target near a slanted scale. This is what they used at the Canon service centre in Sydney to calibrate my gear -



So you're saying that a target at 90° (like the above or Tamron's equivalent) will allow the AF system to work properly, and so will a target like Tim Jackson's at exactly 45°, but that you will get focus errors at less than 45°, and between 45° and 90°. So there's something magical about how a thin black line looks to a perpendicular AF sensor when the white background is at 45° or 90°?

Tim Jackson's chart is designed to be used at 45° so that the scale correctly indicates distance along the lens axis. If you deviate from that angle the indicated distance will not be correct, but according to any theory of how AF works on such a target, and in my experience of doing such tests, the angle doesn't make any difference to where the system will focus. (Where it would break down is at such a shallow angle that the system picks up on foreground texture in the paper rather than the black line, but that's not about the angle itself.)

Can you share some images that show the effect you're seeing? Shots at 90° and 45° would show good focus, and shots in the range of 60-75° would show opposite errors to shots in the range of 15-30°?

Wilba, I don't have the angle photos at 45' and at different angles on hand as I deleted them after the tests. I have on my pbase just a reg 45' shot of my lens under lens test.

If you like over the weekend I can take some photos to show the difference.

This what I mean about my box tests. The lipton boxes are exactly 1-1/8" difference from the focus chart box.

I would make sure that at all focal length that the center box has the best focus. Full exif is there to show 0ma.

I the lens only need a + or - 3 I wont send it in for calibration, but any more than that definitely. The reason is that my 2 young kids are taking photography in high school this Sept and was thinking of buying them a T2i to use or hopefully if Canon releases a 60D.

Since the T2i has no ma the lens would need all my lens to be dead on calibrated. I'm picky with focus. Where I point the camera to focus I want it to. If they are any focus errors I want it to be a user issue and not hardware.





--
Canon 5D II, 7D, 50D, & Fuji 31fd

Yaamon said:

Wilba, I don't have the angle photos at 45' and at different angles on hand as I deleted them after the tests. I have on my pbase just a reg 45' shot of my lens under lens test.

If you like over the weekend I can take some photos to show the difference.

Click to expand...

If you like, sure, please do. I'm not going to accept your claim without evidence. :-D