10D focus test result

[Pekka]

Hi all,

I'm not a guy who has patience to do scientific equipment tests but today I took my 10D and tested all my lenses on tripod with mirror lock, cable release and two 550's slaved (roof bounce). One shot focus only with center focus point only. I took 10 shots on each aperture up to 2.8 and one f8 shot too from each lens.

Results were interesting:

- Canon 70-200/2.8L (@200mm) focuses rock steady and exactly on focus point each time. Same with 500D.
- Canon 35/2 fluctuates - some shots are ok, most front focus
- Canon 50/1.4 fluctuates and all are front focus
- Sigma 20/1.8 fluctuates and front focuses.
- Sigma 14/2.8 fluctuates and front focuses.

Aperture changes did not affect focusing.

Resetting the camera did not change focusing.

Manual focusing worked as expected.

All lenses were sharp as expected: there are no complaints in that area. 10D is not a soft camera - it is very capable of displaying extremely sharp detail and distinquishing lenses and apertures clearly. I was pleasantly surprised how contrasty and sharp 1.4 on my 50mm was and 35/2 was unbelievable sharp and crisp at f2.0!

What I mean about "fluctuation" (did not find a better word): each focus lock (via cable release shutter) was a bit different - looks like the focus points are not confined in what you see graphically in viewfinder but the camera looks for surrounding areas too, randomly. There seems to some fuzzy logic in action, and this is somehow related to front focusing (lens which did not fluctuate did not front focus either). You could also hear this on noisier lens motors - each time you pressed focus lock the lens would whirr a different sound - once in a while same, but at least 4 different ones.

To demonstrate, here's an animated gif of 50/1.5 at F1.4, 12 successive mirror-locked one-shot AF shots (it is a 1MB file, let it download fully to see the animation). Red circle is the focus point:

http://photography-on-the.net/10D/10D_F1_4_oneshot_focus.gif

Other interesting thing I noticed was that card bay halted writing as soon as focus lock was attempted. Odd.

I'm calling Canon Finland tomorrow to find out how they would handle finetuning my camera. I will let you know how it advances.

[mishkin]

I think USM in relatively new lenses (released 3-5 yrs ago max) has finer steps than in old lenses (especially first USMs).

My impressions:
70-200IS - no or very little fluctuations
16-35L - no or very little fluctuations
28-135 - little fluctuations
50/1.4 - large fluctuations


50/1.4 is very sharp lens, but its focus system is terrible.

[Pekka]

mishkin wrote:
I think USM in relatively new lenses (released 3-5 yrs ago max) has finer steps than in old lenses (especially first USMs).

My impressions:
70-200IS - no or very little fluctuations
16-35L - no or very little fluctuations
28-135 - little fluctuations
50/1.4 - large fluctuations


50/1.4 is very sharp lens, but its focus system is terrible.


Right, it could be just focusing motor technology showing design flaws. But I sort of would expect a bad lens motor to focus on same bad spot each time especially when camera is on tripod :)

[volleygerd]

Hello Pekka,
i am very interested in the answer from Canon. I will recieve my 10D in the middle of next week (app. 16./17. April) and will test, if she is front focus. Can You give me some hint, how i can test it quick and easy?
Kind regards to Finland from Germany,
Gerd

[Mickey]

I've found similar issues with my 10D. I've been going back and forth trying to figure out if it's my lens or the body that's causing this inconsistent focusing. Unfortunately I only have the one body and one lens since I recently moved up from a G2, so testing is difficult.

I've posted some samples here showing how much it can vary from shot to shot:
http://forums.dpreview.com/forums/read.asp?forum=1019&message=4784090
http://forums.dpreview.com/forums/read.asp?forum=1019&message=4791838

Is this the kind of variation you're seeing, or is this more extreme?

Michael

[hugodrax]

Weird. Makes no sense. did the D60 and D30 suffer this problem? I heard some people send the lens and body to get calibrated by canon. is there some kind of correction table that exists in the body or the lens itself?

[Roger_Cavanagh]

Pekka,

Your animated GIF illustrates the "issue" nicely. I haven't tested my 10D as systematically as you, but I'm fairly certain that I am experiencing similar behaviour, but with possibly greater variation.

Regards,

Roger

[dab63]

I got mine 10D 1 week ago, and I was warried about AF problems reading so much about it on different forums.
So the second thing I did (first one was deadpixels test) was to perform fredmiranda's reset procedure and take some shots.
I'm using a new 28-200 Canon lense (not USM); all the shots of same sequence have the same (good, IMHO) focus.
Here is one of that pictures http://www.casaborghi.it/canon_eos_10d.html plus some 100% crops before/after applying UnsharpMasking.
As a note, when I bought my 10D, I tested my Canon's Sigma lense and it didn't work (err 99); my local dealer confirmed that there is a firmware incompatibility between 10D body and older lenses, so maybe even the AF problem could be related to lenses' firmware functions.
Ciao

[hugodrax]

That could be true maybe a new lens with a 10D will not have this issue VS an older lens. anyone try this as an experiment.

[Mickey]

See this link for a list of compatible and non-compatible (requiring update) lenses for the 10D: http://www.sigmaphoto.com/Html/news/News.htm

I don't get any errors with my Sigma 28-70 2.8 EX since it's on the "compatible" list, but I still get the focusing inconsistencies.

[Cal Maier]

Pekka wrote:

Right, it could be just focusing motor technology showing design flaws. But I sort of would expect a bad lens motor to focus on same bad spot each time especially when camera is on tripod :)



I am experiencing a slightly different problem with my older 70-200 2.8 L. The focus two inches behind subject at 70mm consistantly and tack sharp at 100mm thru 200mm every time, regardless of f-stop. I have tried this lens with 2 other 10D's and found that one body was very similar to my own and the other body was consistantly front focusing so it was very good at 70mm, but all other focal lengths were approx. 1 to 2 inches in front of focal point.

My other lenses including the 28-70L and 15-30 Sigma are giving me acceptable auto focus on a consistant basis. My 50mm 2.5 macro seems to front focus by 1/2" in auto focus but this isn't a real problem as I normally focus this lens manually.

I would be curious to hear what Canon in Finland has to say about your problem Pekka, I plan on phoning Canon here in Canada and will let you know what they have to say. I have a feeling that my problem with the 70-200 is the age of the lens, but I guess we'll see what Canon has to say.

Just my $.02

Cal

[dab63]

I've just performed a simple test in 'every-day' condition:
1) I'm in my office, with a Cisco's blueprint on the wall behind me
2) I put my cam on pocket tripod on my desk
3) distance from the wall: 2 meters
4) I took four shots, without caring about WB, with same lense (Canon 28-200) at 4 focal lenghts: 40mm, 57mm, 100mm and 200mm.
Here are results (cropped 100% from 6M image):
http://www.casaborghi.it/crop_at40mm.jpg
http://www.casaborghi.it/crop_at57mm.jpg
http://www.casaborghi.it/crop_at100mm.jpg
http://www.casaborghi.it/crop_at200mm.jpg
Note that blue boxes have less than 2cm width.
IMHO focus is quite fine also in this bad conditions, but I'd really appreciate your opinions.
Ciao

[cwhitla]

As a D30 owner for two years, I have been very interested in the 10D since it promised to resolve the low light focus issues of the D30. With the flurry of 10D focusing problems being reported, I decided to test my D30. I found that it's focus also tended to vary and it was frequently front focusing. I also found that, as Pekka suggests, areas around the focus point do seem to contribute to some focussing issues. I got differences in focus results depending on whether I placed other objects closer or farther from the center one I was focussing on. All testing with central point only and a tripod.
I am coming to the conclusion that the Canon autofocus simply has a tolerance that, while varying from camera to camera, will never be "tack sharp" 100% consistantly under all conditions and all the time - at least at a price I am willing to pay. I am still on the fence about buying a 10D and whether it will really focus any better or worse for me in real world circumstances.

[Pekka]

I called Canon repair today (actually Canon Finland does not have their own repair shop any more, they have passed the function to third party). I managed to get a fixed date (22nd) when I take the camera there and they will check it out. They did not know of any 10D problems as they have not repaired/handled any 10D's yet (other than seen the demo one) but they got the repair manuals etc a few days ago. For not repairing a 10D is understandable because to my knowlegde only about 30 has been available in Finland so far.

This gives me time to test things more. Based on comments I had I will do same tests next week in different lighting conditions (indoor low, indoor bright, outdoor low, outdoor bright) and do more real photography too of course.

Nevertheless I have then an opportunity to talk with the guy to has the repair manual and skills and decide what can be made to improve the issue if there is something to be done.

[Pekka]

hugodrax wrote:
Weird. Makes no sense. did the D60 and D30 suffer this problem? I heard some people send the lens and body to get calibrated by canon. is there some kind of correction table that exists in the body or the lens itself?

For example AFAIK D30 has a separate timer adjustment for each shutter speed so there must be tons of other adjustments hidden there - and more on 10D.

[Cal Maier]

Hi Pekka,

I called Canon Canada Service in Calgary this AM and talked to one of their service techs about my particular problem with backfocus at 70mm(see post above.)

He assured me that he hasn't heard of this particular problem with the 10D or a 70-200L lens, but was going to contact Canon in Japan and would probably have an answer for me within a couple of days. He told me that they have yet to perform any service on the new 10D and the only firmware update so far was to fix the Chinese language menu as some things were mis-spelled or something.

I too, plan on performing some more tests while I wait for his call. I was advised that I may have to send the camera or the lens, or both in for service, but not until Canon Japan got back to him.

Just thought you may want to know.

Cal

[Mike Stiles]

Pekka,

It seems that my 10D is exhibiting the exact same behavior as yours. With the 70-200L @ 200mm, the focus is dead on. The camera front focuses and progressively gets worse as I move down in focal length through the, 28-70L, 50 1.4, and Sigma 15-30.

I have both a D30 and D60, and ordered the D60 service manual to fix a slight back focus problem that I had with it. After making a very small adjustment with the with the software calibration routine on the service manual, my D60 is now performing great.

The D60 back focus problem that I had was relatively consistent across all of my lenses. This 10D that I have goes from being perfect at 200mm to focusing 10cm in front of the target when at 28mm from a distance of only 2.5m away.

I am eagerly awaiting your results with the repair on your camera. It sounds like you have an open communication channel with the folks that will be doing the repair, and hopefully you will be able to provide some insight as to what is going on. I would love to have some information to send in with my camera when it goes in for repair.

Mike Stiles

[Ralf Jannke]

Hi Pekka,

I have not counted the number of pages/postings, but in the German www.dforum.de you have a huge wave about (un)sharp pics taken with the new 10 D. I red your last posting with interest. I think that is nothing really new! I had the same feelings with my 1,4/50 mm USM or former 1,8/50 EF. But in my opinion that has nothing to do with the body! I found the same miss-focusing using the 50 millimeters on the 1,3 MP Kodak/EOS DCS 3, the 2 MP EOS DCS 520 (D 2000) and my D 30, D 60 and 10 D. The 2,8/70-200 works best for basketball on ALL bodies - the 50 mm is frustrating. The most sharp shots had been taken with the L Zoom. Even the 1,8/85 mm or 2/100 mm USM is missing more than the zoom.

Bets regards from Bonn/Germany

Ralf Jannke

[P.C. Plod]

You can't adjust each shutter speed individually on the D30. One adjustment applies to all speeds. I have the service CD.

[rodbunn]

You guys at Canon should be looking into the focusing problem! For the cost of this 10D body EVERY LENSE and EVERY SHOT should be clear (unless the user, not the "body" has problems).....

I would have bought a Nikon if I'd have known not all my lenses would work and that each lens was going to focus differently with the 10D.

CANON HELP!!!!! Are you listening ??????

[rodbunn]

Hey, My Sigma 28-70 2.8 lense was not taking clear pics with my 10d and I wasn't getting an ERROR 99. I tested it out and the Apateur wasn't changing even though the camera said it was. I called Sigma and they said that yes, that model won't work the apateur even though it says it is..... They are going to re-chip it for free but MAYBE this is why some people think the pics are SOFT ?????

[beech1]

I have had my 10D for a week and seem to be getting good results. This thread made me test the camera and lenses. I tested my 10D using a:

Tamron 20-40 2.7 -3.5
Canon 28-70L
Canon 70-200L

Focus and exposure was dead on. I checked exposure with a Gossen Luna Pro. All lighting was available light, I shot from wide open to f8. Every shot was on the mark in focus and exposure.

Don Goings

[Torben]

I think your challenging the auto focus too much.
Inside your autofocus frame you're having several contrasty edges at different distances. This may confuse the AF algorithm (like focusing into thin branches of a tree).
Try placing a target perpendicular to shooting direction and make the AF area cover this completely.
Not that I doubt your 10D having problems ...

[Mickey]

rodbunn wrote:
Hey, My Sigma 28-70 2.8 lense was not taking clear pics with my 10d and I wasn't getting an ERROR 99. I tested it out and the Apateur wasn't changing even though the camera said it was. I called Sigma and they said that yes, that model won't work the apateur even though it says it is..... They are going to re-chip it for free but MAYBE this is why some people think the pics are SOFT ?????

Which version of the 28-70 2.8 do you have? I have the 28-70 2.8 EX DF and mine seems to work fine (aperture stops down when shooting and when using the DOF preview button) but I seem to be getting rather inconsistent focus. I did some test portraits last night and one was totally out of focus, some appear to be in perfect focus but the majority of the shots seem a couple of inches front or back focused (often I'd see that the hair half-way back is perfectly sharp but the eyes and mouth were soft). I need to reshoot since this might be user error and I or my subject may have moved slightly between focus locking and shooting. I used to do the focus and recompose technique a lot with my G2 but with this narrow DOF of a 70mm lens at f/2.8 and f/4 I may have to be quicker about it and more careful with what's in the AF sensor area.

Michael

[88 monte]

Mine seems to be Dead on
http://mywebpages.comcast.net/jmoore411/Kid.jpg

[rodbunn]

HI, I have the Sigma 28-70 2.8 (the 77mm one). I don't remember which model that is but Sigma said it wouldn't work with the 10D until they do a chip changeout. The camera worked with the lense but there wasn't any DOF in the pics. I could change the f stop in the camera and shoot pics but when I looked at the pics it looked like it was stuck on f2.8.....

Thanks, Rod