PDA

View Full Version : Sigh - 40D focus issues!



Julian Robinson
01-15-2008, 05:59 PM
I know this is covered in many forums, but I have not seen a considered conclusion, and would appreciate input from bird-centric photographers.

Until recently I was happy photographing with my 20D and the 400/5.6 lens (Canon). 10 days ago my 20D died spectacularly (gasping non-stop, the shutter kept operating even when turned off). I was told it would cost more to repair than it was worth, and I couldn't last the one or two months it takes Canon Aust to repair things (I'm not in CPS obviously).

So I bought a 40D. I have been extremely disappointed with the image quality, not one of my shots looks 'sharp', out of 1800 bird shots tried in two days. I did some basic tests including focus-point and it seemed:

a) focus was all over the place, often in front or behind the target by 3-30mm over a 4m range using my 400/5.6 and a borrowed 100-400L. Never way off target, but enough to make the target point not perfectly sharp most times.

b) with the 100/2.8 lens the focus jittered or juddered continuously (on AI servo), or juddered momentarily (on One-shot focus) then took variably out-of-focus shots -- very out of focus.

c) the 400/5.6 is less sharp on this camera than the 100-400 is at 400, when focused correctly. Subjectively it also appears less sharp than it did on the 20D which is one source of confusion to me - seems to indicate that my current problem is more perception than reality since it is unlikely that the lens actually went bad at the same time as the camera! I can't compare with the 20D/100-400 combination since I didn't have that lens while the 20D was working.

I took the 40D back and got another, and it is in all respects the same. 1000 bird shots the other day, none 'seemed' sharp. I use RAW generally so it is not the jpeg sharpness setting, what is missing is the feather detail, the low-contrast details that no amount of sharpening will bring back.

I know this is a complex problem, and I am doing more tests today (including with another 100-400) but I was wondering if anyone has been through the same sort of thing and come up with a resolution? Of course I am not keen to send my camera plus lens to Canon because of their slow service time at peak bird season and during the time I had set aside for specially for bird photography. And I am not sure if it would achieve anything.

After tests today I am aiming to replace the body yet again, and try a third. This sounds ridiculous, but I had similar problems with a 24-105 lens ... tried 4 before finding one that seemed acceptable in IQ, and I have evidence to show that the first 3 were really not up to the mark.

Any suggestions, even for a plan of attack?

Maxis Gamez
01-15-2008, 06:26 PM
Can you post images?

Julian Robinson
01-16-2008, 10:28 AM
Hi Maxiz - Apparently I can only post one small image a day so it's hard to illustrate this kind of problem. Here's one taken today, as close as I usually manage to get to a bird. It's a full size crop of a jpg straight out of the camera with camera sharpening set to 3. The focus point was on the head, aiming for the eye. This is typical of resolution/crispness I am getting, it doesn't get better. The photo taking situation was near perfect being from my balcony with plenty of time, something to lean against, no haste and a stationary bird. Taken with Canon 100-400 at 400mm, f7. There is no camera movement at all in this shot.

What concerns me is:

- The eye and beak details are soft even though of high-ish contrast.

- The feather texture /plumage detail is almost absent.

- The background noise is evident already, before I do any PP sharpening. This was ISO 320 so I was expecting better.

- By the time I sharpen to show the eye etc in proper detail, the noise is outrageous, and it doesn't do anything really for the plumage which doesn't have much contrast or detail to start with.

Apologies if the attachment is missing, the process is not intuitive to me - uploading it before I send the message.

Thanks

Julian

James Prudente
01-16-2008, 12:09 PM
Julian,

What are the techs on this shot?

Jim

Julian Robinson
01-16-2008, 06:12 PM
James - Not sure what info you are after that I didn't mention, so below is the full EXIF (plus the thing it doesn't tell you is that IS was on). As I mentioned this is a full-size crop (i.e. not re-sized) 750 pixels wide (to meet the requirements of this forum) at the centre of the frame.

Filename : IMG_4227.JPG

Make : Canon
Model : Canon EOS 40D
ResolutionUnit : Inch
DateTime : 2008:01:16 14:11:43
ExposureTime : 1/320Sec
FNumber : F7.1
ExposureProgram : Aperture Priority
ISOSpeedRatings : 320
ShutterSpeedValue : 1/332Sec
ApertureValue : F7.0
ExposureBiasValue : EV0.3
Flash : Not fired(Compulsory)
FocalLength : 400.00(mm)
ColorSpace : sRGB
ExifImageWidth : 3888
ExifImageHeight : 2592
ExifInteroperabilityOffset : 8624
FocalPlaneXResolution : 3888000/876
FocalPlaneYResolution : 2592000/583
FocalPlaneResolutionUnit : Meter
CustomRendered : Normal process
ExposureMode : Auto
WhiteBalance : Auto
SceneCaptureType : Standard
MacroMode : Off
Self-timer : Off
Quality : Fine
FlashMode : Off
Drive Mode : Single-frame
Focus Mode : One-Shot
ImageSize : Large
Easy shooting mode : Manual
Contrast : Normal
Saturation : Normal
MeteringMode : Evaluative
ExposureProgram : Aperture Priority
Focal length of lens : 100-400(mm)
Flash Activity : Off
Color Tone : Normal
White Balance : Auto
Firmware version : FirmwareVersion1.0.4
Owner name :
Serial number : 2592-08394
Sharpness Intensity : 3
Parameters : Standard

Maxis Gamez
01-16-2008, 07:09 PM
HI Julian,

Based on that image, I have to agree with you, there is nothing in focus in that image and the noise is a large amount but you have to remember this is a jpg. What I would do is to do a test with a slow moving subject like a car in RAW and see what you get. Post that image again.

Bill McCrystyn
01-25-2008, 11:56 PM
Unlikely, but possible, it also doesn't tell me if you have your focus limiter set properly on the lens.

Bill McCrystyn
01-26-2008, 01:20 PM
c) the 400/5.6 is less sharp on this camera than the 100-400 is at 400, when focused correctly. Subjectively it also appears less sharp than it did on the 20D which is one source of confusion to me

Actually if you look at the MTF for the 1-4 @ 400 you will see it edges out the f/5.6 - 400, believe it or not.

Dan Elster
01-28-2008, 02:03 AM
I'm glad I saw this. Currently my 40D and 1-400mm lense are both at Canon for the SAME REASON. I've been very dissatisfied w/ my 40D - I upgraded from a 10D. I kept thinking it was my fault that images were slightly soft. But I've been doing this long enough to know when I "nail" a shot. If you want, send me an email and I'll keep you posted when I get my stuff back. The more I tested the stuff myself, the more confused i got. I've been using the 1-400 for a few years and know it takes a razor sharp image w/ a 10D. I'm quite sure there is somehting wrong. We aren't the only ones either, I'm finding out..

Alfred Forns
01-28-2008, 08:14 AM
Sorry to hear about the problems Might be something to the specific camera I would try returning for another and maybe do a quick test before you take it out of the store?

In the old times I used to test lenses before taken them out The variation from lens to lens was high Keep us posted

Robert Becker
01-28-2008, 12:14 PM
This thread is very interesting to me. I have been shooting with the original Digital Rebel (300D) along with the 100-400mm. Almost every shot I make is soft. I have posted two photos so far on this website in the "Wild and Free Bird" forum as well as many others on other websites. Most of the critiques have been that the images are too soft. So now I'm coming to the conclusion that perhaps the lens and camera body need to be calibrated together(?). I suppose it's possible that the quality control at Canon isn't what it should be and each camera has it own personality. I really want to upgrade soon to a "better" body, but I'm a little bit frightened of what I'm going to get out of the box.
I wonder if anyone else has experience any noticeable improvement after having Canon perform a calibration?

Julian Robinson
01-29-2008, 08:27 AM
Dan - we're not the only ones, but there's a lot who are happy with their 40D's! I must admit that all the photo's I've looked at exhibit some aspects of what makes me unhappy, that is, the loss of detail on low-contrast plumage. I have developed a few ideas after much investigating, head scratching and moaning on forums ...

- for sure the anti-aliasing filter on the 40D is 'softer' than on the 20D i.e. it has less artifacts but makes low detail, low contrast areas look like plastic.
- this effect is made much worse if there is even a small lack of focus.
- the central sensor on the 40D is pretty big (covers a large area) and can easily lock on to, or be influenced by, branches and head outlines and other high contrast things that you don't want it to see. One poster on another forum has greatly improved his shots by only using one of the outer 6 focus sensors, which seems strange but apparently works for him because they are smaller -- so you tend to get what you aim at. By contrast when you use the center sensor, aiming for the eye or head is tricky unless the bird fills almost the whole frame since the sensor will often see the back of the bird's head rather than the eye. Or, as I said, a branch or leaf.
- focusing with a 400mm lens on low contrast things like some birds is getting towards the limits of the technology and very often the camera doesn't get it quite right. For some reason, for me, the 40D centre sensor was much more variable than the 20D. I don't know why. So many of my bird and test photos with the 40D were just not exactly focused. Not consistently in front or behind, just not right.

Dan you did write to me before but here's the address again, I am really interested to hear what you find when you get yours back. jrsubsATinternode.on.net

Robert - I suspect you might just have a plain badly calibrated camera and/or lens. Most people report great improvement when they get their camera and/or lens calibrated, the popular strategy being to send them off to be matched together. Then again, I know from experience that operator inexperience can cause 1000's of soft shots, so it would be worth doing a decent test to separate possible gear problem from usage problems -- on a tripod, no IS (with your lens), looking at a detailed target of some kind. Some people use banknotes, newspapers, USAF targets etc. If you put a strong mark for the camera to focus on and tilt the target so it is not face-on to the camera, you can get an idea if it is focusing in front or behind or spot-on.

As well as focus problems, a faulty lens can also be plain ugly soft, no matter how well focused.

Alfred - I've tried 2 and they were the same as far as I could tell, so it seems likely to be a camera characteristic, or something I don't understand from a usage point of view, or a combination of small things as I list above. Or something else.

Bill -focus limiter was set correctly! The pigeon shot I showed in my post was actually pretty well correctly focused as you could see in other parts of the frame by the sharpness of the fence it was sitting on. But the fence was hard-edged and sharpened up ok, while parts of the bird were not, and didn't sharpen up at all. The bird was about 10 m away so the lens was not near the end of the focus range.

Maxis - I don't have the camera to do any more tests! But all my shots were consistently soft, and I could tell the one posted above wasn't movement affected by the sharp fence in the lower part of the image.

Here's a few more taken with the 40D, all of them looking not good to my eye... The Kestrel was standing clear of any interfering branches that might have confused the camera, but I took a dozen shots and they all look like this one - soft but noisy. The pigeons were large and close, but the plumage still looks wrong to me even though I sharpened to the limit then used neatimage to remove the introduced noise from the background (no neatimage on the birds, so no detail removed). The image shown below is considerably down-sized so much softness is hidden. The parrot shot is re-sized significantly too and still looks a mite soft.

http://www.flickr.com/photo_zoom.gne?id=2191580701&size=o Kestrel
http://www.flickr.com/photo_zoom.gne?id=2201472778&size=o Pigeons
http://www.flickr.com/photo_zoom.gne?id=2200678455&size=o Parrots

For comparison here's a few birds in flight I took with the 20D to show how good it was at focusing on relatively tricky subjects - some of them quite close and moving quite fast. These are all better than I could get on stationary subjects with the 40D. If you open a few of these in new windows side by side against the 40D examples, you'll get a good comparison.

http://www.flickr.com/photo_zoom.gne?id=251127842&size=o
http://www.flickr.com/photo_zoom.gne?id=479881383&size=o
http://www.flickr.com/photo_zoom.gne?id=1271457665&size=o
http://www.flickr.com/photo_zoom.gne?id=1482441843&size=o
http://www.flickr.com/photo_zoom.gne?id=376304945&size=o
http://www.flickr.com/photo_zoom.gne?id=483232499&size=o