I have both a 5D (18 months) and 40D (2.5 months) and mainly shoot landscapes, macro and small products. First, the AF systems of both cameras are excellent and blow the doors off my old 10D and 20D--night and day difference.
In most shooting I can't tell a diff between the 5D and 40D AF. Both are peppy and reliable under most conditions. However the 40D does begin to show it's cross sensor advantage when you manually select AF points in low light/low contrast. Oddly the 5D seems slightly faster under the same conditions, but misses more often. Not a huge diff but having all crosses is a real advantage. Using AF from a Speedlite is an easy solution if it's too dark to use outer points (with flash disabled). Otherwise you may have to resort to a slight tilt of the camera or locking on a point of contrast.
I've always preferred manually selecting AF points because lock AF-recompose often misses due to shallow DOF at close distances and/or large apertures. What I'd really like is the choice of selecting AF points with ECF. Loved it on my Elan 7NE and EOS 3. The joystick thingie is chunky and slow.
I must admit I have seen some bad fousing anomalies with teh 40D. This is my second copy, the first had a nagging err99 issue that could not be explained or corrected with all of the "usual" fix's. I have been noticing this but wrote it off to user error.... missed that one and such. It was when I was getting OOF shots on a tripod w/ studio lights and subjects standing still that got me wondering. Yesterday I really took notice of it and started to dig.
Here is the issue:
Manual mode. Auto focus. ISO 400. 580EXII flash. Relatively staionary subjects.
The focus is simply wrong. The focus point is on, the area focused is off. Lets look at some images: (these are screen captures from Canons Zoom Browser EX showing shotting data and AF points.
I have been seeing this more and more (looking for it now) and it is causing me to lose confidence in my gear. I see this with different lenses under different conditions.
With the last one at least, it's clear what went wrong - as has been noted many times, the focus sensors are bigger than the red boxes. Here the camera focused on teacher's leg (or perhaps the floorboard?) instead of the boy's sweatshirt. Higher contrast - the focus jumps. As expected.
I don't have a convenient explanation for the second shot however - that one baffles me.
I didn't have the 40D for too long, but I have been quite impressed with the AI servo when I tried it with my 35/2, yes, that buzzy and not exactly swift arc form drive motor.
BarnDog,
With all due respect, clearly the misfocusing in some of your prints is the operator error, or wrong focusing habits, and not the camera's fault. The first image - the focus square is on BOTH the nose and the black subject behind. The camera decided to focus on the face - lucky you . However, the last image again features the same error -the focus square is both on the shirt AND the teacher's leg behind. This time the camera decided to focus on the leg. Try to make it easier for the camera next time. Pick a contrast are on the same subject
gfiksel wrote:
BarnDog,
With all due respect, clearly the misfocusing in some of your prints is the operator error, or wrong focusing habits, and not the camera's fault. The first image - the focus square is on BOTH the nose and the black subject behind. The camera decided to focus on the face - lucky you . However, the last image again features the same error -the focus square is both on the shirt AND the teacher's leg behind. This time the camera decided to focus on the leg. Try to make it easier for the camera next time. Pick a contrast are on the same subject...Show more →
Yeah, I'm with ya on that one. It does seem the focus area is large. I have noticed more of this, whether my bad learned habits or not, with the 40D vs the 30D or 20D.
One of things that does baffle me is when I get the focus lock beep, snap the shot and find the frame out of focus like it never locked as in shot #2. I have seen this shooting portraits in a studio setting as well. I shoot three frames, a pause in between each and the middle one is OOF?
Alan321 wrote:
The trouble here is that focus speed is an important component of focus reliability. There's no point being more accurate if it always takes longer to achieve than you are willing and able to give it. The result will be poorly focused shots. Focus reliabilty is a useful balance of accuracy, speed and consistency or repeatability.
In my experience the 40D AF is unreliable or at least inconsistent. Too often it grabs the wrong subject even when a single AF sensor is specified and is on target and the target is adequate in terms of contrast. I suspect an AF algorithm defect in which the extra long sensor elements for detecting a highly-defocused condition are used when they should not be. If it isn't that then there must be a technical fault, and yet Canon reckon it is performing within specs. Those specs must be pretty loose.
I don't have personal experience of the 5D needed to address your other questions.
There must be a lot of unit-to-unit variation because my copy (and that of another 40D owner I know) feel very responsive and consistently nail the subject (moving ducks and a variety of stationary objects as well). I'm guessing you have a bad one; it happens but knowing it happens doesn't make anybody feel any better. If it isn't too late for an exchange I'd send the body to Canon.
I must be lucky as well - I have two 40Ds and both just nail AF. Much, much better than all the XXDs I had before and my 5D. And when it misses I know it's my error.
jkurkjia wrote:
There must be a lot of unit-to-unit variation
And a lot of user-to-user variation - the vast majoriity of these issues (though not Alan's, I suspect) are usually pinned down to user error or unreasonable expectations...
I have to say that initially the difference in how the 40D performed compared to my 30D had me wondering if I was going to be one of those poorly-focusing 40D owners that covered the forums in the first month or two after its release but - funny thing! - as I've used it more and more, it seems to have got better and better..!
The fact is, I was the issue: now that I'm used to it, I reckon I've instinctively adapted to the camera and how it works, and now - hand on heart - I can't imagine how its AF could be better. I'm routinely getting sharper, more detailed images from my 40D (I shoot birds) than I did from my 30D.
Some of that might be the extra pixels on the sensor, but much of it is down to the AF, of that I'm absolutely certain.
It's very good for BIFs too.
Oh - I use AI Servo all the time, even on notionally stationary subjects, to great effect.
keithreeder wrote:
And a lot of user-to-user variation - the vast majoriity of these issues (though not Alan's, I suspect) are usually pinned down to user error or unreasonable expectations...
Agree regarding user-to-user variation as part of the problem we have whenever "equipment issues" are discussed.
Additionally, you might be right about "vast majority" but I doubt it. However, I always reserve the right to be wrong; how is that for an out? :-)
I remembered the time when 10D is launched it was being touted as having superior AF to the D60.
When the 20D is out, the old 10D becomes "slow focusing".
I think it's little soon to dis the 40D, the 50D isn't out yet.
That's rather cute. Astute observation.
I don't think the 20D/30D/40D/400D are that different in AF speed and accuracy.
Are you guys actually saying that there is nothing wrong with the second shot? From what I can tell the image is absolutely out of focus for no reason other than the camera had no idea what to focus on. While I admit my "technique" needs work, especially now that I am using Canon gear, I think the OP has valid points that I completely agree with.
treebeard wrote:
Are you guys actually saying that there is nothing wrong with the second shot? From what I can tell the image is absolutely out of focus for no reason other than the camera had no idea what to focus on. While I admit my "technique" needs work, especially now that I am using Canon gear, I think the OP has valid points that I completely agree with.
Oh, there is something wrong with shot number 2, but it is probably not the AF. As trenchmonkey said, is it probably too low shutter speed or maybe just plain camera shake. If nothing is in focus, it is probably not simply focusing at the wrong distance.
jkurkjia wrote:
There must be a lot of unit-to-unit variation because my copy (and that of another 40D owner I know) feel very responsive and consistently nail the subject (moving ducks and a variety of stationary objects as well). I'm guessing you have a bad one; it happens but knowing it happens doesn't make anybody feel any better. If it isn't too late for an exchange I'd send the body to Canon.
Regards,
Joe Kurkjian
Joe, it's already been to Canon for a month and it came back unchanged. And as I think I said earlier, here in the Australian market we don't get the benefit of user-satisfaction returns. Unless Canon recognise that it is faulty then we are stuck with our purchase because the shop doesn't want to be stuck with it and probably cannot sell it again as a new camera.
I finally got fed up with my 40D and took the financial loss (about a third) by selling it. I feel better now. I have a clean concience about selling it because Canon confirmed it was ok. Whether I got a bad one or whether my perception of a good camera has simply been spoiled by having 1-series cameras I'm not sure, but I surely was not happy with the 40D in any operating mode. The bad finally outweighed the good as far as I was concerned.
If you have a good 40D then I am pleased for you and just a tad jealous but please accept that I know enough about using SLR cameras to recognise that mine was not as good as canon imply it is, pretty much like BarnDog's one. However, since Canon tested it and reckon it is within specs then I can only assume it is a generic issue with the somewhat loose standard of the 40Ds (or the technician was incompetent, which I doubt) in which some are far better than others.
For those who think I whinge too much about the 40D you may well be right but that will happen far less now that I've sold mine and it is not irritating me every day that I try to use it
I won't go so far as to recommend that nobody buy a 40D but I will recommend that nobody buy any new Canon model until after the early adopters have had it for a few months to give it a clean bill of health. At least then you'll have a chance to weigh up the chances of getting a dud, and it seems to me that whether it be the 1D3 or the 40D the chance is much higher than negligible.
Mike Lucas wrote:
With the last one at least, it's clear what went wrong - as has been noted many times, the focus sensors are bigger than the red boxes. Here the camera focused on teacher's leg (or perhaps the floorboard?) instead of the boy's sweatshirt. Higher contrast - the focus jumps. As expected.
I don't have a convenient explanation for the second shot however - that one baffles me.
It's not as clear to me. Sure your explanation could be right but the child's shoulder is hardly a low-contrast edge and being closer to the camera the AF system should reasonably have focused on it. I've had such errors happen even after the system had already been focused on the closer subject for a few frames.
Successful AF is about a useful combination of greatest contrast and closest distance. Not just one or the other.
I've also had examples such as that second image and it has happened even though the camera is still being held steady with IS on and after it has already been focused on the subject (by which I'm implying no IS start-up issues).
Dawei, that's bad news for you but please let me know what Canon say about it and what they do about it. If they fix it then ask for details of what was wrong. I was not allowed to talk with a technician when mine was sent to Sydney, and there are none here in Perth. I'm 2,500 miles from the nearest Canon technician and there's a receptionist between us
cheers,
- Alan
Why not try Canon Singapore if you are in Perth ...
Alan321 wrote:
If you have a good 40D then I am pleased for you and just a tad jealous but please accept that I know enough about using SLR cameras to recognise that mine was not as good as canon imply it is, pretty much like BarnDog's one.
I'm well aware there exist body-to-body variations and have zero doubt you can tell what will and will not work for you.
It bothers me that I'm forced to spend a few days testing new equipment before actually using it. My normal routine includes static AF testing, if okay move on to dynamic AF testing shooting constant velocity cars, if okay move on to field testing on know difficult dynamics targets, and finally, if the field test passes I can FINALLY use the camera.
Fortunately I don't shoot humans and normally shoot under conditions of good illumination and therefore don't test flash/camera performance; this is a real time-saver but I have no idea how many sins are buried in the camera because I haven't taken the time to check out the flash system.
IMO it is a real shame sufficient QC isn't designed into either the camera or factory test process and we customers have to do most of the work. Because of this I never buy gear from anybody without a good return/exchange policy that includes free round trip shipping.
I hope you have better luck on your next equipment purchase.
trenchmonkey wrote:
Nothing's in focus in the 2nd shot. Looks like an IS lens was used that didn't have time
to spool up or some serious camera shake going on.
Actually shot #2 was taken @ 125/sec w/ a Canon 85mm prime f1.8. This is not camera shake it is simply OOF. This is exactly what I have been troubled with. I have the camera set to give the audible signal focus when locked, used the center focus point and took the shot.
I also was under the assumption that if given a good contrast area the focus point would lock on the closer object. Shot #1 is an example of this. Shot # 3 is an example where this doe not hold true.
I know this could be a problem with only my camera and I am not trying to add to this post with a "sky is falling" reaction. I am simply showing that I found similar occurences to the OP.
I look through my pics and am disturbed when I have missed a shot due to missed focus. I do this enough without gear anomalies adding to it.
Since posting here, I shot a kids Christmas music ensemble with my 30D under much the same conditions as the shots I posted. I did not notice a single "anomaly"
Venus wrote:
Why not try Canon Singapore if you are in Perth ...
Precisely. Canon Australia, just like Pentax Australia, is staffed with ineffective technicians. I have only heard bad things about bad services there. In Australia, Nikon is the only sensible choice with a decent service center.
. However, the last image again features the same error -the focus square is both on the shirt AND the teacher's leg behind. This time the camera decided to focus on the leg. Try to make it easier for the camera next time. Pick a contrast are on the same subject