Anyone else in here experiencing the Canon EOS 20D backfocusing quite a bit? I have shot in different modes (One-Shot, AI Focus, AI Servo) and with 2 different lens (Canon 135mm f2.0 and 50mm f1.4). If there are background subjects, the lens loves to jump on them. Yet, when I view the images in the Canon EOS Viewer Utility, it shows the focus confirmation square lit for the subject I want, yet, it is out-of-focus, while the background (what should be out-of-focus) is in perfect focus. I use the mutliple points and used the multicontroller joystick to select a single focus point. It doesn't matter.
I am going to send the camera off to Canon tomorrow.
Also, anyone else still noticing horizontal banding in ISO1600 and especially ISO3200? ISO3200 is basically useless, because of this banding. This is with and without flash and after the latest firmware version 1.1.0, which was suppose to fix this problem.
I'm not convinced that your icicle is out of focus. It might be a little bit blown, but looks like it resides in the plane of focus (along with the brick wall and part of the longer icicle beside it).
But also, if you're shooting at f/1.4 at that distance your DOF is pretty miniscule. A tiny movement of your body will result in a change of focus - did you use a tripod?
I am wondering if this is a problem with shallow DOF (Depth-of-Field) and auto-focus. I used both the 135mm at f2.0 as well as the 50mm at f2.0. So I don't think it is the lenses. If so, what's the point of having fast glass with this camera?
Are you sure it isn't a viewfinder problem? Like i'd seem that if initially when you focus on something and it isn't in focus in the actual photo, that the difference is between what your viewfinder shows you is in focus and what is actually in focus. Maybe it goes deeper in, but still thats my initial reaction... It sucks that the viewfinder is only an additional set of lenses and requires its own focusing that you can't truly tell if it is in focus or not... Er something, now I'm just babbling, but yea... Consider doing some tests with the viewfinder focus? Not sure... just throwing out some ideas...
crfrey71 wrote:
I am wondering if this is a problem with shallow DOF (Depth-of-Field) and auto-focus. I used both the 135mm at f2.0 as well as the 50mm at f2.0. So I don't think it is the lenses. If so, what's the point of having fast glass with this camera?
Well with fast glass the DoF is so miniscule that you need a tripod and precise control in order to achieve a perfect shot. That's part of the skill of photography I guess, and if you can't hold it, you move to a tripod. Or use the AF servo and hope it at least stays in focus long enough to get some shots off...
Then again, it could be due to human error - miniscule movement towards/away from the subject that gives a trememdous effect especially at f/1.4. I was pre-focussed and ready to capture the shot, and might have retreated/advanced a bit due to the winds/cold/shivers.
data1ore wrote:
More evidence...
I'd say that's out of focus - unless I'm terribly mistaken.
Okay, but it still looks like the focus on the brick wall is due to it being approximately in plane with the tip of the icicle. Did you use a tripod? If not, I'd say that's your problem. Very slight movement.
given my past back-focusing issues with the 10d and the 1d-mkII, i'm pretty sure your camera is backfocusing. after some searching on the internet and a lot of camera talk, i have gathered that it seems to be an anoying problem, but not problematic enough for canon to acknowledge its existence. my suggestion is to send your camera in with the lenses and have canon calibrate them. it sounds annoying, but it will cure your problem. i drove up to my service center and they had the stuff back to me that same day.
fotophroggie, I believe it's not backfocussing...I've not noticed this issue with my other shots. It's probably due to human error - I was waiting for the moment to shoot the drop of water falling, and might have moved in a tad. My bad sadly, and as a result, an otherwise nice photo's down the drain.
*sobs*
More macro training/tripod for me. That'll teach me to try shooting macro while walking about in the future.
I clearly focused on the icicle, intent on catching the drop of water as it fell. It ended up giving me the brick wall as the focussed spot.
!@##!@$%^^....
What was in focus in the viewfinder? Always check the viewfinder. That's why professional lenses have Full Time Manual focusing as a feature.
If the right object was in focus in the viewfinder, then you have a problem with mirror alignment. Either way, I think it should no longer come as a surprise to anyone that the camera (most likely) or the lens may need calibration. I'd send the camera in first, even if it focuses well with other lenses. It's possible for both the camera and a lens to be miscalibrated complementarily to each other.
well i'd be inclined to believe you if you weren't shooting at 1/1000th of a second. the DOF would be a problem, but even in the event of camera shake, the area under the selected focus point should still maintain the most focus. from your examples, it doesn't appear so. anyway, i hope i'm wrong. if you really just want to test it, just put something about three feet in front of a wall and shoot away. that should answer any questions you may have.
RDKirk, I couldn't really tell from the viewfinder on the 20D. That's what professionals pay big bucks for :-). The camera was beeping happily at that particular focus point.
fotophroggie, the 1/1000 only prevented camera shake. It doesn't prevent me from wavering back and forth. I prefocussed, and held the shutter halfway. That means the focus point was set at the icicle. But I must've moved forward a tad prior to shooting, which gave rise to the brick wall in focus.
My 20D has yet to give me backfocusing issues on my 17-85....and in light of my own recollections as I was shooting that shot, my explanation makes the most sense to me.
it doesn't matter if the shutter is 1/1000s or not. as long as there is user movement between the time the camera locks focus and when the user triggers the shutter, the plane of focus will be moved, especially at such closeup distance and large apertures. it is more likely a classic case of user error and not equipment malfunction.
Another thing. The 20D and 10D AF sensors are actually rather larger than the little red boxes (about three times as big - the centre sensor is actually cross shaped not box shaped - the others are long thin lines about three times as long as wide) - and love high contrast. You can check the exact size of the sensor by seeing where the focus snaps in as you move the sensor box towards a high contrast mark on an otherwise plane (and out of focus!) surface. That icicle has very little contrast (especially in the direction in whcih that sensor is sensitive- remember some of the outer sensors are sensitive to vertical contrast some to horizontal). So the cameras going to have problems - it looks like its actually focussed on the water drop, slightly in front of the ice - but with little contrast and few sharp edges it's hard for us to judge what is in focus, let alone the camera!. It's also really hard to judge where the exact plane of focus runs ( it will be parallel to the sensor, which is of course facing partly skywards, so the plane of focus is not running up and down vertically in the image, if you see what I mean) ) in an image when you are looking at it from a funny angle like this - in fact I think the icicle is only very slightly out of focus but there is very little of the ice which crosses the plane, whereas because of the angle a lot more of the wall happens to be in that plane making the whole thing look very odd. It' s much easier to judge these things when you have subjects parallel or perpendicular to the direction the camera is facing!
i do stand corrected.