AGeoJO wrote:
Daan, that water bottle is sharp considering it was taken at f/1.4 and I assume it was without any sharpening or other adjustment, correct? I am glad that things are looking up now for you, at least so it seems. What did you do different from before then? Are you going to expand the test a tad further by checking the other outside AF points out?
You may have a steady hand and you cranked up the ISO to get a fairly high shutterspeed but my word of caution is always, again, to always use a tripod to do some kind of AF issue tests. At least, so you know that it is consistent and the camera/lens movement variable is taken out of the equation. In general, I deem any AF test without a sturdy tripod as unreliable. I am sure you get my point.......Show more →
Right... no PP. The samples in the first posts were taken with the opposite outer AF point (most to the right). There are AF points in the inner circle that also perform good.
I have another appointment with Canon Service coming Tuesday. Everything will be checked (again ). I will try to hand the tech a list (illustrated by 100% crops) including all my lenses and how they perform on all 19 points (both vertically and horizontally). As a reference to myself and as an aid to the tech. I want this BS to be over with. Don't worry, I will use a tripod and remote release (just like I did in the first posts) for this.
BTW Do you get consistent results on all the 19 AF points?
AGeoJO wrote:
Nope, to be frank, I limit my testing only to front/backfocus issues. I used to pixel peed some 3-4 years ago and by so doing I didn't enjoy photography that much. Now, as long as I know that my lenses do not have any undetected front or backfocus issues, I am not worried, just go out and shoot.
This is a 100% crop of one of the tiger shots at at the zoo last Sunday I posted on your other thread. Yes, some PP was done on it.....
I guess you have the freedom not to worry. I need my gear in order to make a living. I want to know what its limitations are so I can deal with them in work situations. The results I get are very inconsistent and sometimes so bad that they are unusable. Unacceptable. Even more, because I know this camera is capable of delivering. The question is: is Canon Service capable of making the right adjustments? Or maybe something is broken, that needs to be fixed? Or maybe it is just what it is...?
AGeoJO wrote:
Again, what did you do different between the first shots and the last shot though that led to that dramatic improvement? Or was it just inconsistent then?
I used two different AF points. Like I said, there is a lot of inconsistency between all the 19 AF points (with all my lenses).
I think we are just seeing the limitations of the design of the AF system.
With 21 megapixels and lenses like the 24L, 35L and 50L, Canon needs to rethink the whole question of focus accuracy. Focus adjustment is just a way of fobbing the responsibility off onto the abused customer. Meanwhile, when the shot matters, I focus manually.
First I placed the cam in landscape mode on a tripod at about 1,75 meter and shot the following sequence for every AF point: f/1.4 AF/MF, f/2 AF/MF, f/2.8 AF/MF and f/4 AF/MF.
10 out of the 19 AF points were off. That is, too much difference between AF and MF (like in the first posts). I think of these 5 AF points were consistently terrible. The other 5 AF points were inconsistently terrible: sometimes focus was good, but most times it was off. Also, some seem to front-focus, while others had a back-focus.
Secondly I placed the cam in portrait mode on a tripod at about 1,75 meter and shot the same sequence as I did for the landscape mode.
But now "only" 8 out of the 19 AF points were off. Some points that were off, were the same as those that were off when shooting in lanscape mode. But there were also AF points off that were good in landscape mode and the other way around. Here also some seem to front-focus and some seem to have a back-focus.
Now I will have to find out if the AF points that are off, are the same for every other lens (long and wide). If not, then there is of course a major problem, because in that case caliobration of the 19 AF points is useless. There is of course also the possibility that the mis-focus occurs random. But this doesn't match with other test results. Obviously, there are some "bad" AF points. Badly calibrated or maybe "broken"?
The fact that some points give good AF performance, gives me hope that it is possible to get good AF on all the 19 AF points. The question is, what does it take to make that happen?
Daan,
I just tested the 35 on my MkIII. I choose an object with definite lines in it (a vacuum hose).
With the 35L there were definitely AF points that would work in one orientation but not the other and one that would not work at all. I am defining 'Not Working' as acquiring focus lock when it clearly is nowhere near and taking the shot.
One point just perpetually hunted. I tried it on the 85L too, same thing. The 50 1.4 wasnt too bad though.
morganb4 wrote:
Daan,
I just tested the 35 on my MkIII. I choose an object with definite lines in it (a vacuum hose).
With the 35L there were definitely AF points that would work in one orientation but not the other and one that would not work at all. I am defining 'Not Working' as acquiring focus lock when it clearly is nowhere near and taking the shot.
One point just perpetually hunted. I tried it on the 85L too, same thing. The 50 1.4 wasnt too bad though.
Thanks
If two lenses are worse than one other, maybe it isn't a 100% body issue. How else can you explain these differences? Maybe it has something to do with body/lens communication?
I already tested my 24-70L @ 35mm on all the 19 AF points in landscape mode using the same method as with my 35L (see above post). 9 out of 19 AF points were off.
7 AF points that were way off using the 35L are also way off on the 24-70L @ 35mm. However, there is also 1 AF point off that was fine with the 35L and one that is good but was off with the 35L. But in my defence, manual focus has a bit more room with the 24-70L than with the 35L, so I could have messed up somewhere.
morganb4 wrote:
Daan, the thing is with thr 50, I can test it until the cows come home, take it into a real world situation and it falls over in a screaming heap.
Its quite possible that there is a body issue that is exacerbated by some lenses which just dont worry other cameras.
Incidentally, I tried the same experiment on a 1Ds Mk II and it was fine on all points.
I don't know what to think of it anymore...
I tested the 24-70L @ 35mm in portrait mode... 8 out of 19 AF points were off. Only 4 of them resemble points that were off while using the 35L in portrait mode.
So maybe it is random. Sometimes I get a good focus on a point, and the next moment it is off (like a light switch). Some AF points are totally off every time.
That probably explains the difference in AF/MF between long and wide. Still, is this acceptable? I doubt if in case of the 35L we are 1/3 of the DOF wide open (1.4). By f/4 it barely catches up. Canon probably finds it acceptable... But I think that is a sign of weakness on Canon's part.
I think it is a sign of weakness of AF systems in general as Nikon is probably in the same position. I understand the need for tolerances but consider the implications.
* Say your shot is dead on using the center point.
* You take the same shot using an outer focus point and it is off by X but still within tolerance.
* You take another shot and it is off by -X but again within tolerances.
* You compare the first shot with the second and third shot and say they are OOF. They are, but not from Canon's perspective.
* You send the whole kit to Canon which adjusts it (meaning that it could come back worse but may still be within tolerances) or does nothing because "nothing" is wrong.
Seek and you will find the shortcomings of a system, whether it is Canon or Nikon. I share Henry's sentiment here and don't let this issue drive you nuts! Lets face it, how often in real lif do we take shots using an f/1.4 aperture at about 3 feet plus you have to do it really fast? In real life, you stop down the lens for more depth-of-field and/or shoot from a longer distance or do both, again increasing the DOF. If it is coming down to crucial shots, like brainiac suggested, use manual focus in conjunction with LiveView and bracket the focus even. Enjoy photography whether it is a hobby, an obseession or at a professional level.
morganb4 wrote:
I shoot wide open with this lens a lot, often close.
Secondly, how can an AF confirmed shot like this not drive me nuts....
I AM SICK AND TIRED OF PEOPLE TELLING US THAT THE CAMERA IS FINE!!!!
Within tolleramce = wrong
User Error = wrong
That I should just live with it = wrong
Please add something constructive as asking us to just get over it is kind unreasonable and shows that you really dont understand the situation.
Did you see my post above, although it was addressed to Daan after I did my own test? I tried to assist him; I don't care about your camera and your complaining about this here on this forum won't solve the issue. What do we know about this anyway. Take it up to Canon like Daan did and will further do to satisfy his requirements. Maybe you already did that, I care less.
You are trying to focus on a curved surface. What possibly makes you thnik the autofocus system shoul read your mind and pick the same exact spot within the focus sensor area to focus on as you do? Try the test over focusing on a flat surface parallel to the sensor plane.
Thanks Josh, I as do all the other let down MKIII owners appreciate you confirming that the 35L can indeed work with the new generation. It helps me feel that there is light at the end of the tunnel.
To deny however that the tunnel is a scary place is unfair. As I said (yes I stand by it). I routinely shoot wide open with this lens. I could before why cant I now. I dot understand why I now have to compromise my shooting style and the assingments I get, to benefit the camera.
Can you imagine a reporter that is sent on assingment coming back and saying 'sorry guv, wasnt enough light I have to shoot everything at f4 since I got my new super low light machine'.
I apologise if I came on to strong but I really am sick of being told its either not a problem or its my fault. I stand by my point Josh: that it is that not sensible to suggest that I should not use this camera wide open and that you have missed point by asserting this.
I also disagree that discussing it in the forum wont help, I am hoping that I can try absolutely everything that I can and close off every avenue of posibility before it goes back again, that way the techs will have a much clearer picture from me instead of me saying 'it just deosnt work sometimes but I cant put my finger on it'. As a result of this thread and Daans excelent work I have now found that I have many sensors on the left that have un-synched cross type sensors.
AGeoJO wrote:
Did you see my post above, although it was addressed to Daan after I did my own test? I tried to assist him; I don't care about your camera and your complaining about this here on this forum won't solve the issue. What do we know about this anyway. Take it up to Canon like Daan did and will further do to satisfy his requirements. Maybe you already did that, I care less.
morganb4 wrote:
Thanks Josh, I as do all the other let down MKIII owners appreciate you confirming that the 35L can indeed work with the new generation. It helps me feel that there is light at the end of the tunnel.
To deny however that the tunnel is a scary place is unfair. As I said (yes I stand by it). I routinely shoot wide open with this lens. I could before why cant I now. I dot understand why I now have to compromise my shooting style and the assingments I get, to benefit the camera.
Can you imagine a reporter that is sent on assingment coming back and saying 'sorry guv, wasnt enough light I have to shoot everything at f4 since I got my new super low light machine'.
I apologise if I came on to strong but I really am sick of being told its either not a problem or its my fault. I stand by my point Josh: that it is that not sensible to suggest that I should not use this camera wide open and that you have missed point by asserting this. ...Show more →
No sweat; we are civilized ! I understand your frustration and Daan's as well. I assumed that you have done everything and anything that could be done, including AF microadjustment, updated the firmware, etc. Again, my only suggestion after that is to complain vigorously with Canon Australia and give them hell....
I am hoping that I can try absolutely everything that I can and close off every avenue of posibility before it goes back again, that way the techs will have a much clearer picture from me instead of me saying 'it just deosnt work sometimes but I cant put my finger on it'. As a result of this thread and Daans excelent work I have now found that I have many sensors on the left that have un-synched cross type sensors.
I really appreciate your concerns about my mental sanity... Probably justified, because I am beginning to loose it a bit
I am not a tester or pixel-peeper (I hear myself after 4 hours of boring test shots ). I only want to make sure I can communicate my AF problems in a way that the tech will understand them fully and hopefully take care of them. Or declare this camera dead for that matter... There are also others with what seems to be the same AF problems. It is a great way to use a forum such as this to inform one another. And maybe getting some useful insights along the way. So, as far as I am concerned, every input is appreciated
BTW I freaked out over the differences that both my 35L and 24-70L seem to have on the AF points. Later I realized that the 35L is calibrated inside the camera and the 24-70L not. I will have to do some further tests to see how my 85L II and 135L are holding up. Overall the worst AF points seem to be on the right side (in landscape mode). I hope I can detect some kind of pattern.
AGeoJO wrote:
Seek and you will find the shortcomings of a system, whether it is Canon or Nikon. I share Henry's sentiment here and don't let this issue drive you nuts! Lets face it, how often in real lif do we take shots using an f/1.4 aperture at about 3 feet plus you have to do it really fast? In real life, you stop down the lens for more depth-of-field and/or shoot from a longer distance or do both, again increasing the DOF. If it is coming down to crucial shots, like brainiac suggested, use manual focus in conjunction with LiveView and bracket the focus even. Enjoy photography whether it is a hobby, an obseession or at a professional level.
I really like the look of the 35L with 1Dmk3 @ 1.4 @ 3-5ft and I will see myself using it often once Canon get my 1Dmk3 working properly. 3-5 ft is will cover about 1/2 body portraits. And I found out about the problem not from doing test shots but doing actual shots. I was shooting a friend reading a book, focused on the eyes but came out OOF and the book in front is sharp.
). I will try to hand the tech a list (illustrated by 100% crops) including all my lenses and how they perform on all 19 points (both vertically and horizontally). As a reference to myself and as an aid to the tech. I want this BS to be over with. Don't worry, I will use a tripod and remote release (just like I did in the first posts) for this.
! I understand your frustration and Daan's as well. I assumed that you have done everything and anything that could be done, including AF microadjustment, updated the firmware, etc. Again, my only suggestion after that is to complain vigorously with Canon Australia and give them hell....
). I only want to make sure I can communicate my AF problems in a way that the tech will understand them fully and hopefully take care of them. Or declare this camera dead for that matter... There are also others with what seems to be the same AF problems. It is a great way to use a forum such as this to inform one another. And maybe getting some useful insights along the way. So, as far as I am concerned, every input is appreciated