stanj wrote:
My 35L works flawlessly but no amount of prayer etc. will make my 24L work reliably. Granted, even less magnification, but still annoying.
Does your 35L also works flawlessly on the outer AF points? Maybe you can do similair comparison shots as I did, AF vs MF on outer AF points. I would be very interested in your findings
I have the same issue with the 40D and 35mm f/1.4. It will front-focus a tad at f/1.4. At f/4, the front-focusing goes away. All other lenses in my lineup have no front/back-focus issues. This is including my 50mm f/1.4 and my 85mm f/1.8. A big difference between the the 35L and the other two is more chromatic aberration at f/1.4.
In other words, nothing is broken, but "in spec" isn't good enough. If I send the camera in to Canon, which would be foolish because with all other lenses there's no issue, I'd probably get a nice note that indicates that nothing could be done, and I'd be down the shipping costs.
So for the bokeh effect at f/1.4, manual focus with live/view is the only thing reasonable on this lens / camera combo. Perhaps the 5D Mark II when it comes out will have something a bit more updated than the 1Ds III or 40D (which both came out around the same time) in the way that it handles focusing, but I'm not hopeful.
The reason I haven't sold the lens is because it still heads and tails better than the 16-35mm f/2.8 L II zoom I have at 35mm at all apertures sharpness wise. There is no other substitute in the Canon lineup that can compete with the 35mm L, and I'm not fond of 3rd party lenses.
Daan B wrote:
Does your 35L also works flawlessly on the outer AF points? Maybe you can do similair comparison shots as I did, AF vs MF on outer AF points. I would be very interested in your findings
While I haven't done a specific test like you did, I use outer points pretty much exclusively and the yield rate is about what I would expect given the subject matter (uncooperative toddlers), light (not so much) and all other adverse conditions. I'll try to shoot a comparison like yours over the weekend.
stanj wrote:
While I haven't done a specific test like you did, I use outer points pretty much exclusively and the yield rate is about what I would expect given the subject matter (uncooperative toddlers), light (not so much) and all other adverse conditions. I'll try to shoot a comparison like yours over the weekend.
Stan, if I remember correctly you use AI Servo in these cases, right? Please, do the tests by using One Shot. When I use AI Sero in combination with the outer AF points, I often get much better results than when using One Shot. Even on static subjects. And certainly on moving subjects. The above tests were done by using One Shot
Next week I will try to make a similar test with my 1Ds3. On Sunday it will get a good workout with a 300/2.8 IS as I shoot off centre AF points quite regularly for motorsport.
Alistair Watson wrote:
Next week I will try to make a similar test with my 1Ds3. On Sunday it will get a good workout with a 300/2.8 IS as I shoot off centre AF points quite regularly for motorsport.
Daan B wrote:
Stan, if I remember correctly you use AI Servo in these cases, right? Please, do the tests by using One Shot. When I use AI Sero in combination with the outer AF points, I often get much better results than when using One Shot. Even on static subjects. And certainly on moving subjects. The above tests were done by using One Shot
Correct assumption, Daan - and yes I'll switch to One Shot for the experiment.
I am taking Monday off work so I should be able to do the experiment no later than that.
Alistair Watson wrote:
Next week I will try to make a similar test with my 1Ds3. On Sunday it will get a good workout with a 300/2.8 IS as I shoot off centre AF points quite regularly for motorsport.
Will let you know.
Daan B wrote:
Thanks, I really appreciate that
No worries mate. If it works, I will sell you my 35L at a very reasonable inflated price!
Just a thought... my understanding, which is perhaps wrong but comes from a senior Canon Tech, is that the AF has a pretty wide tolerance of what is "acceptable". In fact, I was told that it is equal to 1/3 of the DOF at maximum aperture. If this is the case, then as the lens gets wider the tolerance gets greater as DOF increases as the focal length decreases.
By way of example, let's say you are on a tripod exactly 10 feet from the subject at 35mm and F2.8 you will have 4.36 feet of DOF and thus a AF focus tolerance of 1.45 feet. If you are using the lens at 70mm then the DOF will be 1.03 feet and the acceptable tolerance will be only .34 feet.
Since there is a big difference between sharpness derived from having "acceptable DOF" and critical focusing it would seem to me that you may not notice a 4" or less error at 70mm because it represents only 3% of the distance whereas the potential error with the 35mm is a very noticeable 14.5% (staying within acceptable tolerances).
Hrow wrote:
Just a thought... my understanding, which is perhaps wrong but comes from a senior Canon Tech, is that the AF has a pretty wide tolerance of what is "acceptable". In fact, I was told that it is equal to 1/3 of the DOF at maximum aperture. If this is the case, then as the lens gets wider the tolerance gets greater as DOF increases as the focal length decreases.
By way of example, let's say you are on a tripod exactly 10 feet from the subject at 35mm and F2.8 you will have 4.36 feet of DOF and thus a AF focus tolerance of 1.45 feet. If you are using the lens at 70mm then the DOF will be 1.03 feet and the acceptable tolerance will be only .34 feet.
Since there is a big difference between sharpness derived from having "acceptable DOF" and critical focusing it would seem to me that you may not notice a 4" or less error at 70mm because it represents only 3% of the distance whereas the potential error with the 35mm is a very noticeable 14.5% (staying within acceptable tolerances). ...Show more →
Henry,
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.
Daan B wrote:
You can see how sharp the lens is suppose to be by looking at the 100% MF crops. The "tests" are to show the difference between AF and MF.
BTW did you seen how distorted the troubled 100% AF crops look? Wonder what that is about...
Guys, when you test, try about 3-4ft. I notice on my 1dmk3 with 35L that the AF problem are not that noticeable at greater than 10ft. I get 4inches of front focus between 3-4ft when using the outer focusing point.
HeintjeLee wrote:
Guys, when you test, try about 3-4ft. I notice on my 1dmk3 with 35L that the AF problem are not that noticeable at greater than 10ft. I get 4inches of front focus between 3-4ft when using the outer focusing point.
Testing was done at 4ft (both 35L and 24-70L). I will try again using a MFD of 10ft when I have time.
Daan B wrote:
Testing was done at 4ft (both 35L and 24-70L). I will try again using a MFD of 10ft when I have time.
if your 35L gets better at that distance, then I could be sure that we have exactly the same problem. BTW, when I was showing the Canon tech the problem I was having, I showed him using shot from 2 focusing point and told him i have the same problem with the rest of the focusing points. I got the camera back with the same problem and he told me he only adjust the two focusing point!!! Well, I still have the same problem on that two focusing point though I notice that it tend to get a little bit higher success rate. By little higher, i mean from 1 out of 10 to 2 out of 10 I sent the camera back few days ago and just remembered that I should have also told the tech to check at 3-4 ft..
AGeoJO wrote:
Daan,
You issue bothered me a bit, since my 35mmL is my favorite available light lens. Although I didn't notice any issues using the lens for real life shooting before but I did a test using my 1Ds MarkIII and 35mm f/1.4L at wide open aperture to check it out.
I focused on the orange using the outside AF point to the right while the gear on a sturdy tripod. Again, in the single shot mode and I also used a cable release just to minimize any unknown variables. The distance was approximately 3 feet or 1 meter. I transferred the file to LR, lightened it up a touch, and that is the overall view of the image. No other adjustments were made. I created a virtual copy, cropped that to 100%. I exported both of them onto jpg and here they are..... I am pleased with the result, what can I say, the AF is right on the money. BTW, this lens required a +12 on the microadjustment usinge the center AF point previously. Like I said in your previous thread, this doesn't help your case but my point is the issue with the 35L on the 1Ds MarkIII doesn't seem to be wide spread.
Joshua, thanks for this little test. At least it is good to know that your 35L performs fine on that outer AF point (and maybe on all the others too ).
Inspired by your test I tried another little test myself. Nothing fancy, shot handheld. But now I used the most left outer AF point. It gave acceptable AF performance as you can judge by yourself...
Full Frame, f/1.6, ISO800 (conversion by LR, no PP):
My point is, that not all the 19 AF points show equal AF performance. But the most outer AF points are generally the worst. So I know good AF performance on all the AF points must be possible. Again, thanks for pointing that out However, I beginning to ask myself what the techs are doing in my case, since the @#$% 35L already has been serviced twice to my 1Ds3 (that also has been serviced twice). Maybe it is time to get nasty on them
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 to, 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....
I sent the camera back few days ago and just remembered that I should have also told the tech to check at 3-4 ft..
). 