Hate to admit it, but I have the AF issue. Going to make a case with Nikon. Any advice?

Started Nov 4, 2012 | Discussions thread
Fayard
Regular MemberPosts: 127
Like?
Re: Hate to admit it, but you are wrong.
In reply to olyflyer, Nov 7, 2012

olyflyer wrote:

Fayard wrote:

Forget about Leonard. He is stuck in page 100 on the Nikon booklet. Even if he's right in theory, a long as your target has some details to focus on, you should be fine.

That's incorrect and bordering nonsense. The fact that you don't seem to like Leonard doesn't make him, or Nikon wrong about how the AF sensors and the cameras are working. Regardless what you think, page 100 is very important when you are trying to test focus. Page 100 is true to 100% and nobody with any "focus issue" should ignore what it says. It is very clear that:

"Autofocus does not perform well under the conditions listed below. The shutter release may be disabled if the camera is unable to focus under these conditions, or the in-focus indicator (‚óŹ) may be displayed and the camera may sound a beep, allowing the shutter to be released even when the subject is not in focus."

You may not believe in it, but dismissing it and telling others to ignore it based on your feelings towards Leonard is not helping anyone, not even you.

Careful selection of target is important, just like good light and steady tripod, as well as mirror lock and in fact also a remote trigger to eliminate "AF error" caused by camera shake.

Also what you are saying is wrong when you say that "as long as your target has some details to focus on, you should be fine". This is totally wrong because with some targets the camera must struggle more than with others and while you might be able to focus on certain target at one time, at another the camera may fail or could give you false focus confirmation, just like page 100 explains. Focus tests can not done on any target because that kind of approach can give you any results, good or bad or anything in between.

Olyflyer,

I've done many test on autofocus, and I give my experience. I've done all the testing being careful with everything: target, light, tripod. Here are my comments:

Target: I think that the best target to take is the one available on this page. It is made of horizontal black and white lines, with different width so the Phase detect system won't match one line with another. The fact that the lines are horizontal provides some vertical contrast that can be easily tracked by outer sensor on the D800. My experience is that most targets work, and the barcode of this book works pretty well. Of course, if you take a specific target with a repeating pattern, you might fool the system. It seems that people from Nikon have the same experience. The article available here (Nikon article about Fine Tuning) does not go crazy about the target.

Lightning: Of course you need some good lightning, and the color temperature of your light may have some very little influence on the focus. Of course it you turn your camera to AF-C and you realize that your camera keeps hunting, there is something wrong with your target/light combination. Anyway, my advice is to shoot 6 times the same target with the same conditions. You would get a feeling for the shot to shot variation on the AF system.

Tripod: The tripod might be useful for 2 things. Firstly, you don't want to shake when you take the picture as it might lead to some blur. As there is also the blur problem coming from the fact that lenses are usually weaker on the side than on the middle, my advice is to use the method I've described above where you check on the paper on the ground where the focus has landed. Therefore, with this method, shooting faster than 4 times the focal length, the tripod is useless. Secondly, a tripod can be useful to prevent you from moving in between the time you focus and the time you shoot. If your depth of field is 1cm (0.3 inches), I agree that if you don't use a tripod, your test is worthless. But if your depth of field is 10cm (3 inches), I find the tripod useless.

If you want to do this kind of test, the first thing is too think carefully. And obviously Leonard does not think at all. All he does is keep referring to the same page of the manual. Anybody doing carefully the tests will find that they are problems far more important, the biggest one being shot to shot variation. You'll always find that there is at least a little bit of variation, even in bright sunshine, even with the best target you can think of (the one above, the one given here by Marianne that may come from Leonard, any USAF chart, any siemens star), and this variation gets bigger and bigger when you go to wider lenses and you go too long focusing distance (for instance with a 35mm at a distance of 4 meters (4 yards)).

So, do the tests with something on the floor to know where the focus has landed, try many targets and you'll realize that Leonard goes crazy about the quality of the target.

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow