Autofocus issues

[george013]

Dan,
You may correct me but he is using the backbutton with af-s. Meaning there might be a delay between catching the focus and taking the picture. In bif it's usual af-c as far as I know. I don't use it.

George

[Marconikon]

Yes, I know it's not the best way to test a lens but sometimes it just happens to shoot close to the minimum distance since it is "only" 60cm on this particular lens... and the results are so off I'd just save time shooting manual focus at those distances even at f/5 - f/8. By the way, I'm using AF-C most of the times and BBF, it was set on AF-S only because I was using a tripod, anyway yes it could happen to miss a shot in AF-S because of camera movement, but i find it quite hard at narrow apertures..

[NorthernFocus]

Dan,
You may correct me but he is using the backbutton with af-s. Meaning there might be a delay between catching the focus and taking the picture. In bif it's usual af-c as far as I know. I don't use it.

George

Good point. If hand holding at short range in particular that can be an issue. I have AF-S disabled on the bodies which allow it.

[Marconikon]

Here are two images shot handheld unfortunately with auto iso (forgot it) demonstrating the issue, I know not a great test but still I believe I should be able to get better results from the AF.. 70mm f/5, distance is roughly 70-80cm and cropped more than 100%, jpegs from raw with no adjustments except for a slight exposure correction to make them match.

Contrast detection: https://www.dropbox.com/s/nk3gc41qi1...C0256.jpg?dl=0
Phase detection with AF-C single point (obviously i kept the BB pressed): https://www.dropbox.com/s/sp9k7uzb17...257-2.jpg?dl=0

[Marconikon]

That's utter nonsense. Every professional wildlife photographer that I know uses BBF. Sometimes I wonder if anyone at Nikon is actually a photographer.


Then why bother with phase? Plus this is contrary to Nikon's own marketing mantra. This is true only for a flat target perpendicular to the camera. It may be true that contrast is just as accurate but it is not as precise. At least not as it is deployed in the current Nikon bodies. For 3D targets phase is more precise.


Did you inquire specifically about the alignment problem that some D800s had and whether your serial number falls into that lot? As I recall you can tell whether your body has been checked/corrected by looking at the mounting thread on bottom of the camera(I know it sounds ridiculous but so it is). If there is a black dot at the bottom of the threaded hole then it has been corrected. I had an early model D800E and had to send it in under warranty and it returned with said marking. If is sounds to weird to believe, Google it.


My experience with Nikon is that their shop test procedures/practices do not represent real world shooting. It is extremely difficult to communicate with them. I was having problems with a lens that they could not reproduce. I had to send them my own images to demonstrate the issue and then they corrected it.


No. Sounds like you are having an issue. As George suggested I would try testing at various ranges.

These comments were very useful and I was thinking most of the things you wrote!

[george013]

There's no exif in the pictures.
What I can see is there's ca in the pictures. A green one and a red one. I would say a lens fault. I leave it to others.

George

[pschlute]

Have you tried using a rocket blower in the mirror box? Perhaps you have a dust speck on the AF sensor.

[Stagecoach]

Marco,

1) according to him the AF inconsistency depends on the fact I am using AF-On focusing (back button) and release priority in AF-S single point though I am not recomposing nor I'm letting go the AF-On button before i release the shutter.

You really can not undertake any serious testing to diagnose possible AF inaccuracies (e.g. front/back focusing) whilst holding the camera, even if it's on a tripod. Although this may give you a suspicion that something is not right it is then necessary to undertake controlled methodical testing to reach a conclusion that will consist of a number of 'comparative' tests.

Of the two images you linked to, one shows what the camera/lens is capable of and the other simply shows what was achieved at another specific instance. The images/crops are different with respect to lighting and subject positioning for starters, and with that subject you will never really know what the AF chose within its operational area to lock onto.

Place your camera on a tripod with a remote release (or use the timer), with a subject at a similar distance to that you are concerned about. Take a shot with Live View, this will give you a reference of 'the best focus/result' your camera/lens can produce.

Very carefully (no movement to anything) set your camera to AF fine tune adjust +20. Take shots at increments of 5 through to -20 (9 shots). Ensure that for each shot the camera is de-focused first and has then focused (use beep or green light confirm) and the lighting is constant.

Put that sequence of images (10 in total) into something such as ViewNX, scroll through and look to see if one of the 9 images taken is comparable to the Live View image. Any conclusions should be based upon a number of repeated tests, and different targets.

The below two images are an example of two within a sequence of 9 that show the difference with a range of 20 adjustment of the AF fine tune and are 100% crops. The sharper of the two was equal to what I could achieve with manual/live view focusing. The 'focus' box was placed at the pencil with the sharpening line (paint/wood) in the centre.

[Stagecoach]

There's no exif in the pictures.
What I can see is there's ca in the pictures. A green one and a red one. I would say a lens fault. I leave it to others.

George

I would consider seeing CA in the pictures perfectly normal and absolutely no indication of a 'lens fault' George.

What is significant with the CA is that it is entirely different in each shot, suggesting uncontrolled testing conditions with Marco's linked examples.

[george013]

I would consider seeing CA in the pictures perfectly normal and absolutely no indication of a 'lens fault' George.

What is significant with the CA is that it is entirely different in each shot, suggesting uncontrolled testing conditions with Marco's linked examples.

The two pictures have two different colors. I think that means they focus at a slightly different distance. I would try that lens on another camera.

George

[Stagecoach]

The two pictures have two different colors. I think that means they focus at a slightly different distance. I would try that lens on another camera.

George

George,

Marco has stated, they were hand held, different ISOs and one altered in post. These are all variables/unknowns that are going to give 'differences' in the supposedly same images other than rear/back focus.

[NorthernFocus]

Marco,



You really can not undertake any serious testing to diagnose possible AF inaccuracies (e.g. front/back focusing) whilst holding the camera, even if it's on a tripod. Although this may give you a suspicion that something is not right it is then necessary to undertake controlled methodical testing to reach a conclusion that will consist of a number of 'comparative' tests...

I agree with Grahame regarding any "serious" testing. Doing some rough handheld testing to confirm suspicions that you have a problem is fine. But in order to truly validate, quantify, and correct an AF issue with any degree of accuracy requires a disciplined process similar to that indicated in his post.

Phase detect AF has a bit of inherent variability. In order to accurately tune a lens/camera combination testing needs to provide enough information to minimize the effects of statistical variability in the system. From a practical standpoint testing needs to be rigorous enough to tune within one's own needs/tolerance. But it is an unrealistic expectation that every shot made with phase detect AF will produce an image with the focal point precisely where it theoretically should be within the DOF. So it is very difficult to make any corrections with casual one off test shots.

Simply based on the fact that hand held samples are resulting in perceived problems and the Nikon service shop has not identified any issues with their testing, a disciplined test procedure is advisable before making any decisions.

[george013]

George,

Marco has stated, they were hand held, different ISOs and one altered in post. These are all variables/unknowns that are going to give 'differences' in the supposedly same images other than rear/back focus.

CA is a lens property. And the D800 is a high resolution camera, small sensel size. And two different colors means to me to different focus distances.
From https://en.wikipedia.org/wiki/Chromatic_aberration


Some lenses don't work well on a high resolution camera. Search for "d800 technical guide". I can't link to it at the moment.

George

[Marconikon]

There are so many replies I don’t know whom to answer first, thanks for the interest! First of all I know these shots are not to be considered a test and I linked them just as a “demonstration” of the results I consistently get at short distances! As soon as I have some time I will test it more rigorously... I believe the CA to be perfectly normal since I’m shooting at 70mm 80cm away, so I have a DOF of about 3-4cm and the “target” was in direct hard sunlight.. and the lens is like 10 years old or more. I can’t stress it enough, I’m not trying to asses the entity of my back focus (pretty sure it is a back focus since I can focus closer than the confirmed focus point manually and get better results), I’m just trying to understand if it is a problem Nikon can and should solve or if I have to live with it and just resign to the fact that at 70mm at short distances I’ll have to manually focus

[george013]

There are so many replies I don’t know whom to answer first, thanks for the interest! First of all I know these shots are not to be considered a test and I linked them just as a “demonstration” of the results I consistently get at short distances! As soon as I have some time I will test it more rigorously... I believe the CA to be perfectly normal since I’m shooting at 70mm 80cm away, so I have a DOF of about 3-4cm and the “target” was in direct hard sunlight.. and the lens is like 10 years old or more. I can’t stress it enough, I’m not trying to asses the entity of my back focus (pretty sure it is a back focus since I can focus closer than the confirmed focus point manually and get better results), I’m just trying to understand if it is a problem Nikon can and should solve or if I have to live with it and just resign to the fact that at 70mm at short distances I’ll have to manually focus

I think ca is more visible at longer distances. The internal image distance is bigger and the top angel of the light cone hitting the sensor is bigger so the spreading of the different colors is bigger. Yours is visible at the shortest distance and the maximal focal length. That might be a hint for the quality of the lens. And enlarged by the smaller sensel size of the sensor. Maybe your af is influenced by it too.

If you test it again I would do the test with also a longer distance and a shorter focal length.

Be aware that image quality doesn't stop at a certain point, it's a sliding scale where you've to find your own acceptable point. In your example you choosed for 2 extremes: maximal focal length and minimal focusing distance.

I do think it's a lens issue. But who am I???

George

[Marconikon]

Well it could be a lens issue, not excluding it! Just saying CA is present in the slightly OOF and totally OOF parts of the image for the one in focus (the focus point is on the k towards the front) and much more evident in the missed focused one. That’s why I believe it to be normal, it’s a D lens and on the d800 with such a high resolution FF sensor it’ll show up in any high contrast scene at wide apertures. My problem is I can’t seem to find a way to AF at close distances (0.6-1m) at 50-70mm other than using manual focus and shifting the plane of focus closer, I’m talking about more than 90% misses...

[Stagecoach]

My problem is I can’t seem to find a way to AF at close distances (0.6-1m) at 50-70mm other than using manual focus and shifting the plane of focus closer, I’m talking about more than 90% misses...

Can we take it that your above comment is with respect to the 'Lens AF-Fine tune' being set at -20?

[Marconikon]

No, I’m sorry I didn’t specify, the fine tune was off for the tests, tried -20 and it makes it slightly better (but far from acceptable) at short distances, the only problem is that such a setting compromises the results at longer distances (won’t focus correctly at infinity)

[Stagecoach]

No, I’m sorry I didn’t specify, the fine tune was off for the tests, tried -20 and it makes it slightly better (but far from acceptable) at short distances, the only problem is that such a setting compromises the results at longer distances (won’t focus correctly at infinity)

Thanks for that Marco.

[Manfred M]

Macronikon - have you tested the focus accuracy with the more modern "G" lenses? Both of the lenses that you list as being problematic are the older "D" style that rely on an in-camera focus motor rather than the in-lens ones found in more modern lenses.

The older lenses do have a longer focus throw than the newer ones. Gear lash could come into play (the focus screw diameter appears to be quite small) as could older feedback technology on precise focusing. The focusing issues that you are having seem to occur at or near the minimum focus distances of these lenses.