Even though we ARE Nikon lovers,we are NOT affiliated with Nikon Corp. in any way.

English German French

Sign up Login
Home Forums Articles Galleries Recent Photos Contest Help Search News Workshops Shop Upgrade Membership Recommended
members
All members Wiki Contests Vouchers Apps Newsletter THE NIKONIAN™ Magazines Podcasts Fundraising

Autofocus - a question for the experts

mszak50

US
20 posts

Click to send email to this author Click to send private message to this author
mszak50 Registered since 01st Feb 2009
Sun 17-Jul-11 10:24 PM

I have this interesting autofocus behavior that I can reproduce at will with the D7000, but which does not reproduce with a D90.

1. Daylight, 85mm 1.8D, F2.8
2. about 4' to target
3. Target is a spyder lens focus calibration target
4. Single focus point - center selected
5. AF-S, focus priority

If I aim the focus point on the intersection of the black and white area, my focus is spot on, which is what I would expect.

If I move the focus point an inch up and to the left, to a 100% white area, the camera still locks on, but things are now front focused about an inch.

My D90 with the same lens won't lock onto the all white area. This is really handy, because then I know that I don't have the shot, and can do something right then and there to make sure I do get the shot.

With the D7000, I don't find out until I get home that I didn't get the shot, which is too late.

(BTW, I can also repro this with a Nikon 50mm 1.4D)




Visit my Nikonians gallery.

GroovyGeek

Portland, US
84 posts

Click to send email to this author Click to send private message to this author

#1. "RE: Autofocus - a question for the experts" | In response to Reply # 0

GroovyGeek Registered since 16th Jul 2011
Mon 18-Jul-11 04:18 AM

Assuming that you are indeed in single focus spot mode you may need to map where your focus sensor actually is. It is rarely where the viewfinder markings may seem to indicate, which may explain why the D7k locks differently from your D90. To map the location of your focus sensors

1) Take a toothpick and draw several heavy black stripes on one end
2) Set the camera to AF-C
3) Focus on a busy pattern such as newsprint
4) While in continuous focus mode start moving the toothpick a few inches above the surface of the newsprint and note at what point it "steals" the focus from its original point. You can map the left, right, top and bottom position of the sensor with this method.

The sole cross sensor on my D90 is along the bottom line and to the right of the area indicated in the viewfinder. This can be very handy when trying to ensure that focus is spot on with a narrow DOF.

Acknowledgement for the method goes to Leon Goodman, google "D70 backfocus" to find his full description.

mszak50

US
20 posts

Click to send email to this author Click to send private message to this author

#2. "RE: Autofocus - a question for the experts" | In response to Reply # 1

mszak50 Registered since 01st Feb 2009
Mon 18-Jul-11 08:38 AM | edited Mon 18-Jul-11 10:18 AM by briantilley


Thank you for that suggestion - I did a quick test using described procedure. Unfortunately, on this camera, the focus points are pretty much where they say they area - certainly they are off no where near enough to explain what I'm seeing.

Here's a link to a gallery of test images - I was using the top most (in portrait mode) cross point sensor - the D7000 images that are in focus have the sensor on the intersection of the 4 large black and white squares. The out of focus images have it on the center of the upper left large white square. The D90 ones are focus regardless of where I put the sensor.

http://www.inbalancephoto.com/Other/D7000-VS-D90/18092791_vtP3RF

Visit my Nikonians gallery.

Leonard62

Pa, US
4419 posts

Click to send email to this author Click to send private message to this author

#3. "RE: Autofocus - a question for the experts" | In response to Reply # 0

Leonard62 Gold Member Awarded for excellent contributions and sharing his in-depth knowledge and experience with the community, especially of Nikkor Lenses Writer Ribbon awarded for his contributions to the Nikonians Resources articles library Nikonian since 15th Mar 2009
Mon 18-Jul-11 03:20 PM | edited Mon 18-Jul-11 03:27 PM by Leonard62

Hi Mike,

I'm certainly not an expert, but what you describe is not really unexpected. The instruction manual states that auto focus may not be obtained if a low contrast area is chosen even though the focus dot appears. This would allow the shutter to fire even though focus priority is chosen. It's just not a good way to test for focus accuracy. I don't have a D7000 but I do have the Spyder Lenscal and the the 85mm f1.8D. So I tried to duplicate your test with my D300. I was very surprised that autofocus was accurate whether I used the intersection or the center of the white square as a focus point. I used your setup conditions exactly. The only thing I can think of is that maybe different models will react slightly differently when trying to focus on a no contrast area.

edit: I just tried the same test with my D200 and it wouldn't autofocus at all on the center of the white square.

Len

Visit my Nikonians gallery.

KnightPhoto

Alberta, CA
4969 posts

Click to send email to this author Click to send private message to this author

#4. "RE: Autofocus - a question for the experts" | In response to Reply # 0

KnightPhoto Gold Member Nikonian since 18th Dec 2006
Tue 19-Jul-11 01:08 AM

Hi Mike,

I find your thread of high interest. I have a similar and difficult to reproduce symptom when shooting with my 24-70 lens in theatre.

It seems to me, my apologies to all readers for the subjective nature of what I describe, that occasionally my D7K, when it fails to find a suitable contrast in my intended target, seems to set its focus well into the back of the theatre (possibly to infinity or beyond in my case) and therefore if DOF happens to be insufficient my subject is forward of the zone of focus.

In my case it is an intermittent and probable poor AF target in low light occurring phenomenon, where I will lose a burst or partial burst of photos as a result. Mind you it sometimes only lasts a photo or two and then AF is properly achieved and I generally get good photos in theatre.

Anyhow, long story, similar to you I feel like the D7000 still tries to make a photo out of a poor situation. It feels like the AF firmware has a new twist as compared to three other Nikon DSLRs I have owned.

I do have a test situation at home where I can force the D7000 (pointed at a plain gray pillar) to not achieve focus, just like my D700. So far then, unlike you I haven't been able to reproduce the phenomenon at will.

I don't believe my problem is a focus tuning issue because in my cases focus is WAY off when this occurs. Much more so than a mere tuning exercise could correct. So I am hoping Nikon engineers will with a few reports popping up recognize this as an area for improvement and possible firmware update. I have experimented with setting the camera in FOCUS rather than RELEASE mode, but IIRC this is not helping. I am using AF-C and nine-point dynamic in my tricky situation. Not of direct help to you I realize, but it could be a similar phenomena to yours (and it might be the source of some of the AF difficulties people are experiencing and reporting on with this camera). I am trying to keep an open mind and an eye out for any pattern as I continue to use the camera.

Best regards, SteveK

'A camera is an instrument that teaches people how to see without a camera.' -- Dorothea Lange
My Nikonians gallery
My Nikonians Blog

jamtins

Meadow Heights, AU
87 posts

Click to send email to this author Click to send private message to this author

#5. "RE: Autofocus - a question for the experts" | In response to Reply # 4

jamtins Registered since 03rd May 2011
Tue 19-Jul-11 08:17 AM

Hi Steve,

I am glad you posted about your theatre issue, for I too have been somewhat perplexed by the behaviour of focus, both AFC and AFS on the D7000.
I shoot small birds in a grey forest. Lens is a 500 F4 (non VR). Tripod mounted, locked down, bean bagged.

Just occasionally, I get results where the head/chest is blurry, (but that is where the focus point is shown in software) and just a small distance behind is rendered sharp. So I can get a head/chest/wings every so slightly soft and then mid along back, or legs pin sharp.
On other occasions, perhaps on the same shoot, the bird eye is pin sharp.
I have contemplated focus tuning, but like you, don't see the point.
If every shoot was a problem, then I could go for a fix, but it is unpredictable. Even large birds, falcons or kites on a branch against the sky at say 25 yards can cause the same sort of result.

I came from a D90,D200 , and shot most of last year's season and never saw this happen once. (previously was using a wonderful old 600 f5.6 manual focus)
The images are used for magazines, and wall prints.

I do think it may be a low lighting problem (low contrast) even on say a bird with red , white or grey chest feathers but as they are part of my body of work, it is not much help to say shoot only in bright light on contrasty subjects.

Thanks again for posting and being quite definitive with the problem, I was hopelessly unsure how to describe it.

Interestingly enough if I shoot video, it looks crisp and sharp. But I am not a video person.

DJ

.. I set out to discover the inventions of God. -John Muir

Visit my Nikonians gallery.

Visit my Nikonians gallery.

mszak50

US
20 posts

Click to send email to this author Click to send private message to this author

#6. "RE: Autofocus - a question for the experts" | In response to Reply # 5

mszak50 Registered since 01st Feb 2009
Tue 19-Jul-11 07:37 PM


Just an update - after seeing my test images, Nikon wants the camera back (for the second time).

Visit my Nikonians gallery.

jamtins

Meadow Heights, AU
87 posts

Click to send email to this author Click to send private message to this author

#7. "RE: Autofocus - a question for the experts" | In response to Reply # 6

jamtins Registered since 03rd May 2011
Wed 20-Jul-11 06:17 AM

Hi,
Be interested to hear how it goes with Nikon, because if it makes a difference, I will take mine to the local Nikon repair centre.

I am going to try to shoot some test shots in the field, keep complete records of each shot and try a variety of focus settings. But in the end, I suspect now that the camera is putting an extra turn or so on the autofocus motor, even though it is passing the focus point.

I also wonder if it only happens in one direction?
If it goes from infinity to the subject distance or if it goes from closeup to the subject distance?
Will also include that in the tests somehow. FInding a co-operative bird will be the first point of order.

Hope the trip to the doctor cures the problem

DJ

.. I set out to discover the inventions of God. -John Muir

Visit my Nikonians gallery.

Visit my Nikonians gallery.

Len Shepherd

Yorkshire, UK
12722 posts

Click to send email to this author Click to send private message to this author

#8. "RE: What you are describing is an AF limitation" | In response to Reply # 0

Len Shepherd Gold Member Nikonian since 09th Mar 2003
Wed 20-Jul-11 07:53 AM

>If I move the focus point an inch up and to the left, to a 100% white area, the camera still locks on, but things are i >front focused about an inch.
>
>My D90 with the same lens won't lock onto the all white area. This really handy, because then I know that I don't have
>the shot, and can do something right then and there to make
>sure I do get the shot.
What you are describing is part an AF limitation.
As AF develops (as it has between D90 and D7000) it becomes more advanced, but there are always a few subjects that an older system handles better than the new.
The D90 better handles your scenario, but the D7000 AF can do many things the D90 cannot do.
Phase detect AF takes a segment from the picture area. The segment is projected onto 2 AF detection lines.
AF then selects an area of good contrast on each line and attempts to focus it.
If the selected detail is in the center on each AF dection line focus should be good.
If the detected detail is outside the center on each line focus is behind the selected detail.
If the detail is inside the center on each line focus is in front of the selected detail.
The camera uses this information to decide which way to AF or, if using manual focus, to indicate which way to turn the focus ring.
If there is no detectable detail at the center of the AF mark the AF system does not know that
As you demonstrated if the camera can detect a subject outside the AF mark it can focus at the wrong distance.
The D3 series and D700 AF work in a similar way to the D7000.
On a detail your D7000 can detect AF detail even more off center in group dynamic AF than in single point.
The dection zone overlaps onto the next AF viewfinder mark in group dynamic - to make passing the subject from 1 AF point to another easier when focus tracking.
Page 93 of your D7000 lists 6 types of subject which may cause similar focus errors. As many factors affect AF ability often Nikon's list of 6 provide good AF - but sometimes they do not
Nikon's guidance is also at
https://nikoneurope-en.custhelp.com/app/answers/detail/a_id/4585
For critical work it can be wise to check AF accuracy on the rear monitor. If AF is not accurate Nikon suggest switching to manual focus.
In Example 2 Nikon caution high contrast lines are not always good for AF.
If you get a front or back focus indication with Spyder Cal my advice is double check with a different AF target as, as Nikon say, single high contrast lines can in themselves result in focus error.

briantilley

Paignton, UK
30235 posts

Click to send email to this author Click to send private message to this author

#9. "RE: Autofocus - a question for the experts" | In response to Reply # 0

briantilley Gold Member Deep knowledge of bodies and lens; high level photography skills Donor Ribbon awarded for his support to the Fundraising Campaign 2014 Nikonian since 26th Jan 2003
Wed 20-Jul-11 08:33 AM | edited Wed 20-Jul-11 11:04 AM by briantilley

>Phase detect AF takes a segment from the picture area.
>The segment is projected onto 2 AF detection lines.

For those who may be interested in an alternative view regarding the detail of AF systems, I would suggest checking the posts from Nikonians Team member benveniste in this recent thread in our Nikkor Lenses Forum.

Brian
Welsh Nikonian

mszak50

US
20 posts

Click to send email to this author Click to send private message to this author

#10. "RE: What you are describing is an AF limitation" | In response to Reply # 8

mszak50 Registered since 01st Feb 2009
Wed 20-Jul-11 10:46 AM | edited Wed 20-Jul-11 11:01 AM by briantilley


There is no question that target selection affects the ability of the D7000 to auto focus. It affects the ability of the D90 to auto focus as well.

...but back to what I'm really trying to solve...

I would expect that I can use the top most cross point sensor (in portrait mode) to focus on the subject's eye at F2.8, and get a sharp, in focus photo. Specifically, I expect to see eyelash separation on the near eye, and for a bust up shot or tighter, expect to be able to see clear reflections in the eye (often seeing some old guy holding a camera).

It may be that the designers of the D7000 elected to "give up" some efficacy in that area to gain ground in another area, but there are a lot of single focus point portrait shooters out there - while it may not be a stretch, it's not a slam dunk conclusion that makes complete sense (that is, there's no particular evidence that suggests that is exactly what happened).

In the images that I sent to Nikon, the area in focus was quite a distance away from the selected focus point - it's almost like the full matrix of sensors was active, and it just was focusing like it was in AF area mode.

What I was hoping to get from Nikon after sending in the images was something along the lines of "Hey dummy, you need to do X instead of Y". Or even "This is how all D7000s work" (and a pointer to the page you pointed me to). The former would been perfect - I could adjust and move on, the latter would have been acceptable...I could then know I had to sell the camera and find something more suitable for the type of portrait work I do.

I did look at the link you sent - I really think that page is really meant for new shooters that get images that have obvious focus issues. Every one of the scenarios they outlined there at the scale that they imply I can pick out in the view finder (and with screw type lens that I use, I can close my eyes and *hear* that things aren't right). I don't even press the shutter. (And in general, with single point focus, I don't run into issues that the camera was focused on the person as opposed to the skyscraper in the background).

I'm hoping that this is just a "my particular D7000" - Nikon wants the camera back after seeing the RAW images from both my D90 and D7000 under identical tests - that at least gives me hope that there really is something wrong with the camera, and that it can be remedied.


Visit my Nikonians gallery.

Len Shepherd

Yorkshire, UK
12722 posts

Click to send email to this author Click to send private message to this author

#11. "RE: What you are describing is an AF limitation" | In response to Reply # 10

Len Shepherd Gold Member Nikonian since 09th Mar 2003
Wed 20-Jul-11 06:05 PM

>I would expect that I can use the top most cross point sensor (in portrait mode) to focus on the subject's eye at F2.8, and
>get a sharp, in focus photo. Specifically, I expect to see eyelash separation on the near eye,
On a detail I presume you know only the central block of AF points are the more efficient cross type, and the outer points which are located for what is often the best point for focusing on an eye are the less efficient single direction type - reading left right with the camera in portrait mode.
Sometime the AF can locate an eye perfect, sometimes not.
The yes is more likely with directional sidelighting, eye make up and false eye lashes - all of which contribute to helping AF detect the intended target.
The no is most likely to be soft lighting, smooth skin, fine short eyebrow hair, and no make-up - all of which make it more difficult for AF to detect an eye.
Digressing to Canon, many Canon bodies (and the discontinued Nikon D2 series) have a more efficient cross type sensor near top or bottom with the camera in portrait mode.
When the D4 series come on my wish list is more cross type AF points. Telling a D7000 user (instructions page 300) to consider manual focus when the camera cannot auto focus is not ideal.


Photography is a bit like archery. A technically better camera, lens or arrow may not hit the target as often as it could if the photographer or archer does not practice enough.

Len Shepherd

Len Shepherd

Yorkshire, UK
12722 posts

Click to send email to this author Click to send private message to this author

#12. "RE: Autofocus - a question for the experts" | In response to Reply # 9

Len Shepherd Gold Member Nikonian since 09th Mar 2003
Wed 20-Jul-11 06:11 PM

>For those who may be interested in an alternative view regarding the detail of AF systems, I would suggest checking
>the posts from Nikonians Team member benveniste in this recent thread in our Nikkor Lenses Forum.

My explanation in this thread, also in all good textbooks, has the advantage of explaining how the AF system knows whether focus is on the subject, in front of the subject, or behind the subject

Photography is a bit like archery. A technically better camera, lens or arrow may not hit the target as often as it could if the photographer or archer does not practice enough.

Len Shepherd

G