D800 Firmware Update Focus Improvement!

Started Apr 2, 2013 | Discussions thread
m_appeal Veteran Member • Posts: 3,434
Re: Another example of misfocus

michaeladawson wrote:

m_appeal wrote:

michaeladawson wrote:

m_appeal wrote:

GroWeb wrote:

m_appeal wrote:

Focus point was clearly positioned on the eyebrow, yet the camera chose to on the eyeglasses...

Did you take these shots in AF-C rather than AF-S focusing mode? Some people -- myself included -- have found in AF-C (even in single point mode, and definitely in a dynamic area mode), that if the camera does not find the target under the selected AF sensor to be sufficiently contrasty, and perceives that there is a better, more contrasty target under a neighbouring focus sensor, it will use that sensor rather than the one you selected.

Yep. It was taken in AF-C. Arent you supposed to use AF-c to compensate for movement? The point is - it's not accurate.

But perhaps working exactly as designed.  That's the point.  If the camera decided that the eyeglass frame was a better target it may have chosen it instead.  Is the eyeglass frame in perfect focus?  If the camera's logic as programmed was to shift focus to the eyeglass frame and it achieved perfect focus on that target then the AF system is working perfectly.

??? So basically it's useless to select a focus point because a camera will pick something else to focus on? That's working as expected?

What you or I might expect has no relationship at all to how a software programmer or designer may have specified.  I totally see where you would say it is not "working as expected".  But it may be working as designed, regardless of what you may think of that design.

Having said that, I'm not sure that the camera did shift focus to the eyeglass frame.  Where did the red square come from?  Is that from the image viewer reading the focus point in the EXIF?  Or did you draw it in?  In my experience, when the camera decides to change focus points to an adjacent point it will also move the red square to show what AF point was actually used (not what point was selected by the user).  So in your case it would seem that the AF point stayed on the eye and eyelid and the camera missed focus.  But I'm not positive on how your image viewer works.

It comes from View NX (Nikon software) and shows the point that was used for focusing. I certainly didn't draw the red square in. I'm using SINGLE point Af-c, so the camera should not switch focus to another point.

I assumed you didn't draw it in, but had to ask.  All I can say here is that if you use AF-C with 9 or more points the camera DEFINITELY will chose one of the adjacent focus points if it does not like the contrast under the user selected AF point.  You can easily test this yourself with a featureless wall with some high contract object close by.  If the AF spot is on the featureless wall and the high contrast object is underneath one of the 9, 21, or 51 points the focus will shift to that point.  Is this expected behavior.  Too many it is not.  But it is in fact how the camera AF system is designed.

Now you say you used SINGLE point AF-C.  I would agree with you that the focus point should not shift (although others debate this) in this case.  In fact, View NX is confirming this by showing the red AF box over the eyelid.  So  I would agree that your camera appears to be mis-focusing.

That's exactly why I'm using single point and not 9 points. I don't know if I should send it in to Nikon to have a look at or if it's within spec. I don't think these results are acceptable though.

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