D800 Focus Theory

Started Apr 18, 2012 | Discussions thread
em_dee_aitch Veteran Member • Posts: 3,675
Re: It works...

russbarnes wrote:

OMG it works. Focussed over target with D800 and 50mm f/1.4G at f/1.4 using outer left, locked with AF-L, recompose to centre and it's as sharp as when focussed with the centre point. I've just done that three times over three different objects. Looks like my 50mm f/1.4G needs a service doesn't it? I'm still going to wait and see what Nikon have to say though because there doesn't seem to be anyone out there talking about a real right-sided weakness as others have said, which by the law of averages would show itself as much as the left if all things were equal.

You should not use focus and recompose when testing AF. When you do that you are not testing AF but rather are just exploring the shape of the lens' field of focus. For a focus and recompose test to work, you would actually need a lens with a field of focus is curved in a manner to correspond to the arc proscribed by your recompose maneuver. If you were using a true flat-field lens, you would be blatantly decreasing your focus, and a good result would in fact be a "false positive" indicator of "good focus" that is actually faulty.

But back to your overall theory: It is blown out of the water by one simple test, that for symmetry. We should not expect our lenses to have totally flat focus fields, but we should expect them (and the body mounts) to be aligned with a reasonable amount of symmetry, meaning that when using the center AF point that identical subjects placed symmetrically on the left/right sides should be equally in or out of focus. For that reason symmetry is relatively simple to evaluate, though tracking down the source of it can be complex.

Anyway, my point being that "blame it on the lens" is a bogus theory for cases in which the lens is shown to be symmetrical on a different D800 sample or on a D3x or even a D3s. Believe it or not, a D3s is perfectly adequate for detecting symmetry issues smaller than some of the huge errors that D800 users are displaying here. In these D800 cases it should be easy to isolate the symmetry issue to the AF system in the D800 so long as you have a second body on which the lens tests correctly. Comparing the CDAF result at a single point is also an undeniable test.

So, if you:

1. Show the lens is symmetrically good with even performance, by testing it on other camera bodies, so that you know the lens is not the likely problem, and

2. Subsequently perform your phase detect (viewfinder) AF test on the D800, and

3. Verify the focus miss by comparing Live View CDAF, and

4. Take some center AF point shots of symmetrical subject matter on the D800 to verify the symmetry of the lens on the D800 itself.

... then I think you can be reasonably confident that a D800 problem is real and not "blame it on the lens."

I totally believe those of you who claim to have perfect D800 copies, and I also believe those of you showing clear problems, because it back up "my" theory, which is the usual horrible QC effort by Nikon

-- hide signature --

David Hill
San Francisco & San Jose, CA | Austin, TX
Wedding Photographer and Apparent Gearhead

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