Ismail N writes:
D4 & D800 AF and 35 1.4G fine focus adjust and (Nikon) AF in general—
I have found exactly as you do in your comments re Nikon AF and its inability to focus properly with fast lenses.
* There are only two deterministic events in the AF system - the transition points into and out of the AF zone of focus - where the AF system deems focus to be achieved.
* With fast lenses (f/2 or faster) this AF zone is actually wider than the depth of field, let alone the zone of sharpest focus. Therefore, when using a fast lens (in MF mode) it is possible without changing anything other than the focus ring position to take a shot which is front focused, in focus, and back focused all which the AF system indicate as being 'in focus'.
* The AF system simply seems to place focus somewhere within this 'AF zone' leading to incorrect focus with fast lenses.
* With manual lenses (such as some f/2 Zeiss zf2 lenses I have) it is possible to skew the AF fine tuning value so that one of the two deterministic events mentioned above in the AF (I chose the transition into the AF zone turning from camera) falls sufficiently within the 'AF zone' so that it can reasonably consistently indicate sharp focus.
* with AF lenses, the best that can be achieved is to fine tune the lens so that front focus occurs equally as often as back focus. This can further be helped by using AF-C so that the system hunts within the AF zone and success rate is increased.
As distance increases so do the errors.
The AF system is simply incapable of correctly determining focus with fast lenses. The issue exists with 1.8 lenses as well, it's simply worse with 1.4 lenses.
I switched to Nikon this year but how has this gone unnoticed for so long?
And your emphasis on sharp focus and focus shift (beyond f/2.8 which I had previously ignored) has improved performance obtained from my 14-24 in particular - thanks!
DIGLLOYD: It has not gone unnoticed by me and has made me rather grumpy about the whole matter when shooting in the field— and I’ve been reporting on it for years. People see what they want to see, and often test incorrectly (e.g. a high contrast black and white chart at close range, a fast telephoto extrapolating to a short fast lens, etc).
Gunnar P writes:
On the topic of D4 vs. D800 and Roy P’s view on autofocus accuracy. Not surprised that Roy thinks D4 has better AF even though officially AF systems are "the same".
I owned the D3S. (I sold it recently). It was a heavy but ergonomically unbelievably pleasing. Most importantly, I almost never had an out of focus image. I take a lot of pics of my kids who are "fast moving objects". In a series of 10 frames, maybe the first two were unsharp but then AF locked on and I always had tack sharp results.
With my D800, this is now different. And I am not talking about small errors that are visible due to to higher resolution. I am talking about a very high rate of "throw away" pictures that are very, very unsharp. And it is clearly an AF issue -- when the d800 locks on -- e.g., me doing a manual focus on my zeiss lens - the results are amazing.
No matter whether I use 2.8 pro zoom lenses or my beloved 28-300 -- the D800 AF produces a good amount of through away pics even though AF seemingly "locked on". There seem to be many moments of back/front focus situations, irrespective of the lens. I am only avoiding this in MF mode but that just doesn't work with moving kids.
So, I am now waiting for a D4X .
DIGLLOYD: Nikon autofocus is the pits in the real world from what I can tell from my use in the field. I’ve had far better luck with Canon autofocus on the 5D Mark III and 1D X. Shooting black and white charts to test AF is photographic masturbation. Getting focus accurate in the field on real subjects is the real deal and while I’m not going to go “prove” it, in my repeated experience, Nikon autofocus is screwed in the field on real subjects. But since I almost always use manual focus (preferably with Zeiss lenses), I have an alternative that works for me.