Nikon: auto-focus correction required?

[Dave Humphries]

Hi Nikon people,

I know there are a fair few Nikon D90 owners out there and some that shoot wildlife with long lenses.

What I need to know is this; have you found it necessary to use the auto-focus offset memory feature* when swapping between lenses?

The problem I have had with my Sigma 18-250mm on my new D5000 was that if I focused on something, I got the peep-peep, but it focused say 5% behind where it should. i.e. the calibration of auto-focus was off. It took a few weeks to convince myself this was bad enough to complain about and that it wasn't just me using an inappropriate focusing mode, or the subject had moved, etc.

This was only an issue at the 250mm end and occured whether I was shooting across a 50m wide river, or a butterfly at 18"; if I went to 100% pixel view, my shots were clearly mostly NOT focused where the should be, fortunately the Nikon software can display (in PP) which focus point was in use to confirm.

Here's a gallery of real world test shots: http://www.pbase.com/dhumphries/testdemoshots
The 'test_100' shots are screen snips of the 100% view, showing the focus point (red box). Very telling are the butterfly and the Heron, where I focused in front of the subject deliberately, just look where it's sharpest.

Anyway, as I'd shunned the internet in favour of a dealer ('thank heavens' as it turns out), I went back and eventually we got some tripod mounted, self-timer taken test shots off with 3 lenses; a Sigma 18-200, a Nikon 18-200 and my 18-250. These, of an A4 sheet of paper shot at an oblique angle, with an obvious focus point, the text was sharpest both fore and aft of the focus point on the other two lenses, but not mine, which was sharper behind the supposedly focused point. With those shots in the bag, (or on a CD) the lens has gone back to Sigma for calibration, return eagerly awaited.

The dealer has kindly lent me a s/h Nikon 18-70mm older kit lens to keep me going, but imagine my dismay when, even at 70mm, this too is apparently displaying a smaller but noticeable tendency to focus past where it should. So, even if the Sigma comes back OK, this makes me fearful of buying any other lens without careful checking in future, as I don't want to be here again.

Now, the D5000 doesn't have any memories for lenses, but I gather the D90 does and this includes a focus offset (true?), if so do you use it?
Or am I just unlucky (twice), or too critical?

Thanks in advance, and I apologise for the long and rambling post.

Cheers,

[Mossy]

The D90 doesn't have the feature you're talking about (I certainly can't seem to find it in mine), only the D300, D700 and D3x has it as far as I know, maybe the D3 as well.

Secondly, you could try the ruler test instead of the thing with the paper. Taking a shot along the ruler will give you an idea of just how far the front/back focusing issues go.

[Dave Humphries]

Thanks Mossy,

The paper test was pretty conclusive and had something easy for the AF to get a hold of. Easier than a ruler where you might argue what it latched onto - although I accept something in inches or mm is easier to quantify.

I'll put the paper pics up later.

I have, since writing the above, just reviewed what I shot yesterday, did I say (above) the Nikon lens at 70mm was less affected, I may have been wrong more later.

The shop tests, with a ratio of 2 lenses agreeing and one not, pointed to the lens being at fault rather than camera body, but now this other lens is showing similar results, that makes it 2-all. However, as I'm only a week and a half away from foreign holiday, I don't want to be sending the camera away just yet! I shall just have to work around it if the Sigma comes back the same as it went.

Thanks for prompt response, at least I can feel better that I couldn't have afforded a body with that feature

[paulphoto]

I saw on youtube that nikon 200 has the fine-focus adjustment

[Markvetnz]

I also suspect that my new 50D needs some microadjustment to the AF. At 100% pics are not sharp, especially with the ef 16 -35 L II. A stop of 22 should give excellent DOF with everything nice and sharp - correct? I'll spend some time tomorrow focus calibrating each of my lenses to the camera. I found a link on the net how to go about it.

[Colin Southern]

I also suspect that my new 50D needs some microadjustment to the AF. At 100% pics are not sharp, especially with the ef 16 -35 L II. A stop of 22 should give excellent DOF with everything nice and sharp - correct? I'll spend some time tomorrow focus calibrating each of my lenses to the camera. I found a link on the net how to go about it.

Hi Mark,

Are you shooting JPEGs or RAW?

The microadjustment results are only visable at 100% magnification ("pixel peeping") - and then you'll probably only be able to see a small difference in steps of about 5. Many (myself included) have concluded that it's not worth the effort.

On the other hand - the thing that's makes a FAR bigger difference is correct sharpening. If your evaluating your images at a pixel level (ie 100% magnification) it's vital to apply capture sharpening to RAW images - Canon recommend an unsharp mask of 300%, 0.3 pixels, and a threshold of 1 (I prefer 0) for low ISOs.

If your shooting JPEGs then the overall sharpness will be a function of your picture style setting.

[GillR]

I also suspect that my new 50D needs some microadjustment to the AF. At 100% pics are not sharp, especially with the ef 16 -35 L II. A stop of 22 should give excellent DOF with everything nice and sharp - correct? I'll spend some time tomorrow focus calibrating each of my lenses to the camera. I found a link on the net how to go about it.

I won't pretend for a second that I understand the science behind it, but is it possible that the softness you're seeing at f22 is a result of diffraction? Have you tried comparing the same scene taken at, say f11?

[Dave Humphries]

"Great minds ..." Gill,

That thought crossed my mind too!

Obviously, any such calibration is going to have to be done wide open (I assume).

Cheers,

[Colin Southern]

I won't pretend for a second that I understand the science behind it, but is it possible that the softness you're seeing at f22 is a result of diffraction?

In my opinion, diffraction - as a source of image degradation - is somewhat over-rated; I'm finding that with a little sharpening shooting at F22 / F32 really doesn't degrade an image that's printed up to 44" wide all that much. It is visable when "pixel peeping", but usually isn't anything overly dramatic even then. Personally, I don't normally even worry about it.