manual focus problems on k10d

Ok, three different lenses, playing with the diopter till the view through the viewfinder really does look as sharp as it can. After a long day what I now see is that I need to focus about an inch in front of the subject for the subject to be perfectly in focus. (this at a distance to subject of about 6 yards/metres away)

Can anyone explain this problem?
Is there something I'm doing wrong with the diopter?
Has something gone wrong with some internal process in the camera (this problem may have existed since I got the camera but never noticed, although I'm sure I would have noticed this before as bad as it is.

Grateful for any suggestions as I don't relish the idea of shipping the camera back to the shop...

point your camera at a white/bright solid colored wall. (focus at this point is irrevelent)
set your diopter so that the af markers [ ] in the center of the viewfinder look razer sharp.
now it's set up right, leave it alone.

Perhaps it is not the diopter's fault??
I don't know if it is a common issue but I seem to have heard similar issues. My K10D definitely focuses in the wrong spot.

Take a newspaper, book or other evenly patterned object in front of the camera at a 45 degree angle. Get a real small DOF and focus on a single line.
You should see some amount of text in focus both behind and in front of the specific spot that you focused upon. View images on screen to make sure.

For my K10D, the line I focus on is just barely in focus, but the line immediately behind is out of focus. Several lines in front are in focus. I believe this is not how the focus system should work.
It is very frustrating at times, but I am not quite yet willing to loose my camera for 8 weeks while it gets shipped off for repairs.

It's quite common to find people who have difficulty with manual focus. Experiments with MF cameras showed that when a given subject was focused on by different, or even the same, people, the amount of error was surprisingly large.

This is why MF cameras have focusing aids such as microprism spots or split image screens, or a mixture of the two.

I wouldn't assume that there's a camera error - if there was then all your pictures would be blurred. MF or AF.

Very careful setting of the dioptre adjustment is essential and for the short sighted it may be helpful to use reading glasses for cameras with no such adjustment. The optical distance of the image in the viewfinder is about 1 metre and you need to be able to focus at that distance.

Hope that helps.

I downloaded an excellent focusing tool from a N-k-n site somewhere (happy to email it to anyone who wants it), tested the focus (same 45 degree angle sort of trick. It's exactly as amoringello says, the line just past the one I'm focusing on (and a bit of the next line) are crystal clear.

It's not my eyes or ability to focus, in any event, and after the great tip above on setting the diopter that was also sharp as can be so does that leave a camera problem?

John I'll check the AF tomorrow, it does sound logical that that would have to show the same problem for it to be camera error.

Thanks all.

Testing autofocus using the groovy little chart from http://focustestchart.com shows the same results, what's in best focus is consistently a few millimetres to a couple centimetres behind what I'm focusing on. Please tell me there's some way to fix this that I've overlooked...(outside of posting the camera back down to the dealers).

I don't know how accurate your testing is, I don't mean that negatively, I just don't know what sort of error we are talking about in relation to any experimental error.

Focusing screens will be in correct register within certain limits, like every manufactured item there will be some leeway, but small errors will be covered by depth of focus (as opposed to depth of field). As with all experiments, there is also our own experimental error as mentioned above.

The key question is, are your pictures critically sharp? If they are, don't worry. If they aren't then you know what to do.

I am bemused.

You claim you can detect an error of one inch in six yards. That's approximately 0.5%.

First, I'm amazed you can measure such a small focussing error so accurately, and second, at that distance the DOF will ensure that there is no problem. Or are you the fortunate possessor of a lens with a focal length of 1:0.2?

I am beginning to suspect a wind-up.

G

Photos are at: http://www.flickr.com/photos/10737481@N06

The focusing chart shots were done with a 16-45mm on autofocus and the leaf shot was manual with a 90mm lens at f2.5. I manually went out of focus each time, then let the autofocus zero in on the centre bit. They were taken from about 1 meter away.

In the garden shot I focused very very carefully (from about 2-3 meters away) on the 5 leaves together that are somewhat brighter than the rest and approx. in the centre of the pic. Notice its the leaves beyond that that are the sharp bits, and the focusing chart should speak for itself.

I'm not having you on, and I've had the camera for 6 months and probably should have noticed the problem by now. Where it finally came to light was using a tripod and f2.5 aperture and the subject was just not quite right in any of the pics. I had him hold a CD with text on it about an inch or two in front of his face and I focused on the text. After that all the pics were perfectly focused.

OK. I was going on your statement about 1 inch in 6 yards.

G

But that isn't a precise measurement. A "held" item will be in reality waving about anywhere and everywhere. Your camera would need to be on a tripod and so would the CD for this to be meaningful.

Quite so.

I find it disappointing that when people understandably worry about whether it's their equipment or their technique at fault, the 'tests' they do to establish which it is are often very poorly designed, and allow no firm conclusions to be drawn.

I'm not saying that DXG is wrong, just that his tests are not conclusive.

Remember - the more unlikely the claim, the more rigorous the test needs to be to prove it beyond doubt.

In this case, the test is trivially simple to do, providing you're careful. Equipment required: chair, tripod, camera.

Ensuring that the focus is extremely accurately in the centre of the region desired is not, however, easy, since it's very easy for slack in the tripod or even the carpet to result in slight movement between the positioning and capturing the test image.

I had a go at this just now. Sadly, it seems I must continue to blame only myself when my pictures are unsharp:




Edit: I should have said, focus was set to spot, and I got the green hexagon when the very centre of the frame was aimed at the "this text should be perfectly in focus" part of the chart, which was stuck to a flat piece of card.

In DXG's case, the chart doesn't seem to be straight in the first one at least, and with such a short focal length lens, it would be a lot harder to point it accurately than with the 300mm I was using. I'd like to know a bit more about his setup before I believed the results.

Chris, I get very similar results as you with my FA 50mm 1:1.4 lens. The lower 2mm marker is sharper than the upper one.

More testing to do . . . using a cable release this time.

Not entirely convinced. The CA shows green below and magenta above, so the lower one looks sharper, with the magenta above making it look less sharp. But I think the green blur extends as far as the magenta one - or if not it's beyond the precision of the experiment. It's only a 600 dpi laser printer, and ordinary paper.

Having repeated the experiment with FA 50mm 1:1.4, FA 100mm Macro 1:2.8, FA 28-200mm lenses (all wide open) and a seismically rock solid tripod together with cable release . . . no focussing issues here. The 50mm produces slightly softer images wide open but that's expected from this lens. The 100mm Macro is tack sharp.

Interesting afternoon.