focus problem

I've been having a lot of trouble with the accuracy of focus on my Nikon D100. I'm pretty

certain that this is a technical issue rather than my own error, but I'm open to suggestions.

If I'm able to upload an example I will.

 

Basically, when focussing for a portrait, I always focus on the eye, reframe, and shoot.

What seems to happen regularly is that the photos look great until you blow up close, or

print. The eye is out of focus, but (for instance) the skin between the eye and the ear is

not.

 

I have paid in the past to have the auto focus checked and re-set, but the issue has never

really improved.

 

This is not an issue I had pre-digital and is *extremely* frustrating. Any thoughts or

advice welcome.

 

Angus<div></div>

Somewhere between the process of reframing and shooting, the focus has shifted.

 

Can you post a blown up (to the same extent as this digital capture) example from film where this has not happened?

Angus,

 

I have a D70 with similar problem. I have tested it using the method provided here:

 

http://md.co.za/d70/chart.html

 

I would suggests that you do this test. In my personal experience the kit lens has problems but other seem to work fine.

 

My D70 has been to Nikon twice under warranty but the problem persists, so I have learned to live with it.

 

Haris

Are you sure that your initial focus is locked in? Check your settings and your actions. It sounds like your camera refocuses after the reframing. Focus locking is often dependent on an appropriate camera setting followed by appropriate usage by the shooter. Some cameras lock focus by pressing the shutter release halfway down.

 

On a D70, the focus lock instructions are in the manual. If you are in AF-C mode, you have to also have to press the AE-L/AF-L button to lock in the focus. Most people do not need AF-C unless they want to shoot movement of some sort. If you are in AF-S mode, the lock occurs when you press the release halfway down AND keep it there. If you let up, the lock is gone unless you have pressed that AE-L/AF-L button. Does the D100 work the same way?

 

I hope I have not confused the issue.

 

There is also the possibility of subject movement. Stand up a book at an angle to you, focus on the spine, reframe and see if the focus shifts to somewhere on the cover. If not, your settings and technique are OK. Of course, in a test, the photographer may be more aware of proper procedures and vary from normal practice.

 

If the problem were the autofocus mechanism, you would see it on the shots you do not reframe.

 

Bill

There are three independent systems that should be calibrated to be the same: the actual image on the sensor, the image on the viewfinder, and whatever is seen by the AF system. Then there is the different issue of what the AF system locks onto.

 

Here are some things to try:

 

1. See if AF and MF agree in the viewfinder. AF on a flat subject with sharp lines, see if you can touch it up to look better. If so AF and viewfinder do not agree.

 

2. See if AF and image sensor agree. AF and expose, slightly tweak focusing and expose. Repeat several times. See if you can do better than AF.

 

3. Make sure the AF is actually locking onto the point shown in the viewfinder. (I have an N8008 that focuses on a spot above the rectangle shown in the viewfinder.) Set up a scene with a background and a small well defined foreground object. Play with the AF system, and find where the object has to be to get the AF to lock into it. You might discover that the AF is actually focusing on a point to the left of the rectangle. (What happens if you AF on the right eye (subject's left eye), do you get the nose sharp?)

 

4. Finally there is the issue of what the AF does when the region contains things at different distances. Try to set up a scene with things at different distances. (Maybe a flat target turned at an angle.) See what the AF focuses on. It might be that it goes for a point between the closest and farthest object in its senstive region.

 

5. If the D100 has different modes for AF that could conceivable affect this, try them all.

 

-- David Jacobson

Sounds like you may have a back focus issue. WHile this has been written about mostly regarding the D-70, I believe it can be an issue any dSLR, EVEN CANONS. There are some good web sites on how to check for this. I'd do a web search for back focus. If that's you're problem, Nikon can fix it for you.

They didn't fix your autofocus problem. You should send it back to the place where they were supposed to re-set the autofocus.

 

Have you ever tried to manual focus with the same lens and camera?

Test by focusing on something well lit & static that has very fine detail, use a tripod to make absolutley sure and use the timer or the remote release (Nikon ML-L3). Use ISO 200 and RAW. Have the object at a similar distance to your portrait subjects and use the same aperture. Remember that most DSLR images viewed at 'actual pixel' size will look a tad soft 'straight out of the camera' so just use your normal amount of post processing sharpening and then compare.

 

There is a firmware update coming up in mid May that is supposed to address one or two focussing issues so check the Nikon website to see if they are relevant.

 

Maybe (with your portraits) by the time you have locked focus then re-composed and taken the shot, you, the camera and the subject have moved enough to cause this. Perhaps a different focussing technique could work.

Forget the firmware upgrade, that was for D70. Sorry I forgot you said D100.

Technical issues are possible, I don't know about them. However, as Trevor suggests, movement of subject may cause this in extreme situations. Also as Vivek suggests, is it not conceivable that during reframing the distance of the region of focus from your camera changed? I would think this possible only when you are operating with a very shallow depth of field, such as at the close focusing limit of a telephoto (at wide apertures). In any case I guess you could at least eliminate this possibility by shooting the same subject by focusing using your normal method at the same point and not reframing.

As others have suggested, I'd check focus lock first. Most dSLRs have a bewildering array of options so it's easy to lose track of which function is set to which option.

 

As for backfocus, the simplest method I've seen is to set up a test target on horizontal surface (desk, countertop, whatever) and aim the camera at an angle. I used an ordinary desk calendar - plenty of contrast between the black horizontal and vertical lines on white paper. Set the lens to maximum aperture for shallow DOF - you don't want DOF to affect your reading of the results. Use a center spot focus setting and lock focus on a set of crossed black lines. If the resulting image shows the focus locked where you aimed it, all is well. If not, there may be a focus error.

 

When I first got my D2H I was concerned that it had a backfocus problem. I'd never had any complaints about focus with my Olympus C-3040Z P&S digicam, but it didn't have many options for in-camera sharpening. After getting the hang of the appropriate in-camera sharpening for the D2H, which varies according to the format (NEF vs. JPEG), along with careful post processing, I could only blame my crappy photos on myself.

I wonder if the CCD sensor "in-camera crop" is causing this to be more of a problem than it would be with a full-size 35mm imager (i.e., film)?

 

Is the smaller sensor size magnifying this "problem"? Leaving less room for focusing error.

One other possibility is that the area with most contrast difference gets focused (either during manual or auto focus). Hair on the head and the skin on the face offers this difference rather than the eye. This is less of a problem in film cameras with proper prism finders and a split focus screen.

 

The tiny viewfinders in cameras like D100 and D70 exacerbate this problem.

Vivek has a good point, tho' I'd think that if selective spot focusing was used there would be enough contrast between the iris and white of the eye. If a less selective focus method was used it would be fairly easy to misfocus.