Thanks, everyone, for your thoughtful comments to the many questions posed on this site. I have learned a great deal about my D100, and photography in general.
I recently purchased a D100 and have found several images where the pic is sharp in the upper third and soft in the lower third when shot horizontally and sharp on the left while soft on the right when shot vertically. I'm using a Tokina 28-80 ATX Pro lens. Has anyone seen this type of problem before? I've wondered if maybe the CCD is slanted a bit, the lens is defective, or maybe the photographer is not quite right;)! I've followed the many good suggestions about shooting only RAW, using no sharpening except in Capture or PS, etc.... In fact, when the images are sharpened, the sharp areas get too sharp while the soft areas look great. I am going to try to attach two of the images where this problem is seen - notice my daughter at the lower part of the image compared to my son at the upper part, and in the other image my son's right shoulder sleeve compared to his face.
Thank you in advance for your comments and suggestions.
Do you know what aperture these were shot at? If it was wide open, the shallow dof is going to make this even worse.
To see if there is a problem with the lens or CCD, try taping a newspaper up to a wall, carefully squaring the camera to the wall, and taking some shots at wide open aperture. See if you can still see a difference in sharpness across the frame.
Either I'm going blind (can't really see a problem), or when you compressed the image dramatically (which you have to do to meet the Nikonians prerequisites), it's difficult to see what you are talking about.
As Victor asked, what aperture setting did you use? That would play a key roll. If you shot at let's say...2.8 or 3.5, then the first photo could be slight off because your daugher below is not necessarily on the same plane as your son on the top - albeit the differences may seem minute.
Both photos looked pleasantly exposed. Of course, I'm not looking at the original image that you have on your PC.
The pic with the four children is at f4.5, and the other is f2.8. The images are much more dramatic when looking at RAW/NEF.
I read some reviews on this particular lens which suggest it can be soft at f stops <5.6. This brings another question to mind - do you believe Nikkor glass is better than Tokina? (Sales reps at Ritz and B&H both suggested this lens over the Nikkor.)
I went back and shot a newspaper (per Victor's recs) at f2.8, 3.5, and 7.1, cropped and magnified. The camera was on a leveled tripod. The best comparison is at the f7.1, and I've included two images - one from the upper part of the original, and one from the lower part of the original. I did appply "sharpen" via PS to both images once, and applied some brightness (+10) to the lower image (I was using the built in flash which didn't seem to cover the lower part of the frame as well. Anyway, you (I) can see a subtle but definite difference - with the upper image a bit clearer than the lower image, though seemingly less pronounced than with the previous images that generated this thread.
Again, many thanks for your helpful comments
PS The upper image on the forum is actually the lower image of the original - sorry for the confusion - I uploaded them out of order.
Hmm.. that definetly is strange. One thing I notice from the top image (lower section of newspaper) is that the newspaper doesn't seem to be flush with the wall. The letters seem to skew slightly upwards...aren't as perfectly horizontal as the bottom image (top section). Are you sure the newspaper was flush against the wall? What focal length did you use? You might try zooming in...
The camera is about six feet from the wall with a focal length of 68mm. The newspaper is actually two pages taped to the wall, with the lower page slightly tilted upward (left slightly lower than right. I didn't notice this when taping them to the wall.) The portion of both pages captured in the magnified images is flush with the wall.
Does anyone else besides me see the difference, or am I crazy?!
I just got cable internet service today with a 25MB web page. If I can figure out how to use it, I'll try to upload larger images to that page and maybe it will show this better.
Very strange, Bill. You are definetly not crazy. I do see a slight difference. I tried to replicate your expirement, and I'm not sure what to make of the results. I have different camera settings, different light, different subject, different lens... so many variables it is hard to tell what to make of it. Here are my results. I think I also detect an ever so slight "softness" in the slice taken from the bottom of the image. Do you see it?
I'll have to go back and repro it with more careful lighting and leveling to be sure.
As a double check, try another lens. Does it have the same problem?
This is just a theory; I don't design digital cameras. If a CCD is at an angle or warped it could have different chromatic aberration across the CCD. I don't know if this would show up, but a possibility would be to stop down the lens and take pictures rotating the camera around an axis from front to back. If you see a difference and another D100 using the same lens doesn't I think this would show something wrong with the shape of the CCD.
Note: I made the above suggestion in case you really want to play with it yourself. It would probably just be easiest and best to have Nikon look at it if using another lens shows the same problem.
Another possibility, and more probable in my uneducated opinion, is that the anti-alias filter over the CCD could be slightly warped rather than the CCD itself, which probably undergoes more rigorous quality control. This is assuming that the filter is a thin sheet of some polymer based material. I would suspect the lens and the aperture setting first though. f/2.8 is going to make even the slightest deviation of the newsprint from the wall appear out of focus. I would try two things before faulting the camera body. Take a long exposure so you can stop down, and if that still shows problems, try another lens. A prime would be good, as they are crisper and faster.
The AA filter is a thin slice of Lithium Niobate (a hard but brittle crystalline substance) which is bonded to the front of the CCD; given its function it is the obvious suspect if you get the same result with different lenses, but given its construction it seems unlikely that it could be warped.
I may have missed it somewhere but are all these pictures taken with the same lens?
Have you tried other lenses and do you get the same thing?
If so and if you get the same result, I would definitelly rush to a Nikon service center since you may have a problem with one of many things (alignment of CCD (very unlikelly) problem with the filter in front (also unlikelly in my opinion) or possibly a problem with the lens mount and as a result the lens positioning). Regardless, this call for an immediate trip to your nearest service center.
I've seen a bunch of things on the Internet and with my own D100 but this is nothing that I've read about or seen
I finally had a chance to reshoot with a different lens, and it looks like the problem is in the lens. The new pics are evenly sharp, even at an f stop of 2.8. I shot with a Tokina 80-200 ATX Pro, 80mm focal length, f2.8, with a SB-80DX flash. The images are cropped and magnified to 200%. No sharpening applied, though I did add +2 exposure comp.
So, the question goes back to the origin of the problem. If it is the lens, what do you think could be causing the first lens to produce these uneven images? Also, would someone comment, in general, on the quality of Nikkor vs Tokina glass? Thanks again everyone for your comments and suggestions.
It could be many different things in a lens. One of the elements could be misshapen, it could be at a different angle then it should be etc.
It is uniformly said that a camera's manufacturer has better lenses then a third party like Tokina. So a Nikon lens would be considered better. Having said that, Tokina does make good lenses--good enough for all uses. If you want their best get one from their pro series.
"...would someone comment, in general, on the quality of Nikkor vs Tokina glass?"
Speaking from personal (though admittedly not recent) experience, Nikkor glass is much better. If I owned top-of-the-line gear like the D100, IBM 1gig microdrive and SB80DX flash, the last place I'd economize would be with the lenses.
Just a personal opinion, no offense meant to users of other lenses.