What's this edge softness?

Hi all

I wonder if someone could give me some opinions. Some of my photos seem to have what appears to be edge softness and I don't fully understand it. This photo was taken with a 40D and 24-105L lens at 1/640 @ f/8 at ISO200. It looks sharp enough except when you look at the golden spire right at the top of the minaret. This doesn't look sharp. Can anyone explain or give any comments? Thanks.

It appears it's a depth of field (DOF) limits issue. The closer/foreground elements appear to be very much in focus. So, as elements get further and further from the plane of critical focus, they are going to go more and more OOF. If you look closely at your image, you will note that it gets softer as you move up the building.

I got to agree, it looks more like a DOF issue than a lens focus issue. The spire is the farthest object from you.

I'm not really smart enough to know what's going on, but some observations:
The dome is starting to have a little softness, and the spire is just a little softer.
The sky is really noisy, I think that's contributing to the problem.
The histogram is fractionalized, by this I mean it's not a smooth curve anymore. Instead of a smooth curve, the curve is broken up into a bunch of vertical lines next to each other that still form the curve, but with gaps in it. These gaps suggest missing data to me. Missing data probably means some sort of softness.
I agree with the others here, the spire is the farthest from the lens and probably out of the plane of critical focus.
Maybe the edge of your lens isn't sharp, a lot of them aren't, but it certainly should be at f/8.

Thanks for the comments folks. I would have thought that f/8 was enough to cope with the necessary depth of field. Is this not so?

casaaviocar - can you explain what you mean about missing data etc? This was from a RAW file and was just converted for uploading to flickr. Would that account for the missing data? The out of focus spire seems to be the same on the original RAW file.

I believe there are two things working against technical sharpness toward the top of the frame. Let me say, I don't think it's a depth of field issue.
1 - the artifacting of the sharpening tool is evident and will undoubtedly create distortion at this level, especially in thinner lines.
2 - there is a slight bit of color fringing. Though loaded with 1-S-UD element and several Aspherical elements, the 24-105 does a good job limiting this phenomenon, but with any lens toward the edge of the frame it can be more noticeable in close inspection.
I recently calibrated the focusing on my 50D with my 24-105. When I looked real close, even when stopping down to f11, I noticed a slight bit of chromatic aberration. That is normally not noticeable, except when you look at it beyond 100% in PS.

Ed Harp wrote in post #9979773
I believe there are two things working against technical sharpness toward the top of the frame. Let me say, I don't think it's a depth of field issue.
1 - the artifacting of the sharpening tool is evident and will undoubtedly create distortion at this level, especially in thinner lines.
2 - there is a slight bit of color fringing. Though loaded with 1-S-UD element and several Aspherical elements, the 24-105 does a good job limiting this phenomenon, but with any lens toward the edge of the frame it can be more noticeable in close inspection.
I recently calibrated the focusing on my 50D with my 24-105. When I looked real close, even when stopping down to f11, I noticed a slight bit of chromatic aberration. That is normally not noticeable, except when you look at it beyond 100% in PS.

Thanks for those comments Ed. I too doubted that it was depth of field, simply because I would have thought that, at that distance, an aperture of f/8 would have covered the slight difference in the distance of the spire.

There was CA, but I thought I had cured it (in LR3). Do you think that it is CA that is causing the apparent softness in the spire? I don't think it can be general edge softness, as the picture seems to be sharp at the bottom and sides.

Another thought has just occurred to me re you comment about sharpening artifacts - I'll take all the sharpening off (again, in LR) and see if that makes any difference.

When I put a copy of this in LR the histogram had a lot of gaps in it. I've noticed this effect when I make fairly large exposure adjustments. There isn't enough data to stretch the DR out so the histogram shows gaps between the data.

HTTP response: 404 | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

The odd thing about this is that I didn't make any exposure adjustments!

It's not just sharpenign effects, but also slight overexposure towards the top of the tower, especially in the doem and spire.

Also, how much sharpness do you expect? The top of the spire is only 3 pixels wide. To define any details is really impossible in that case. If you look at the grass in the foreground just beyond the stairs you see exactly the same thing. Do not that at 100 % you are looking at a rather large magnification, and you should stay away from th epicture twice the diagonal for normal viewing.

It is also debatable how much DoF you get from digital. I reckon it is about 40% less than from film (on FF), and no DoF calculators take this into account (yet), although shooters all over now start realizing that this is the case. On APS-C it is only 50% of film on APS-C, but it is less noticeable there due to the automatic increase in DoF due to the smaller medium.

However, IMO, the main culprits are slight overexposure, sharpening artefacts, and the small size of the details.

HTH, kind regards, Wim

Thanks Wim. I've checked the original RAW - without any sharpening - and the difference is the same. Your point about over exposure is interesting, but the histogram does not go completely to the right - is it possible for it to be overexposed with the histogram touching the right edge?

Roy I know this is off target of your original inquiry but I noticed that your sky has a lot of sharpening artifact in it. If this is a problem a common solution is to mask off the sky before sharpening or select the sky and use Gaussian blur to reduce the noise.

Hi Roy,

Roy Mathers wrote in post #9980030
Thanks Wim. I've checked the original RAW - without any sharpening - and the difference is the same. Your point about over exposure is interesting, but the histogram does not go completely to the right - is it possible for it to be overexposed with the histogram touching the right edge?

Yes, especially if it is just slight overexposure, and only partial slight overexposure for that matter.

Kind regards, Wim

jetcode wrote in post #9980772
Roy I know this is off target of your original inquiry but I noticed that your sky has a lot of sharpening artifact in it. If this is a problem a common solution is to mask off the sky before sharpening or select the sky and use Gaussian blur to reduce the noise.

Thanks Joe

wimg wrote in post #9980932
Hi Roy,

Yes, especially if it is just slight overexposure, and only partial slight overexposure for that matter.

Kind regards, Wim

Wim - if it is only partial slight overexposure, would it make that much difference to the sharpness?