[bxf]

But why do you press the SR button if you dont like the result?

If this is somehow relevant, I'd like to know how. Please explain.

There's the the Live View, you can zoom it and see exactly what you have in focus and how sharp it will look. Did the image look good before you pressed the SR? Why not? Or what went wrong after in case if it did.

Yes, LiveView seems to provide better focus, but I don't like to use it much. I find it easier to look through the viewfinder, especially since I don't walk around with my reading glasses, which I'd need in order to see the LCD properly. Also, using LV would be a workaround, not a solution to the problem.

BTW Are you sure the Tamron is well calibrated and focusing properly?

Well, now we're getting closer to my main problem. Am I sure? No. I've performed several tests, and I've had difficulty in achieving consistency. However, I did finally settle on AF adjustment of -5, which gives me reasonable results and consistency indoors, under specific lighting conditions. However, these do not appear to hold when shooting outdoors. Also, my zooms appear to require very different AF settings at the wide and tele focal lengths. I'm not sure, at this stage, if I should be blaming the camera or the lenses.

For sharp image you need good light. Check under good conditions if you get sharp images.

I'm not being overly demanding with these snapshots, so when I see something bad, it's an observation made without pixel peeping and without being overly critical. Some of my shots are bad even by P&S standards.

[bxf]

f 9.5 corresponds to an airy disk diameter of about 12.75 micrometers. The pixel spacing on the K-5 is 4.8 micrometers.

At 100% display each diffraction spot is occupying at least 2-1/2 display pixels therefore it looks soft. You are viewing it from about 1/2 meter I suppose. View it from a couple of meters back and it will look much better - or display it at about 40%.

The problem is that "100% display" implicitly assumes resolution is limited by the sensor's pixel pitch but in this case the limitation is likely the diffraction spot size because of the high f-stop relative to the camera's pixel pitch.

Hmmm... the pixel spacing on the K20D is only a little more than the K-5 so unless they have differing AA filter strengths one would expect about the same diffraction softening at f:9.5

F:diffraction.onset = Pixel.Pitch/(2.44wavelength) ~ 3(sensor.width.mm/sensor.width.pixels)/4

Roughly speaking you can expect to see diffraction softening at 100% when the f stop is greater than the sensor's pixel pitch in micrometers.

My primary intention in this thread was to tackle the inconsistencies I get with AF, some of which defy any attempt to define a pattern of behaviour. Based on some of the comments in this thread, I am able to conclude that 1. SOME of the softness I've attributed to bad focus is actually lens softness/diffraction, and 2. my expectations are unrealistic with respect to what a lens can do at a certain FL and subject distance.

Basically, you are confirming that there's something wrong with my K-5 shots. When I look at my K20D shots, I see reasonable (relatively speaking) results, even with the lens wide open. I don't see anything like this K5_S2654.JPG - File Shared from Box.net - Free Online File Storage which I seem to be getting a lot with the K-5. I know the sky is blown, but that is easily corrected in ACR. The face and hair, however, are very unsharp. Some of the background, again, seems more in focus than the subject. I'd really like to understand why. I suppose the easy answer may be that I'm using incorrect AF Adjustment setting, but then I'd have to wonder how one is supposed to cope with lenses that have AF Adjustment requirements that are something like 14 steps apart between the wide and tele settings.

[Dr_who]

Sorry for the delayed response, but I've spent all this time scrutinizing the circled areas, trying to locate the dust. My screen and my glasses have more visible spots Just kidding. I do see something but very faint, and I never would have thought anything of it. The spots in the link you provided are certainly more obvious. I'll try Dust Removal, and keep my eyes open. Thanks.

I wouldn't have noticed it either had I not been fooling around with the original sized raw. Also your image was shot at f9.5 and 1/500 so certain settings make it more noticeable. Set your camera to f22 and take a picture of a clear sky or a flat white or off white wall. That makes it easier to spot. In the end if you don't notice the spots. It doesn't take away from the photo in this picture and it is indeed faint like you said.

[jstevewhite]

My primary intention in this thread was to tackle the inconsistencies I get with AF, some of which defy any attempt to define a pattern of behaviour. Based on some of the comments in this thread, I am able to conclude that 1. SOME of the softness I've attributed to bad focus is actually lens softness/diffraction, and 2. my expectations are unrealistic with respect to what a lens can do at a certain FL and subject distance.

Basically, you are confirming that there's something wrong with my K-5 shots. When I look at my K20D shots, I see reasonable (relatively speaking) results, even with the lens wide open. I don't see anything like this K5_S2654.JPG - File Shared from Box.net - Free Online File Storage which I seem to be getting a lot with the K-5. I know the sky is blown, but that is easily corrected in ACR. The face and hair, however, are very unsharp. Some of the background, again, seems more in focus than the subject. I'd really like to understand why. I suppose the easy answer may be that I'm using incorrect AF Adjustment setting, but then I'd have to wonder how one is supposed to cope with lenses that have AF Adjustment requirements that are something like 14 steps apart between the wide and tele settings.

Honestly, I'd really like to see a K20D shot at ISO 1600, f8, 180th with the same subject size that is "sharper". At 100%, I can make out the jump-ring that holds the "B" on the necklace; I can see what is eyebrow and what is pencil; I can see the lip piercing, etc. The white appears a little overdriven for the scene, and it looks like the plane of sharpest focus might be about at her left collarbone (see the edges on the purse peeking through near her belt). Poking this image through an unsharp mask cycle in photoshop would make it look sharper.

Now, this is not the same as saying you're not having focus issues; I just don't think this is necessarily one. This is a pretty sharp image for the circumstances.

I have one zoom lens from Sigma that is similar to what you describe; that is, if I set the autofocus fine adjust at the wide end (18mm), it's off at the 50mm end, and vice versa. The solution everyone has given me is "Send the camera and lens together for calibration", but since it's just the one lens, I don't really wanna mess with it - all my others work properly, including two other Sigma zooms and all my Pentax glass. The 18-50 still works properly on my K20d, though. Odd.

[newarts]

I suggest you do a tripod test with the K20D and the K-5 with the same lens on a target like a tilted newspaper for a sequence of f-stops.

Then you'll know for sure about relative sharpness, focus etc. I suspect f:5.6 might look best at 100%.

[bxf]

Honestly, I'd really like to see a K20D shot at ISO 1600, f8, 180th with the same subject size that is "sharper".

OK, you got me there. Although I do have some clear and sharp K20D shots at ISO 800 and 400, that's as far as it goes. However, your comment did wake me up.

I just spent literally hours reviewing my photos, in an attempt to narrow down the things that displease me. One thing that became obvious is that I've "forgotten" that I was quite careful with my ISO settings with my K20D (now gone), rarely exceeding 400. With the K-5, I've gotten carried away, and started shooting a fair bit at 1600 & 3200, forgetting that "better" does not mean "excellent". This is a point you made earlier (post #7) "it's a matter of usable images, not the best possible images". The fact is that none of my K5 shots at ISO 400 or less are horrible (at least within the context of this thread ). Fine, this use of high ISO explains quite a few of my bad images.

Now, this is not the same as saying you're not having focus issues; I just don't think this is necessarily one. This is a pretty sharp image for the circumstances.

I have one zoom lens from Sigma that is similar to what you describe; that is, if I set the autofocus fine adjust at the wide end (18mm), it's off at the 50mm end, and vice versa. The solution everyone has given me is "Send the camera and lens together for calibration", but since it's just the one lens, I don't really wanna mess with it - all my others work properly, including two other Sigma zooms and all my Pentax glass. The 18-50 still works properly on my K20d, though. Odd.

Rather than bore you with more images, let me state with full confidence that I have consecutive shots, taken seconds apart, where #1 & 3 are reasonably focused, and #2 is $hit. I have this problem with more than one lens. I certainly had some badly focused shots with my K20D, but there was more overall consistency.

So, I am satisfied that I've found explanations for a good percentage of my fuzzy images. What remains is the inconsistency in the optimal AF Adjustment across the zoom range. Aside from the possibility (wishful thinking here) that FW 1.10 resolves some of these issues, does one blame the camera, or the lenses? If it's the lenses, is this a fixable problem? Would Tamron or Pentax consider this to be a problem that they should repair?

I sincerely appreciate your having taken the time to add your comments. Thanks.

[bxf]

I suggest you do a tripod test with the K20D and the K-5 with the same lens on a target like a tilted newspaper for a sequence of f-stops.

Then you'll know for sure about relative sharpness, focus etc. I suspect f:5.6 might look best at 100%.

K20D is gone.

I've done several sets of tests trying to determine optimum AF Adjustments. Aside from the difficulty I encountered in achieving consistent results during the tests, I finally managed to come to some conclusions. Unfortunately, a couple of my lenses show a very wide gap between the wide & tele settings. Also, I believe I've permanently screwed up my eyes in the process So much so, that when I look at my test results now, I ask myself how I came to the conclusions I did I'm hoping FW 1.10 will show some benefits in this area. Thanks again for your input.