[ychousa]

Hi, I'm wondering if my lens has been calibrated correctly with my K-3. Could you check the images and let me know what you think? Based on previous analysis, I've calibrated my 35mm f2.4 to be +5 in AF fine adjustment-the following two shots are the result after +5 adjustment. Not sure if it's other issues, but I find some not-so-nicely focused shots some times even after calibration. (Example shot with some bad focus at f11 below-I believe I used center focus with re-composition.) Should I try another calibration method? Could you share your thoughts? (For the test chart shot, I used a tripod at 45 degree with a 2-sec delay.)

Thanks for your help!

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[Rimfiredude]

Looking good to me. You have a little more depth of field in the rear of the picture. Look at the 10s. The one in the rear is just a tad sharper.

[ychousa]

Third pic is a 100% crop and shows pine trees on the left not in focus. This is just an example and I find some focusing problems.

---------- Post added 02-14-15 at 12:33 PM ----------

Looking good to me. You have a little more depth of field in the rear of the picture. Look at the 10s. The one in the rear is just a tad sharper.

Thanks for your reply! In that case, what number would you think is appropriate in AF adjustment? (Currently +5.)

[MadMathMind]

In regards to the pine tree example:
1) Where did you aim your focus? The focus points are rather large, although less so on the K-3.
2) You shot at f/11. The diffraction limit for a 24MP camera is actually around f/9, so the lack of sharpness may actually be due to that rather than lack of focus. Plus, the lens probably doesn't match the resolution of the K-3 either. To my eyes, your issue just looks like lack of resolving power in such small fine details, not lack of focus.

[ak_kiwi]

Is my lens in focus?-sample shots included

Focus adjustment looks pretty good to me going by the top and bottom edges of the 14mm mark. You could also check for possible de-centering by shooting a sheet of newspaper parallel to the sensor plane and comparing edges/corners.
The photo you posted may be soft on the left hand side because your focus point is too far into the scene. For the f11 used I came up with a hyperfocal distance of 7.35m 5.6m giving a near limit of about 6.54 3m and infinity.
Do you have a link to the full size file of this shot?

[jpzk]

The diffraction limit for a 24MP camera is actually around f/9

Interesting post/thread.

How do you end up with this figure?
Does it mean that diffraction is lessened while closed down (stopped down) when using a 16MP sensor instead?
Just curious.

JP

[ak_kiwi]

2) You shot at f/11. The diffraction limit for a 24MP camera is actually around f/9, so the lack of sharpness may actually be due to that rather than lack of focus. Plus, the lens probably doesn't match the resolution of the K-3 either. To my eyes, your issue just looks like lack of resolving power in such small fine details, not lack of focus.

Good point MadMathMind.
ychousa - here is the Photozone resolution graph for the DA 35 f2.4 (on a K-5) if you havn't seen it. (I guessed the lense based on the exif and CA on the focus test - but don't worry about the CA since my 31 Ltd does the same thing)

Attached Images

 

[MadMathMind]

Interesting post/thread.

How do you end up with this figure?

There's a tool online to do it: Digital Camera Diffraction – Resolution, Color & Micro-Contrast

Scroll down the bottom. It's just above the CONCLUSIONS section.

Does it mean that diffraction is lessened while closed down (stopped down) when using a 16MP sensor instead?
Just curious.

JP

Yes. The smaller the pixels, the smaller the aperture for the diffraction limit. For an APS-C (1.5x) 16MP sensor, the limit is around f/11. This is about the same as what it is for a 36MP full frame sensor, by the way.

There are multiple levels of "diffraction limit." I use the "standard grayscale" because it most closely mirrors the luma channel of the image, where most of the detail is concentrated.

It's also worth noting that diffraction limiting is gradual. It may start at a given f stop but it may not affect the final image. It also depends on content. Shooting something with really fine details from far away (like in this example) is inherently not going to have the sharpness of something closer, although that has more to do with the resolving power of the lens than anything else.

[ychousa]

1) Where did you aim your focus? The focus points are rather large, although less so on the K-3.

I believe I aimed at the far mountain on the center.

2) You shot at f/11. The diffraction limit for a 24MP camera is actually around f/9, so the lack of sharpness may actually be due to that rather than lack of focus.

Does that mean any aperture beyond 9 would cause diffraction? I'm a bit confused.

Plus, the lens probably doesn't match the resolution of the K-3 either. To my eyes, your issue just looks like lack of resolving power in such small fine details, not lack of focus.

OK, this is a little tough topic for me to understand :-) Could you elaborate a little more for easier understanding? Should I choose another lens such as 40mm limited instead of this 35mm?

---------- Post added 02-14-15 at 02:01 PM ----------

Focus adjustment looks pretty good to me going by the top and bottom edges of the 14mm mark. You could also check for possible de-centering by shooting a sheet of newspaper parallel to the sensor plane and comparing edges/corners.

If I find a decentering, does that mean the lens is defective? (Sorry a novice here )

For the f11 used I came up with a hyperfocal distance of 7.35m giving a near limit of 6.54m and infinity.
Do you have a link to the full size file of this shot?

Well, the left pine tree was surely more than 7mm apart from shooting point if that's what you mean. And I've uploaded the DNG file in the following link. Hope the link works.

https://drive.google.com/folderview?id=0B6lZi8NF91BxfnpneTNXVl9hN05EMDdONXJm...Fk&usp=sharing

[ak_kiwi]

I believe I aimed at the far mountain on the center.

See this link to give you an idea show where you should be focusing for a shot such as you posted.

---------- Post added 15-02-15 at 11:29 AM ----------

If I find a decentering, does that mean the lens is defective?

If you get a noticeable difference the lense may be defective but I don't think that is the problem.

Well, the left pine tree was surely more than 7mm apart from shooting point if that's what you mean.

You should be focusing (corrected info from other post)on a point 5.6 meters from the camera (camera sensor plane to be exact) to get everything from 3 meters to infinity in acceptable focus.

[ychousa]

See this link to give you an idea show where you should be focusing for a shot such as you posted.

---------- Post added 15-02-15 at 11:29 AM ----------



If you get a noticeable difference the lense may be defective but I don't think that is the problem.



You should be focusing (corrected info from other post)on a point 5.6 meters from the camera (camera sensor plane to be exact) to get everything from 3 meters to infinity in acceptable focus.

Thank you so much! I'll try the suggested way and see how it improves.

[Bruce Clark]

Third pic is a 100% crop and shows pine trees on the left not in focus.

In general terms. trees are not a good subject to check focus as they move, even in the slightest breeze. Far better to test in streetscapes.

[MadMathMind]

OK, this is a little tough topic for me to understand :-) Could you elaborate a little more for easier understanding? Should I choose another lens such as 40mm limited instead of this 35mm?

It's not a knock on the lens. The DA35 has very good resolution. But sensors now have a lot more than the lenses do. Basically, you can't expect that any lens is going to provide "24MP" of detail. In other words, when you zoom to 100%, there's going to be some blur or other loss of fine detail. If you view it 50%, everything may look sharp, however, because that may be the limit of the lens's power to resolve.

You will also lose detail as you move farther away from the target. The atmosphere is not clear. There's dust, pollution, etc. in it. The more air between you and the subject, the more that is going to affect the photo because light is scattered by all those particles (including the air molecules) as it moves through the air--it's one reason why closer objects appear sharper to you and very distant objects will always appear hazy.

Time of day also affects things; it is generally clearer in the morning than the afternoon. If you want to see this effect in extreme, go to a high place and take a photo in the late afternoon. It will look something like this:



The subject here was taken from 550 feet in the air (top of WaMo) from a half mile away. The effects of late afternoon Washington pollution are incredibly obvious.

In your example, it's not as extreme, but if you're looking for your lens to resolve individual leaves from 100 yards away...that's just too much to ask.

[ychousa]

That makes total sense. A little bit confusing thing to me is that it looks like the lens front-focuses in some cases even with the chart being shown correct. I took some pics this afternoon. The first one within 4~5 feet and second one in around 10~15 feet. All two shots focused on my dog's eyes. The second one shows it's front-focused on a bit left front side grass. I took about 10 identical shots in each location and all of them show the similar results-farther shots not in focus while close shots are well-focused. Should I add a couple more number to make AF find adjust to let's say to +6 or +7 from +5? I'd like to hear some opinions. Thanks!

[ak_kiwi]

One of the things I too found with my K3, when compared to the K-5 series cameras, was it really showed up small errors in focusing. After many hours shooting my trusty old focus target (and cat dogs and kids etc) I came to the conclusion that the targets I was using were just not good enough to give repeatable accurate results. I found it was imperative to have a focus target parallel to the sensor plane - and the ability to confirm this parallel was critical to developing accurate evaluations of tests. In the end I purchased a LensAlign focus target which in about 5 minutes told me what I needed to know to fix the problem.

However,just before you buy a LensAlign - try one of these targets on a repeatably well lit wall. Square off the camera to the wall and make sure it is on center with the target also. Test your close auto focus Vs further away auto focus by comparing with shots using live view (with focus peaking) and manual focus.
If you want to get real keen arrange 4 more targets to mark out the corners of the frame. This will give you a check on possible decentering and replaces the "newspaper check" method.