[Viking42]

So I did some field tests the other day with my recently purchased (used) DA 12-24 lens on the K5ii. Using AF-S and center spot focusing, it seems that the lens is not latching onto distant scenes (e.g. 1 km or farther), but is focusing closer than infinity. A search on the forum on adjusting this came up blank -- I only found methods for setting closer distances, but not for infinity. Can someone point me in the direction of a good description of a method for this? BTW, I am aware of the in-camera utility for tweaking focus, so I know how to tweak it there -- I'm just looking for a field method, is all. I'm hoping there's a good step-by-step guide for this somewhere, rather than just trial and error tweaking.

FYI -- the lens AF seems to focus well on nearby subjects, so I am not having any issues with that - just infinity.

Many thanks in advance for any help.

---

Last edited by Viking42; 03-03-2021 at 01:31 PM.

[sergysergy]

Wide angle lenses are known for having trouble AF to infinity. Can you reach infinity if you focus manually? You need to determine whether it is missing focus or if it cannot reach infinity at all.

[Viking42]

Wide angle lenses are known for having trouble AF to infinity. Can you reach infinity if you focus manually? You need to determine whether it is missing focus or if it cannot reach infinity at all.

That's a really good question, and the answer is I don't know. But I will check some images and/or test it this evening and post back.

[stevebrot]

That's a really good question, and the answer is I don't know. But I will check some images and/or test it this evening and post back.

Attempt manual focus in magnified live view. That should provide you with confirmation as to whether the lens can attain far focus. Be aware that "infinity" is a slippery concept* and that that wide angle lenses provide significant challenge even with manual focus in magnified live view. Also note that AF fine adjust should not be done for far focus. Doing so has potential to screw up near focus badly.


Steve

* In practical terms, infinity is that distance where light from two closely spaced adjacent points arrive at the lens essentially parallel to each other. For wide angle lenses, that distance is often not that far away. An additional consideration is the mechanical limitations of the lens were on many wide angles, the difference between 10m and infinity is often less than 5 degrees of arc.

[Viking42]

Attempt manual focus in magnified live view. That should provide you with confirmation as to whether the lens can attain far focus. Be aware that "infinity" is a slippery concept* and that that wide angle lenses provide significant challenge even with manual focus in magnified live view. Also note that AF fine adjust should not be done for far focus. Doing so has potential to screw up near focus badly.


Steve

* In practical terms, infinity is that distance where light from two closely spaced adjacent points arrive at the lens essentially parallel to each other. For wide angle lenses, that distance is often not that far away. An additional consideration is the mechanical limitations of the lens were on many wide angles, the difference between 10m and infinity is often less than 5 degrees of arc.

Thanks for the input Steve. I just installed a new focus screen in the camera to help with manual focus, so between that and live view I should be able to tell if infinity focus is not possible with that lens. During my quick AF test this past weekend, I did check the focus ring on the lens to make sure it had moved out to the infinity mark, and it did so every time, right to the full-stop. I moved the ring to closer focus, then repeated the test twice. And yet the subject at infinity was out of focus, and nearby trees (~15 - 20 metres away) were in focus. I used a DA 50-135 lens on the same outing, did some long-distance shots of the same scene, and all was well -- no issues with that lens. This makes me think it's not the new screen or a general camera problem, but rather a camera-lens problem with the 12-24, or just a lens-only problem.

I'll post back once I've done a couple of quick tests at home later.

Cheers!

[kh1234567890]

Can someone point me in the direction of a good description of a method for this?

It is very simple. Camera on tripod, SR off, short timer release, centre point autofocus, wide open aperture. Find a distant contrasty object (at 12mm it won't have to be that distant), tree or a building, at the centre of the frame. Defocus the lens and shoot a series of shots, stepping fine tune by say 2 or 3 between shots and defocusing each time. Pixel peep on a monitor to find the best fine tune range. File size should also give you a hint, it will be largest for the most detailed frames unless things have moved or light has changed. Then repeat at single fine tune steps over the smaller 'best' range. Done.

AF on my DA12-24 on a K5ii is pretty reliable.

[Viking42]

Testing complete -- I took some shots of a bank of trees about 100 metres away from my balcony to check infinity focus; then another set of a print on the wall to check near focus (a high contrast graphic of some palm fronds, about 1.5 metres from the camera). Camera on tripod, self timer mode; aperture wide open; six shots each subject; first two were AF in spot mode; next two were manual focus through the viewfinder; last two were manual focus in LV, using 6x magnification. Here's what I found...

For the distant trees:
- Frames 1 & 2, AF, focus was off, trees not in focus; a car about halfway between camera and trees was in focus.
- Frames 3 & 4, manual focus in viewfinder, focus was worse; nothing was in focus; (this is very surprising as I am confident I nailed focus in the viewfinder).
- Frames 5 & 6, manual focus in LV, were spot on, perfect focus.

For the wall art:
- Frames 7 & 8, AF, focus was slightly off, print not perfectly in focus; close, but no cigar.
- Frames 9 & 10, manual focus in viewfinder, were again worse; focus was slightly off, but still worse than AF.
- Frames 11 & 12, manual focus in LV, were spot on, perfect focus.

A further note -- in all cases for the infinity shots (trees), the lens focus ring was not all the way over at full stop, even for the LV shots.

So what the heck is going on here? Clearly the lens is capable of infinity focus based on the LV shots of the trees. But why am I seeing perfect focus in the viewfinder but the shots are off, and in fact worse than AF? I thought I did a good job of shimming the new focus screen, but now I am questioning whether that might be the issue here. However, I am not having any such problems with other lenses, which baffles me. Perhaps I need to test them more thoroughly to see if they are off as well, in which case I need to check the focus screen shimming again and tweak. I mostly shoot at smaller apertures, so the DOF makes up for some error. But I still suspect it's the 12-24 lens that is problematic, especially when trying to hit infinity.

Any wisdom gleaned from the above tests is gratefully accepted. I'm scratching my head here a bit.

---------- Post added 3rd Mar 2021 at 20:38 ----------

It is very simple. Camera on tripod, SR off, short timer release, centre point autofocus, wide open aperture. Find a distant contrasty object (at 12mm it won't have to be that distant), tree or a building, at the centre of the frame. Defocus the lens and shoot a series of shots, stepping fine tune by say 2 or 3 between shots and defocusing each time. Pixel peep on a monitor to find the best fine tune range. File size should also give you a hint, it will be largest for the most detailed frames unless things have moved or light has changed. Then repeat at single fine tune steps over the smaller 'best' range. Done.

AF on my DA12-24 on a K5ii is pretty reliable.

Thanks - that is indeed easy and straightforward. Upon further pondering, it seems I have a bit more testing to do to rule out some other factors before I go ahead with that -- first of all to see if other lenses are also missing focus; and if so, do a more rigorous test of the new focus screen. If all is well on both counts, then I will use your advice and have a go at the 12-24.

[kh1234567890]

Assuming that you've shimmed the screen well (there are two drop down frames on the K5ii) also check for crud and dust on the AF sensor at the bottom of the mirror box.

[stevebrot]

I just installed a new focus screen in the camera to help with manual focus, so between that and live view I should be able to tell if infinity focus is not possible with that lens.

Live view is your gold standard unless you have confirmed the calibration of your new screen using live view (most require a change from the factory shims).

Edit: To elaborate, if you had changed out the screen on a film camera, you would confirm calibration and alignment by comparison to a ground glass at the focal plane in the film gate. Ideally, confirmation will be done using a fast prime, say a 50mm f/2 or wider maximum aperture. If your new screen is the S-type screen mentioned in other posts, this is doubly important due to the characteristics of that screen (see Canon Knowledge Base article on EOS Focus Screens LINK) not being recommended for use at viewing apertures slower than f/2.8.


Steve

[Apet-Sure]

Since your viewfinder manually focused shots are the worst, you may want to check the diopter slider on the viewfinder. Your eye auto-focuses as well, and if the diopter is off your eye will compensate and make the image look sharp. However, if manual viewfinder focusing is not a problem with other lenses, then this is probably not the issue.

[Viking42]

Live view is your gold standard unless you have confirmed the calibration of your new screen using live view (most require a change from the factory shims).

Edit: To elaborate, if you had changed out the screen on a film camera, you would confirm calibration and alignment by comparison to a ground glass at the focal plane in the film gate. Ideally, confirmation will be done using a fast prime, say a 50mm f/2 or wider maximum aperture. If your new screen is the S-type screen mentioned in other posts, this is doubly important due to the characteristics of that screen (see Canon Knowledge Base article on EOS Focus Screens LINK) not being recommended for use at viewing apertures slower than f/2.8.


Steve

Damn it Steve, I wish I'd read this before buying the S-type screen: "A Super Precision Matte screen such as the Ec-S is optimized for wideaperture lenses, specifically those that have a maximum aperture of f/1.8 – f/2.8....Note, however, that with lenses that have a maximum aperture smaller than f/2.8, a Super Precision Matte screen will appear dark and grainy" (from the Canon page you linked). That was the first thing I noticed when installing the new screen, was how much dimmer the S-type was vs. the stock Pentax screen. Given that several of my lenses are f4.0, including the 12-24, this may not be the best screen after all. Back in goes the original I guess....sigh.

But that begs the question, if the screen is too dark, how will it affect AF on a lens that has max aperture of f4.0? I get the manual focus bit -- too dark and all -- but how is AF affected?

---------- Post added 3rd Mar 2021 at 22:28 ----------

Since your viewfinder manually focused shots are the worst, you may want to check the diopter slider on the viewfinder. Your eye auto-focuses as well, and if the diopter is off your eye will compensate and make the image look sharp. However, if manual viewfinder focusing is not a problem with other lenses, then this is probably not the issue.

The diopter slider is OK -- I adjusted it after buying the camera, and confirmed it was still good after putting in the new screen. The screen etch marks are nice and sharp! But thanks for the head's up. I will check it again to make sure it hasn't been bumped out of adjustment.

---------- Post added 3rd Mar 2021 at 22:28 ----------

Assuming that you've shimmed the screen well (there are two drop down frames on the K5ii) also check for crud and dust on the AF sensor at the bottom of the mirror box.

Great tip -- thanks for that. I will check that out. As for the shimming, I swapped out the factory metal shim with a thinner plastic one that came with the screen. Focus was off with the factory shim, and bang-on with the plastic one.

[stevebrot]

But that begs the question, if the screen is too dark, how will it affect AF on a lens that has max aperture of f4.0?

It will not affect AF at all, neither PDAF (used with optical viewfinder) nor CDAF (used with live view). Since your focus screen may not a good point of comparison, don't include optical manual focus in your testing. Here is a protocol to use for evaluating your DA 12-24/4...no photos needed!

• Using manual focus and magnified live view (8x is about 1:1) with camera set on tripod, attempt manual focus on your distant subject at 12mm, 16mm, and 24mm focal lengths, evaluating at center of frame. Defocus between attempts. The lens should be able to attain manual focus at all three focal length settings. If not, there is a problem with the lens; otherwise, the lens should also be capable of auto focus to the same distance.
• Remaining in live view, switch to AF, center point and attempt AF of the same subject and at the same focal lengths as above, defocusing between attempts as before. If the camera signals success for each focal length, it is mechanically capable of focusing to that distance for those settings, at least to its satisfaction. If not successful, try with a higher contrast subject. If still unsuccessful, note the position of the distance scale on the lens; if at or near the distance to the subject, the lens is mechanically capable to that range, but CDAF is having trouble perhaps due to subject or lack of mechanical precision at that distance (note the short throw between 2m and infinity).

At this point, you may want to do some focus testing in live view by comparing photos taken wide open (center focus) to best manual focus attempt, say best 4 out of 5. If the live view AF results are all over the place for a subject with reasonable contrast, check the lens for internal haze or fungus.


Steve

[kh1234567890]

But that begs the question, if the screen is too dark, how will it affect AF on a lens that has max aperture of f4.0? I get the manual focus bit -- too dark and all -- but how is AF affected?

AF has nothing to do with the focus screen, auto exposure is affected though.

I have the Canon screen in my old K-7 body, it is great with old manual focus lenses (50mm f1.4's) and to a lesser extent with my little 300mm f5.6 mirror. I was going to swap it to my K-5ii body but decided not to - with slower, short focus throw, undamped AF lenses there is little advantage over using liveview.

In these Covid lock-down times, when I have taken shots of just about every tree and path in my local parks and walks, I've been shooting with a little Panasonic TZ100. Going back to Pentax AF now feels like a trip to the stone age ...

[Viking42]

It will not affect AF at all, neither PDAF (used with optical viewfinder) nor CDAF (used with live view). Since your focus screen may not a good point of comparison, don't include optical manual focus in your testing. Here is a protocol to use for evaluating your DA 12-24/4...no photos needed!

• Using manual focus and magnified live view (8x is about 1:1) with camera set on tripod, attempt manual focus on your distant subject at 12mm, 16mm, and 24mm focal lengths, evaluating at center of frame. Defocus between attempts. The lens should be able to attain manual focus at all three focal length settings. If not, there is a problem with the lens; otherwise, the lens should also be capable of auto focus to the same distance.
• Remaining in live view, switch to AF, center point and attempt AF of the same subject and at the same focal lengths as above, defocusing between attempts as before. If the camera signals success for each focal length, it is mechanically capable of focusing to that distance for those settings, at least to its satisfaction. If not successful, try with a higher contrast subject. If still unsuccessful, note the position of the distance scale on the lens; if at or near the distance to the subject, the lens is mechanically capable to that range, but CDAF is having trouble perhaps due to subject or lack of mechanical precision at that distance (note the short throw between 2m and infinity).

At this point, you may want to do some focus testing in live view by comparing photos taken wide open (center focus) to best manual focus attempt, say best 4 out of 5. If the live view AF results are all over the place for a subject with reasonable contrast, check the lens for internal haze or fungus.


Steve

Got it! Thanks Steve. I'm still learning about the inner workings of dslrs . Good to know that the new screen just affects manual focus - so if I can live with it's dimness with slower lenses I will keep it, because it is much better for manual lenses than the stock screen.

I will run your tests in the next day or two, as soon as I have a free bit of time again. I found a lens focus test chart out on the web, so I will print that off and do some more detailed closer-focus testing this weekend to confirm that the new screen is properly shimmed. And then there's the infinity adjustment posted by kh123..., as well as checking focus with my other lenses. Lots to do!

AF has nothing to do with the focus screen, auto exposure is affected though.

I have the Canon screen in my old K-7 body, it is great with old manual focus lenses (50mm f1.4's) and to a lesser extent with my little 300mm f5.6 mirror. I was going to swap it to my K-5ii body but decided not to - with slower, short focus throw, undamped AF lenses there is little advantage over using liveview.

In these Covid lock-down times, when I have taken shots of just about every tree and path in my local parks and walks, I've been shooting with a little Panasonic TZ100. Going back to Pentax AF now feels like a trip to the stone age ...

How is auto exposure affected? I had read that it was only an issue for screens with split image or microprisms. Mine is a simple matte screen with a few etched lines around the center area, which supposedly is fine for AE. Or did I read that wrong?

As for live view, I almost never use it unless I have to hold the camera away from my eye to get a shot. That's why I love using a dslr -- for the big bright pentaprism viewfinder . I really don't like composing on a screen - whether a rear one or EVF....doesn't feel natural. Especially EVFs....can't stand using them....they make my eyes hurt and give me a headache after an hour of shooting.

And yeah, I hear you about lockdown photography...by June of last year I think I shot every patch of ferns and moss in our local conservation areas

[stevebrot]

How is auto exposure affected? I had read that it was only an issue for screens with split image or microprisms.

A split image screen is pretty much useless for spot metering, but the dimming issue with the S-type affects all metering modes. The drop in screen brightness is precipitous when going from f/2.8 to f/4.0 (viewing aperture) with the meter indicating more than the expected +1 EV increase in exposure due to excess dimming. That said, if yours has the optional laser-etched lines, things may be different, but in a stranger way. The lines may continue to shine bright even as the screen dims. Both screw with the metering, when compared to live view.

I should mention that I gave a negative review of the S-type screen from focusingscreen.com on this site based on the metering behavior. I can temper that review by giving full support for use with faster lenses and no grid lines.


Steve