[James78]

It was not clear from your description whether this is still the case when you are focusing both near-to-far and far-to-near? The reason I ask is that there is often a fairly large amount of focus ring...
Steve

Thanks Steve. Yes I meant both focusing far-to-subject and near-to-subject end up ALWAYS focusing ahead of actual subject with no exception on both manual lenses. Now I understand I cannot expect accurate focus on fast apertures but I thought I could expect at least 1 reasonable (read non-blurry) focus out of 50 attempts

I spent few hours exercising this in different ways including putting subject against tape measure to detect where exactly focus point is. For 85mm on f1.4 and subject distance around 2 metres focus confirmation range is around 30-40cm ahead. With AF Fine Adjustment I could reduce this but still not get to subject.

Initially I thought my Ronikon has major calibration issues with Pentax but now I purchased Pentax-M 50mm f1.4 with similar (but slightly better) results which proves it's sine thing inside the body.

I must say I just recently bought this body from eBay Japan as used with shutter count 17k. Just saying I don't have history of this K5 body.

---------- Post added 09-19-14 at 07:22 AM ----------

Calibration? Again, that's of the camera's AF settings, not the lens.

I know I sound like I'm not putting enough thoughts on responses so far there's something that really doesn't make sense to me. I try things on carpet because I can evaluate exact focus point on carpet texture, put small objects every 10cm, and also tape measure so that I could see where peak focus point goes. I have already noticed it's inaccurate which is the point you're making and I confirm so. But we seem to be discussing focusing issues at whole different levels. My out of focus problem is massive. So many attempts with different scenarios and ALL of them with zero exception cone significantly ahead of subject. Even at f2.8.

As I said I developed the skill to focus with hexagon light and then blindly focus slightly further (with practice I find how much) and then every time I get better focus. How can this be explained knowing that the basis of my manual focus trick is hexagon light yet I get better results. So we have inaccurate focus by design that has been discussed here plus another bigger focusing issue that apparently only I have.

Interesting my DA 50mm which is autofocus works sharp. Multiple other FA lenses that I have work good too. Only on manual focus things go very wrong.

Maybe I should upload an example picture to demonstrate the level of out of focus issue I have so you can evaluate whether it's by-design inaccuracy or something bigger.

---

Last edited by James78; 09-18-2014 at 02:24 PM.

[stevebrot]

Ronikon

Is that the same as a Rokinon?

So we have inaccurate focus by design

Inaccurate is not quite the right word. The AF system lacks adequate sensitivity for lenses faster than f/2.8. This can cause poor precision (the ability to get the same focus point twice). An accurate calibration can still be very imprecise. Similarly a very precise system can be very inaccurate.


Steve

[James78]

Is that the same as a Rokinon?

Yes, I meant Rokinon

Inaccurate is not quite the right word. The AF system lacks adequate sensitivity for lenses faster than f/2.8. This can cause poor precision (the ability to get the same focus point twice). An accurate calibration can still be very imprecise. Similarly a very precise system can be very inaccurate.

Fair enough. I hope I could explain what the real problem is.

---------- Post added 09-19-14 at 10:46 AM ----------

Ok, I just tried an SMC Pentax-FA 50mm F1.4. I already have DA 50mm F1.8. I put both on F1.8, ISO 100, and shutter speed 1/100, then use auto-focus and take shots. I repeat this at-least 5 times for both lenses by trying to focus at closer/further objects and automatically re-focus on the subject at 3 metre distance. Once again, I get consistently out-of-focus results with FA whereas DA gets accurate focus. The Pentax-FA 50mm gets around 20-30cm closer focus making subject out-of-focus which is exact same behaviour with the other manual lenses I have. Now I'm getting even more confused. But definitely this cannot be focus sensor sensitivity otherwise how come it perform good with DA but hurrible with FA? Please note this time I simply used auto-focus.

---------- Post added 09-19-14 at 11:07 AM ----------

This is cropped sample of Pentax-FA 50mm at F1.8 vs Pentax-DA 50mm at F1.8. Distance to subject is around 5 metres.

Pentax-FA 50mm at F1.8
Pentax-DA 50mm at F1.8

As you can notice even the umbrella ahead of subject is not in focus because the actual focus for Pentax-FA is around 1 metre ahead of subject. This seems very off limit and I don't think it is a focus sensor sensitivity conversation anymore.

AF Fine Adjustment is off and I am using centre focus only. Doesn't matter how many times I repeat this I will get very similar results.

[stevebrot]

This is cropped sample of Pentax-FA 50mm at F1.8 vs Pentax-DA 50mm at F1.8. Distance to subject is around 5 metres.

Both are front-focused. Either that or the focus point was set on the wooden object at lower left. Unfortunately Photoshop stripped the maker notes so the exif was not as useful as it otherwise might have been. Can you drop the original file to your dropbox folder and post a link?


Steve

[James78]

Both are front-focused. Either that or the focus point was set on the wooden object at lower left. Unfortunately Photoshop stripped the maker notes so the exif was not as useful as it otherwise might have been. Can you drop the original file to your dropbox folder and post a link?


Steve

Thanks Steve for looking into this. I just put multiple lenses and run tests again. All photos taken with 1/100 F2.5 ISO100, except the 85mm all are 50mm. For those with manual focus I tried both focusing from near-to-far and then far-to-near and in both cases I shot right at the edge of focus confirmation to get a sense of both boundaries.

Here's list of original files as below:
Pentax-DA 50mm - autofocus
Pentax-FA 50mm - autofocus
Pentax-M 50mm - manual focus near-to-far
Pentax-M 50mm - manual focus far-to-near
Rokinon 85mm - manual focus near-to-far
Rokinon 85mm - manual focus far-to-near

[stevebrot]

Thanks Steve for looking into this. I just put multiple lenses and run tests again. All photos taken with 1/100 F2.5 ISO100, except the 85mm all are 50mm. For those with manual focus I tried both focusing from near-to-far and then far-to-near and in both cases I shot right at the edge of focus confirmation to get a sense of both boundaries.

Here's list of original files as below:
Pentax-DA 50mm - autofocus


Pentax-FA 50mm - autofocus


Pentax-M 50mm - manual focus near-to-far


Pentax-M 50mm - manual focus far-to-near


Rokinon 85mm - manual focus near-to-far


Rokinon 85mm - manual focus far-to-near

Thanks! The FA 50 and DA 50 shots are reversed. The one labeled FA 50mm (taken with DA 50/1.8) is better than the others, but still does not look right. The art work and frame are not sharp. DOF at that distance and f/2.5 is a full meter with the near limit being 45 cm in front of the point of focus, yet the threads on the umbrella have about the same sharpness as the pressed wood border on the artwork. It appears the point of focus is at about the plane of the metal tip of the umbrella.

The shot with the FA 50/1.4 at f/2.5 is definitely focused at or near the tip of the umbrella, but the frame is softer over all, possibly due to LaCA visible throughout much of the frame or camera motion. The Rokinon 85/1.4 at f/2.5 shows similar front focus and some evidence (I think) of camera motion. Is it safe to assume these shots are hand-held?

The M 50 shot is simply nasty with no apparent point of focus within the field.

It is likely that you have got a significant calibration or other issue with your AF system that is not limited to manual focus. I suggest we step back a little and do a bit of basic testing. Here is the setup:

• Camera on tripod
• Drive mode on 2s delay. This disables shake reduction should you want to take a test exposure.
• High contrast focus target taped to a wall. I use a lens test target* printed on glossy photo paper.
• Focal plane of camera 20x the lens focal length from the target with care taken to make sure that the focal plane is parallel to the target
• Moderately bright natural light. The AF system has issues with some types of artificial lighting.

What we have done is to make it easy for the camera to attain focus on an unambiguous target without concern for target topography or AF point size or target contrast. Now to do a quick evaluation:

• Mount your DA 50/1.8 to the camera. Place a piece of tape on the focus ring to allow you to mark the focus point.
• Switch to manual focus and rack the focus to infinity
• Slowly move the focus ring until the AF system beeps and the green hexagon displays
• As soon as focus is attained remove your fingers from the lens and place a pencil mark on the tape immediately opposite the "1" in "DA 1:1.8 50mm" on the lens body.
• Switch to magnified live view
• Attempt to obtain a sharper image using live view and write down whether you turned the focus ring to the right (towards infinity) or to the left to make the correction. It makes no difference how far.
• Repeat several more times (at least 5 total) focusing from infinity to 1 meter, noting the direction to correct each time
• Do the same number of attempts working from minimum focus distance to one meter, writing down each time the direction to correct.

The tally should clearly indicate whether you have detectible front or back focus with that lens.

The reason why I suggested the DA 50/1.8 is that it is a relatively new, sharp, and high contrast lens. We want to give the AF system every advantage. You can repeat the test now at a longer distance (1.7 meters) with your Rokinon 85/1.4. With any luck, you will see a similar pattern as with the DA 50.

In regards to your FA 50 and M 50 lenses. I would suggest that you do a careful inspection of both lenses in dim light using a small flashlight (torch) shining obliquely through both the front and back of the lens while observing for fungus or gross damage to the lens elements. Your example image for both those lenses did not look right**, regardless of focus and it would be good to make sure the optics appear clear and clean inside and out. There may be a physical reason.


Steve

* Camera Lens Testing- Sharpness, Chromatic Aberration and Distortion - Bob Atkins Photography

** My M 50/1.7 is marginally sharper than my DA 50/1.8. I would expect yours would have similar performance when compared side-by-side.

[clackers]

I suggest we step back a little and do a bit of basic testing. Here is the setup:

Amazing, Steve. Am putting all that into Evernote. (thumbup)

[stevebrot]

Amazing, Steve. Am putting all that into Evernote. (thumbup)

Thanks.


Steve

[James78]

Thanks Steve very much for the detailed response. I will spend time to follow steps you mentioned and I'll get back with results.

[James78]

Hi Steve,

Here's results all for 1/160s at f1.8 ISO100 in cloudy Australian morning at my sunroom:

• Pentax-DA 50mm focusing infinity-to-subject in order to get sharper focus with LV I turned focus ring at 5 attempts: Right, Left, Left, Left, Same - Inconsistent mainly because sharpness was fine and I won't trust my eyes on LV that much. I could upload samples if you like.
• Pentax-DA 50mm focusing from near to subject: Right, Same, Same, Right, Same

Since first focus was reasonably fine and my LV sharpness improvement wasn't consistent I decided to continue testing on other lenses. Here's results:

• Pentax-FA 50mm: Right, Left, Right, Right, Right and difference was noticeable - I didn't try near-to-subject though
• Rokinon 85mm: Every single time was Right both for infinity-to-subject and near-to-subject. This time huge difference and I had to move focus ring significantly to compensate for better sharpness in LV.

I could upload samples if you like. For now let me share one viewfinder/LV of DA50mm followed by focus confirmation sample of Rokinon 85mm.
Pentax DA-50 - focus confirmation
Pentax DA-50 - Live View
Rokinon 85mm - Focus Confirmation

I don't have a sample of Rokinon using Live View but I could clearly get much better results from Live View.

[James78]

As I'm waiting to better understand what exactly is wrong I also googled and found this: K5 Back Focus issue solved: Pentax SLR Talk Forum: Digital Photography Review

Basically the -10 to +10 calibration limit can be changed which means worse coming to worse I can increase that limit and go for something like -15 or -20 to calibrate. My Rokinon couldn't be calibrated because default -10 to +10 was not enough. As long as I know K5 body is not faulty. I am also expecting my DA 70mm 2.4 and will give that lens a go too.

[stevebrot]

It looks like you are making progress.

I was wondering if you could edit your comments above to link to the high resolution images rather than displaying them. The time to load the page is excessive.

I am surprised that your Rokinon has such a large bias.


Steve

[rawr]

to link to the high resolution images rather than displaying them

Amen. Putting a series of large 5000x3000 pixel, 7MB+ JPGs directly into this thread means to obtain a single page view, I am loading over 30MB of large JPEGs each time. It's painfully slow, and very wasteful of bandwidth.

Post a download link for the images, not the images themselves.

[kh1234567890]

Isn't it a bit futile to expect the phase detect AF which on the K-5 has an effective aperture of f/5.6 to work accurately with a f/1.4 lens ? A decent focusing screen, properly shimmed and skilled manual focus are the only answer. It is certainly possible with the M50/f1.4

[James78]

Isn't it a bit futile to expect the phase detect AF which on the K-5 has an effective aperture of f/5.6 to work accurately with a f/1.4 lens ? A decent focusing screen, properly shimmed and skilled manual focus are the only answer. It is certainly possible with the M50/f1.4

I am suspecting my K5 having major calibration issues. Today I got my DA 70mm F2.4 and even though I have not spent enough time with it seems like -10 calibration makes it sharper. And seems like built-in -10 to +10 fine tuning is not enough. Doesn't it mean body needs pro calibration?

p.s. all posts with embedded images updated to link to images now.