Web Analytics RiceHigh's Pentax Blog: Focus Adjust the K-5 and Pentax AF Lenses with Live View

Thursday, January 26, 2012

Focus Adjust the K-5 and Pentax AF Lenses with Live View

In this post, I am going to brief about a new method and the general procedures to calibrate (focus adjust) the AF accuracy of the (Pentax) DSLRs which are equipped with the Live View function.

Before going on, there is something to note about the Live View function (called "LV" thereafter) of those Pentax DSLRs, the aperture in the LV is controlled automatically by the camera body, in doubled f-stops per step from wide opened, according to the brightness level as metered, but NOT according to the set Av by the user nor by the camera in Program mode. Just say for example that a F1.4 Pentax prime is mounted onto the K-5, the stepping is from wide opened and then to f/2.8 and then to f5.6 and then to f/11, as brightness rises. This is different from other ILDC cameras and systems like the Sony NEX, which always stopped down to the exact set Av no matter how.

So, here we go, the following are the detailed procedures which I have designed and recommended. Read in full and strictly following these for the best results:-

1. First of all, we need to find an ideal AF target. Find something that is flat in surface, with white background and dark but coarse *irregular* patterns, which shows the highest contrasts which are easily distinguishable, like this:-



I focused on the "Kleenex" words. On the other hands, objects with coloured patterns are not recommended so as to avoid the effect of Chromatic Aberration of a lens that might exist, e.g., red colour would front focus more for some lenses.

2. Mount the camera body with the lens you want to calibrate/focus adjust on a sturdy tripod. Put it in the environment that you want it to focus correctly, e.g. at a particular subject distance and the light source you used. For the sake of standardised calibration basis at my side, I put the object at a distance at about 2 metres from the camera and used an artificial light source at about 5200K to simulate the natural daylight.

3. Turn on the Live View of the camera, do the Live View CDAF, observe the focus scale. Repeat the CDAF three times to see if focus arrives at the same position. If it is (as it usually does), note down the exact position of it, take the picture and then switch off LV. Check if the picture is correctly focused (it usually is).

N.B.1. If the lens does not have a focus scale, stick on some marks by yourself.

N.B.2. Do observe about the aperture chosen by the LV by watching through the lens from the front which is according to the brightness level as what I have briefed in the beginning. If you want the lens to be wide opened for the calibration, the lighting level is required to be turned down a bit so as to make it!

For illustration purpose, say for example, I achieved the LV CDAF focus with a perfectly sharp picture with the lens focused at this position, according to the focus scale of the lens:-



4. Go to the Custom Menu number 26 for "AF Fine Adjustment", select "Apply One" and you should see the "Unset" setting if the lens has not been calibrated before. Do note that the camera only supports Pentax lenses with an identifiable lens ID for the Pentax F series AF lenses or later.

5. Do the Phase Matching AF (PMAF) and see if it is front or back focused in comparison to what the Live View AF has determined. Do the PMAF for several times until the AF motor doesn't turn anymore.

If it is front focused, like this, key in a negative (-ve) value for the focus adjust:-



Otherwise, if it is back focused, like this, key in a positive (+ve) value instead:-



6. Key in the compensation focus adjust values by trial and errors until the PMAF completely agrees with the LV CDAF, i.e, turn on the LV mode, did the AF and then switch back to PMAF and re-do the AF, the lens focus ring should NOT turn, anymore.

So, we did it. Below are the results for some of my Pentax AF primes. I've noted that all my zoom lenses (no matter DA or FA) do not require any focus adjust before they were accurate but fast primes would. This may be owing to the lack of a f/2.8 AF sensor in any Pentax DSLRs, including the K-5. Thus, focus adjust by individual lens is essential and crucial:-

Lens
Focus Adjust Value as Measured
Conditions
FA 28/2.8
-2
2m, 5200K, f/2.8
DA 35/2.4
0
2m, 5200K, f/2.4
FA 43/1.9
+5
2m, 5200K, f/1.9
DA 70/2.4
-5
2m, 5200K, f/2.4
FA 85/1.4
0
2m, 5200K, f/2.8
DFA 100/2.8
0
2m, 5200K, f/2.8

Good luck and happy measurebation! I hope that you would get razor sharp images with your Pentax primes (which are mostly capable of when focused correctly) after calibration with my above method! :-D One final reminder: Forget about all those "45 degree" inclined chart methods, which is not going to help you to do the job but would just create more confusions! Believe me.. The Live View will tell you the truth, really! Try it out by yourself! :-)


Related:-

K-5 and K-r Debug Mode Unleashed, Finally!

K-x Debug Mode Tutorial - AF Adjustment

Finally, I am Done! :-D

Comments (23)

Loading... Logging you in...
  • Logged in as
I would also suggest that, when you calibrate your lens, use both a close by object AND a far away one that's close to infinite (but not quite).

My own experience is, if you calibrate lens for FF for objects close by, it could throw them WAY OFF for objects further away, so I always do both near and far calibrations.

Most times it will be okay, occasionally I find a lens that FFs close but BF further, so I have to compromise.
10 replies · active 687 weeks ago
Focus errors for something that is near infinity is usually unimportant due to the huge DOF when focusing at such distances. So, I say RH's method should be fine and his 2m distance is well chosen. Of course you can select 2 distances like 1m and 3m to get more accuracy over a longer range of distances, but I won't try to calibrate using a focusing distance that is close to infinity.
Its not about the DOF, but rather, the difference micro adjustment can make when focusing to infinite. (answer it's HUGE amount)

Why don't you try it? see how much difference micro AF adjustment makes - just to make a point, calibrate using close object, then throw it off by 2 or 3 degrees. this tiny amount hardly makes a difference in objects close by - now try infinite. Be shocked how blurred your pics will be.
Why do you want to "throw it off by 2 or 3 degrees" and not just focus on a distant object and let the AF determine the number of degrees to turn? I have no problem at all focusing at distant objects and getting sharp focus on them after micro adjusting the lens to a focus distance. That was my point and that is all you want to do when taking photos. If some misalignment should exist at near infinity, will it really affect the abiliby to get sharp photos of objects at infinity?
You still didn't get it: AF micro adjustment does not make *AS MUCH OF A DIFFERENT* in closer objects compared to further away ones.

So, lets say if you make a adjustment of +1, in close by objects, this adjustment will probably translate to the focus point being moved for a few micrometers forward, but, this same +1 adjustment, meant the focus point is moved up to a feet or so when it's focusing on a far object.

When I said "throw it off by 2 or 3 degrees" is MEANT as a deliberate emphasised DEMONSTRATION of this effect, I am not telling you to throw adjustment by 2 or 3 in ACTUAL fine adjustment.

BECAUSE the adjustment makes such a small difference when tested close by, you *may* mistakenly adjust it by off a degrees or two, and these degrees will translate to a huge FF or BF problem when you DO try to focus far. That's what I meant by, after you did you close by object adjustment, use a far object AS A SANITY CHECK just to make sure your adjustment is not adversely affecting focusing near infinite.

There are also the issue of non uniform lens misalignment, my FA43 exhibits a fraction of BF on objects close by - BUT if I make the adjustment based on this BF and make it perfect for focusing short distence (I need to dial in a +3 for this lens), this caused a HUGE FF issue when focusing far. So in the end I had to compromise and choose to live with the tiny amount of BF, compared to the un-usable far focusing.
I think both Mage and you have points that are valid. Btw, I shall try to do the "infinity sanity check" on a bright day as suggested. Thank you to both of you!
In other words, the infinite is more of a sanity check.
The sanity check has just been done once, the infinite focus at f/2.0 for my 43 Limited has been dead accurate with my last calibration setting at +5. Lucky me~ The weather is misty and hazy though and will check again on a bright day.
Very good! Sounds like you got a nice copy of the 43/ Like I mentioned, my 43 really needs a +2/ 3 adjustment close by but if I apply the +3, it will front focus terribly at infinite. Bad luck for me :(
Mine is early production unit from Japan.
I see what you mean now, Ying. I was concentrating on the calibration, i.e. adjusting FF/BF, and it is more important to be accurate at close distances. I assumed the lens is not faulty or flawed, but you were concentrating on testing the lens with the second infinity distance re-check. You found a problem that should not cause a problem if built according to specs or expectations. I also have not noticed any problem in focusing at infinity once my lenses have been calibrated, usually at 5 ft. In fact, I think your lens could be treated as faulty and something should be done by Pentax to address that.
I was also using this technique to check if my lenses are well calibrated. LV is always correct :-) BTW, it would be cool if the camera could calibrate the lenses automatically... Use contrast AF to focus precisely and then compare with reading from phase AF and correct for the difference.
3 replies · active 687 weeks ago
You are right, if this kind of feature is built in feature like custom white balance it could help to get over various Pentax AF problems. Someone could propose this to RICOH who has enough loud voice.
I did it but I got no response (Hoya).
I asked if it is possible to have an option to correct the AF errors based on the differences between LV and regular AF. The camera is certainly better in picking up the difference between two positions of the lens; first LV and second normal AF. The difference could be computed and used as 'correction factor'. It costs nothing because it is software and does not sound that complicated.
Great idea!
LV shots are not quite as well-focused as phase detect:
http://forums.dpreview.com/forums/read.asp?forum=...

I discovered this while taking various near/far shots with all lenses to determine optimal AF fine adjustment settings, and I used LV contrast detect as a standard, assuming it to be the sharpest image, which it was not.
1 reply · active 687 weeks ago
Re-try with a more contrast object in B&W and do several times of the LV AF until the AF arrives at the same place and check.
I followed your directions. It turns out my 55-300DAL lens was so out of focus that I had to enter De-Bug mode to fix it. I did that (using your K-x debug mode instructions) then re-calibrated all my lens'. WOW!!! Thank you. That worked like a charm. Now my macro actually focus' where I want it too and my long zoom is much sharper too! Thanks so much... me and my K-5 say THANKS!
Thanks for sharing this technique. Just used it on my DA 35mm macro (no adjustment) and DA 15mm (+4). This corresponds to the relative performance of both lenses prior to running the tests. Haven't tried them in the field again yet, but looking forward to improved performance from my 15mm.
Hello!
Do you have to try this adjustments at the lowest F-number or one is like another?
Tahnk you
István Szalatnyai 's avatar

István Szalatnyai · 559 weeks ago

Hi, I fine tuned my zooms except 50-200.
This has big back focus at 50mm, small back focus at 80mm and perfect between 100-200mm.
Any suggestion?
Thanks for your help and your efforts you put into this post,
1 reply · active 559 weeks ago
So, I suggest to focus calibrate your lens at 100mm something then.
István Szalatnyai 's avatar

István Szalatnyai · 559 weeks ago

Thanks a lot for your prompt answer. I understood the fine calibration from you.:) I really appreciate your help!
For my 16-45 I did the calibration at 2 meters as suggested but after that checked at infinity and different adjustments are needed. My 16-45 needed +10 calibration on 2 meters (same in the whole zoom range) but -5 at infinity.I use that lens mainly for landscapes and visually I experienced front focus on my pictures for that reason I kept the results I measured at infinity (-5). As a consequences now my K-5 and 16-45 is back focused on 2 meters:( I strongly advice you checking your 28 mm prime at infinity. It is hard to say that Panasonic is on the right track with their CDAF on their MILCs.

Post a new comment

Comments by