[DeadJohn]

I do focus calibration by starting with my thinner DoF lenses and proceed from there. My FA 43 came first. Focus was way off. I went all the way to -8 until I got consistently sharp results at f1.9. That was not a good start, and had me fearful whether my other lenses might exceed -10.

The DFA 100 macro WR was a pleasant surprise. Perfect focus at 0. The DFA 28-105 similarly seems perfect at 0. Whew, my camera is fine.

Last time I calibrated my K-5 I used a diagonal focus chart from Yvonne Bourque (apologies for spelling). This time I used a computer monitor. IMO it's much easier to determine the sharpest spot using tiny pixels. Hold the camera at an angle to the monitor, use only center point AF, then zoom in. If the sharpest pixels aren't in the center of the test photo you can adjust AF and retry.

[lightbox]

Very confusing process for someone who is just starting out, I have to admit. I've seen half a dozen different methods and procedures... some more scientific and involved than others. I have more questions at this point than answers.

Are screen-based tests better than the dedicated tools, like Spyder LensCal?

Does it matter what distance you set up the lens, as long as it's at least 10x the focal length? Will you get inaccurate focusing at certain distances if you calibrate for a specific one?

Since the camera only takes one AF adjustment value per lens, how do you calibrate a zoom lens? Do you split the difference and use a middle focal length?

Since many sources state that DoF extends from 1/3 of the way I'm front of the subject to 2/3 of the way behind it, then why do we try to set the focus exactly in the middle of the target?

[pschlute]

Since many sources state that DoF extends from 1/3 of the way I'm front of the subject to 2/3 of the way behind it, then why do we try to set the focus exactly in the middle of the target?

Because for near focussed distances, DOF is nearer 50/50, and anyway DOF does not display the same level of sharpness over the entire range. The point of focus is sharpest and starts to get progressively unsharp as you move towards the boundaries.

[DW58]

Since the camera only takes one AF adjustment value per lens, how do you calibrate a zoom lens? Do you split the difference and use a middle focal length?

I've wondered about this as well. Have read a few threads discussing fine tuning AF but didn't find an answer.

[stevebrot]

Since the camera only takes one AF adjustment value per lens, how do you calibrate a zoom lens? Do you split the difference and use a middle focal length?

You can do that, but it may prove inadequate. Why per-lens fine adjustment is needed is not immediately obvious and a little paradoxical. While poor physical placement of the PDAF sensor may require application of a global bias, per-lens variation in PDAF accuracy has a different cause related to design and manufacturing tolerance. A well-centered prime with few optical aberrations and traditional (non-internal) focus will seldom require fine adjustment. Add centering defect, field curvature, astigmatism, or LaCA to the mix and fine adjustment is usually needed.* Complicate with internal focus or zoom and the adjustment required may be different according to distance and/or chosen focal length.

Owners of recent Sigma products have the option of purchasing the Sigma USB Dock which may be used to upgrade firmware and do multi-point AF fine adjustment. It has been awhile since I read the general instructions (actual steps vary somewhat by lens), but if IIRC, one must calibrate at 3 or 4 different distances and at 3 or 4 zoom settings per distance. Yes, it sounds tedious, but reports are that it works quite well.


Steve

* These produce an ambiguous image to the sensor that reduces precision and may bias accuracy.

[caliscouser]

Since the camera only takes one AF adjustment value per lens, how do you calibrate a zoom lens? Do you split the difference and use a middle focal length?

You could split the difference and go down the middle. Or you could calibrate for the focal length you use most, or calibrate for the focal length where you use the narrowest DOF.
For example, if it was a 16-50 2.8, I would most likely use 50mm at F2.8 perhaps for portraits and the 16mm for say landscapes at F8. So the 50mm would be more critical.

*YMMV depending on the lens and your usage.

[dfp771]

There's an app I use to calculate depth of field for any lens/camera combo. It's called Digital DoF. It doesn't have the K-1 but I just use the generic 35mm full frame. For my Rokinon 135mm F2 the DOF @ 6 ft is ~ 3.1 inches. It also says what DOF in front of the focus point and what DOF behind. In this case it was about 50/50. This app appears to be reasonably accurate. Recall the DOF is the region where focus is "acceptable". Within any DOF the sharpness will vary slightly. I use the "DSLRKit" available on amazon for Fine Focus adjustment. With my K-1, the 31mm limited I own is set @ -2. For the 77mm limited it is +2. Both the Samyang 24mm and the Rokinon 135mm manual focus lenses are at -2. FYI.

[lightbox]

I'm sorry - I feel like I maybe derailed DeadJohn's thread. However the replies have been very helpful. Thanks!

[DeadJohn]

I'm sorry - I feel like I maybe derailed DeadJohn's thread. However the replies have been very helpful. Thanks!

Thanks but you didn't derail anything. I wasn't asking any questions with this thread, it was just some observations.

[TonyW]

I guess like many here I have seen several methods including using various focus charts some commercially available some home made. As yet I have not had the opportunity to look at my Pentax lenses.

FWIW my opinion and preferences:

If using a chart at some angle to the lens (e.g. 20-45%) then IMO it should ideally have another part of the card that has the focus point and plane parallel to the camera senor.

One of the very good and free methods was published by Bart van der Wolf, involves using the effect of moire to assure accurate focus. The beauty is that you use a computer screen to focus on using the published test pattern Camera AF Microadjustment - for free Canon, Nikon, Sony

Another method that seems to be well received is the Dot Tune method - something that I have not tried with Pentax systems

You Tube

You Tube

[Class A]

Hold the camera at an angle to the monitor, use only center point AF, then zoom in. If the sharpest pixels aren't in the center of the test photo you can adjust AF and retry.

This method is way too imprecise. You don't have defined target and hence your results will be very inconsistent. Please use a good AF chart or an equivalent way of providing the camera with a defined focus target.

I've put together some AF adjustment hints quite a while ago. Hope, you'll find them useful.

[DeadJohn]

This method is way too imprecise. You don't have defined target and hence your results will be very inconsistent. Please use a good AF chart or an equivalent way of providing the camera with a defined focus target.

I've put together some AF adjustment hints quite a while ago. Hope, you'll find them useful.

Why is a target pattern on the screen less precise than a pattern on paper?

[Class A]

Why is a target pattern on the screen less precise than a pattern on paper?

You didn't mention a target pattern on the screen before.

It sounded like you were not using a tripod ("...hold the camera...") and as if you were focusing on random screen content.

If you use a precise screen target (Tim Jackson's thick bar isn't ideal, AFAIC) then a screen can probably work as well as paper.

[stevebrot]

Why is a target pattern on the screen less precise than a pattern on paper?

Which pixel are you telling the camera to focus on? After all, the target is not parallel to the sensor. That is the Achilles heel of all slanted target methods. If you want to use your monitor, the moire' method suggested above is very sensitive, though I still prefer a high detail focus target. I personally use a variation of the DotTune method.


Steve

[Class A]

That is the Achilles heel of all slanted target methods.

A slanted target with a single transition from white to black should be indistinguishable from a non-slanted target with the same transition.

Yvon Bourque's focus chart has a good target pattern, but it might be too big, depending on the distance to the lens. One does not want to read out FF/BF too far away from the centre. I think only for very flat field macro lenses one need not worry where to check for FF/BF.

I once created a focus chart as a combination from Tim Jackson's and Yvon Bourque's designs that I quite liked.