does K-5 AF accuracy depend on certain menu settings?

[Kerusker]

RetLaEnvEmp came up with this:

Next changed the setting of the "Program Line" (page 102 in my K-5 manual) to DOF Priority (shallow); this put more of the picture in focus as I was shooting a F/2.8 lens and the default camera setting out of the box was DOF Priority (DEEP) that put everything in a blur behind the subject I tightly focused on and a looked like a front focus problem when it was just the settings. Took me a long time to figure this out, I am sad to say, and I checked every mode in the K-5 to make sure there were any other things I needed to reset for front focus.

in his thread http://forums.dpreview.com/forums/read.asp?forum=1036&message=38305792 (bold markings above by me)

RetLaEnvEmp's observation made me think and I decided to ask all other K-5 users what they think about it (with this thread).
What do/did we have regarding the K-5 Phase Detect Auto Focus (PDAF):

• (1) Front Focus (FF) problem in low light, especially in tungsten light

Firmware (FW) 1.03 has (?) solved the problem although falconeye's paper update states something else.

• (2) PDAF accuracy micro adjustment per lens

For each lens type (up to 20) menu item 26 'AF Fine Adjustment' (AF-FA) allows for a micro adjustment value (+m to correct BF and -m to correct FF). There is also one AF-FA value available for all lenses which can't be identified by the K-5 or weren't introduced by the user. The latter AF-FA value also applies for all manual focus lenses by observing the focus beep etc.

• (3) What's the aperture of the PDAF system?

And how does it compare to the widest aperture of the mounted lens? If the PDAF aperture is about 1:4 or even 1:5.6, how can it accurately focus for a lens having aperture 1:2?

^^^^^^^^^^^^^^

If RetLaEnvEmp's observation is correct - which to me makes sense because to get a deep DoF, Pentax may pull the AF a bit into the front (FF), to be on the save side i.e. the object surely is inside DoF. But this strategy may be bad for fast lenses which have apertures of 1:1.4 to 1:2. And this strategy may be very bad for finding the best AF Fine Adjustment values for fast lenses.

So, if any of you have done AF Fine Adjustments for fast lenses, could you please report your experience? Does the setting of "Program Line" have an impact on AF accuracy?
Are there (more) menu settings which influence AF accuracy?

I don't have fast AF lenses, but I'll have a look at it with my lenses asap.
--
][.Kerusker
we don't see that we don't see (eye's blind spot)



my albums: http://picasaweb.google.de/Kerusker

 

[steephill]

You are reading too much into this and I suspect the OP is mistaken. The Program mode deals with exposure not AF and works just as well with MF lenses. The two DoF options are there to skew the program towards shallow DoF by favouring wider apertures and deep DoF by favouring smaller apertures.

The assertion that DoF shallow setting put more of the subject in focus is wrong - only a smaller aperture will do that. It may well be the case that for the exposure used the program line chose a smaller aperture.

I don't know if Pentax has published the program line info for later cameras but the original *istD manual showed how these operated. You can download a copy and see these.

--
Steve

http://www.pbase.com/steephill

 

[awaldram]

I think there is some confusion here on baseline and AF accuracy.

How wide the baseline is defines the point AF stops working for slow lens.

i.e is the baseline is f5.6 then AF would not and could not function for lens slower i.e the tamorn18-250 would not AF .

Now given its not a hard and fast point I guess the actual PD baseline is around f7.2

As to how this affects fast lens this depends on if enough phase variance can be detected to achieve decent repeatability for fast glass.

It is well reported that Canon scrxwed this up royally and had to use wide baseline additional sensors to mask their design error.

I've seen no evidence that any other manufacturer made this design error and everybody else seems perfectly capable of getting the required data from their sensors to support fast glass.

At a guess I'd say Canon comparators had less resolution than competing solutions so required wider data points to compensate .

--
My PPG

http://www.pentaxphotogallery.com/home#section=ARTIST&subSection=1471087&subSubSection=0&language=EN
My Photo Stream
http://www.flickr.com/photos/awaldram/