[niels hansen]

I think that the single point area in kp is too big. Is it possible in a future software upgrading to get a significant smaller singlepoint AF area?

[Ole]

If you are referring to live view it should theoretically be possible to make it smaller (i.e. read out a smaller area of the image sensor when determining the focus point).

The size of viewfinder AF point is, however, based on physical properties of the hardware based AF sensor in the bottom of the camera and can't be changed.

[niels hansen]

thank s for the information, unfortunately I seldom use lw

[normhead]

There are times when MF is a must if you want a great image.

[stevebrot]

This is an intriguing request. I have seen several recent threads where missed focus was attributed to the area sampled by an individual AF detector point (ambiguity due to less precise sampling). On the KP's center column, most of those points are actually four detection units working together (vertical and horizontal with f/2.8 and f/5.6 sensitivity for each). I guess it is safe to assume that they all share a common objective lens, but how large (width/diameter) a spot is "viewed" has long been a subject of conjecture. The common response is that it is "larger than the red indicator light" or "larger than one might think". A recent addition is that the center point is wider than the rest of the points in the array.

Last night, I was pondering possible test setups that might be used to determine the working diameter (degrees of arc or radians) for the various points. The aim would be to determine minimum diameter of a high detail target for acquisition and maximum diameter of a blank target in a high detail field for failure. It is important to remember that even if the second figure were known, making that number smaller does not remove the ambiguity presented by a receding curved surface (e.g. a peach in profile) or other situations where "large" sampling diameter is blamed for missed focus on fine examination (pixel peeping). Ambiguity of aiming replaces ambiguity of sampling.


Steve

[normhead]

A small flower, trying to get the stamen sharp, it doesn't matter what you use, you can miss focus, and i did. (although that is somewhat masked by the application of definition and heavy sharpening to the stamen area.)



It doesn't matter what you use, if you want to focus on something that small in the frame, you aren't going to know what in the frame is in focus. In this case I got the petal beside the stamen. You either take the time to use MF, or you take the chance the focus won't exactly as you want. It doesn't matter what camera you use or what AF system is in play. You can have everything set up exactly the way you want, and that AF system can adjust the focus as you take the image, so something you didn't want. That's why people who do a lot of this stuff use back button AF and disable the shutter release button activated AF. But, take the AF out of the equation and shoot MF and you get what you focussed on.

I know people want to blame this on the camera, but, blaming the camera doesn't get you anywhere, learning the appropriate technique does.

[SSGGeezer]

As I told another person in one of the threads Steve is referring to, With the dragonflies I have been shooting, I have to choose between what I want the KP to focus on and sometimes that area is too low of contrast to catch a skittish critter in a sharp fashion. then I have to turn my camera to manual focus since I have it set for focus priority while in AF-S mode. Every lens has it's own quirks as do auto focus systems, so shooting is often a learning experience. Some of my better dragonfly images came when I turned off the Autofocus and used manual focus in a more deliberate manner.

[niels hansen]

I agree that MF is the safe bet, I use it rather often, it works, but sometimes I could use a more welldefinated AF where I know where the focus is
I am 75 and sometimes when I am tired, MF doesnot work, especially with short distances , telephoto with full aperture. The same with moving objects such as playing children. I am not fgast enough. In most cases I prefer shallow DOF - except landscapes

[niels hansen]

Steve
I dont remember where I got the information, maybe the old mans memory is faulty, but if I am right the circle covers the single point area. Can you confirm that?
The best arguments for AF is aging eyes and laziness, and only single point. IMO many automatic AF points is a bad joke. No algoritm can decide what is the important part of the motive, average calculations may be acceptable but not good. in danish we say" good enough is not good"

[jimr-pdx]

I was shooting baseball through chain-link fence, and it's clear that my K-50 phase AF center point is not perfectly within the VF box. Most of my shots were manual focus prefocused to the bag I expected action. I tried to test where the AF boundary was but it wasn't an ideal time to be geeky about it

[stevebrot]

I dont remember where I got the information, maybe the old mans memory is faulty, but if I am right the circle covers the single point area. Can you confirm that?

I seem to remember something similar, but it has been about a decade since I last used the stock focus screen markings and back then, the camera was the K10D and it had many less points. Diagrams for recent models seem to show the center point being the sole point within the center "( )" brackets and with neighboring points immediately at those edges. Some marketing materials show a larger rectangle at center. The ability to "find" a high contrast edge (vertical, horizontal, and 45-degree slant) might show the edges of sensitivity. Alternatively an actual photo showing the lens array for the PDAF sensor might give a good idea of size too.

At the very least, one would hope that the diameter is wide enough to minimize dead spaces within the frame.


Steve

[stevebrot]

No algoritm can decide what is the important part of the motive, average calculations may be acceptable but not good. in danish we say" good enough is not good"

Things have changed, but it used to be that the first point that had sufficient signal became the active point for acquisition. Now several points are referenced and monitored and input from the meter RGB sensor might also be evaluated in real time (it can actually form a fairly detailed image with only 86 Kpx). Still though, given three dark blobs that might be the subject, unless the user has provided clues in the form of which point to use, the camera will fall back to a default selection scheme.


Steve

[niels hansen]

Steve
thanks for the information, my next quistion is: has single point AF problems with motives that are not in a right angle on the lensaxis? example given, a fence made of bars? Does it focus on the bar with the red indikator light. I this case MF is reliable and I find AF unpredictebel

[stevebrot]

Steve
thanks for the information, my next quistion is: has single point AF problems with motives that are not in a right angle on the lensaxis? example given, a fence made of bars? Does it focus on the bar with the red indikator light. I this case MF is reliable and I find AF unpredictebel

It might and it might not automatically focus on a bar rather than the background. I have not explored that case. I usually focus and recompose and for a fence or railing I use manual focus.


Steve

[niels hansen]

My conclusion on this thread
I am a little more wise, I am happy to know that I am not the only one with missing AF focus, I am a little disapointed over the"restrictions " on AF single point,I dont Know if it is realistic, but besides Bracketing exposure and DOF I should like to have a focus plane bracketing. Thanks to all inputters-
The next 14 days I will have no AF issues, on the holliday I use my Mamiya c 330 and Ilford FP4