AF Zone Experiment

*NeverSatisfied* · 03-21-2009, 06:13 PM

Alright fellow Pentaxians, there has been a lot of discussion lately about "AF problems" with the K20D (and probably covering the K10D as well). The following information may be old news to some of you, if so then please humor me! (And hey I'm still off work from my surgery so I had to do -something- to kill the time!

)

A while back I had thought there was something wrong with the AF system in my K20D, and people like Marc Sabatella were gracious enough to point out that the actual center AF zone in spot-focus mode is quite a bit larger than the little flashing red box we see in the center. How much larger?

Well, another contributor (sorry I forgot the name) mentioned an easy test to find out. So what I did was run a strip of black electrical tape across a sheet of printer paper, about 2/3 of the way up. I taped it to a wall, then put the camera on a tripod and aimed at the white zone underneath the black bar. Then all I had to do was attempt to AF, elevate the camera a bit, try again, repeating this procedure until the camera would reliably autofocus on the black line. I then rotated the paper 180 degrees and started from the top, working my way down until focus would consistenty lock. I also checked the zone from side to side.

The result: The center AF point in spot focus mode corresponds almost exactly with the semi-circular outline in the center of the viewfinder. I repeated the test with a couple of different zoom lenses, at varying focal lengths, and the results were consistent.

Perhaps this information will help some of you who (myself included) are/were scratching your heads trying to diagnose an "AF problem". Personally I think now knowing the actual size of the AF zone will help eliminate some of the problems I was having with inconsistent focus.
Paul

Sean Nelson · 03-21-2009, 06:24 PM

I did a similar experiment with my K100D Super, except I just put a small black dot on a blank white piece of paper and then hand-held the camera, half-pressing the shutter button and attempting to get a focus lock as I reframed the black dot at various positions. My results with the camera set to center AF only showed that the AF sensor would lock onto anything within the following portion of the image:

This area is a lot larger than the little red LED in the viewfinder would lead you to believe. I suspect that some folks who believe their camera is "mis-focusing" are really seeing the AF module quite legitimately lock onto something they're not expecting it to.

*NeverSatisfied* · 03-21-2009, 07:06 PM

Yea I was one of those folks! I'm certainly learning more about photography, can't say it's translated into any better pictures yet.

**Ole** · 03-22-2009, 12:24 AM

Paul and Sean, interesting findings, you should consider submitting it as an article

Articles - PentaxForums.com

for the benefit of new users.

**Marc Sabatella** · 03-22-2009, 12:40 AM

When I tested mine, it looked more or less like Sean's, but extending *above* the "( )" more than *below*.

Maffer · 03-22-2009, 09:06 AM

Yea, interesting stuff indeed. I have mostly shots landscapes and macros using MF since I bought K200D and then K20D later on. AF has not been that important at all - not until I bought this Sigma 50-150 HSM II zoom. I have recently been pondering why the hell has it been so inconsistent with AF at times. For example trying to shoot our doggie by aiming her eyes with center AF point selected. This randomly results in hitting the nose with focus instead of eye (nose would be just on the edge of the center circle) when using wide aperture like 2.8 @ 100mm. Placing an even surface at the same distance like some contrasty book cover would lock the focus spot on every time.

So in theory some camera with, let's say, 50 AF points would hit the aimed spot more reliably? I guess the AF point areas would at least be a lot smaller and focusing on small subject (like dog's eye) more easier.

mattdm · 03-22-2009, 09:24 AM

Originally Posted by Sean Nelson

This area is a lot larger than the little red LED in the viewfinder would lead you to believe. I suspect that some folks who believe their camera is "mis-focusing" are really seeing the AF module quite legitimately lock onto something they're not expecting it to.

And so removing that indicator from the K2000/K-m might not be as bad a move as it seems. (Leaving off the orientation sensor is still cheap and lame, though.)

*NeverSatisfied* · 03-22-2009, 10:03 AM

Originally Posted by mattdm

And so removing that indicator from the K2000/K-m might not be as bad a move as it seems. (Leaving off the orientation sensor is still cheap and lame, though.)

Interesting! I might try disabling that red focus point indicator, and get used to using the the circle for a while. That would be less misleading visually. Does anybody else do that?

Maffer · 03-22-2009, 12:26 PM

I will most likely get split prism focusing screen to make sure I nail the focus where I want it to. Works wonders with my old Chinon camera.

*NeverSatisfied* · 03-22-2009, 04:42 PM

Originally Posted by Ole

Paul and Sean, interesting findings, you should consider submitting it as an article

Articles - PentaxForums.com

for the benefit of new users.

Hi Ole, well I don't know about Sean, but for myself, not sure exactly how that is done. But your comment is much appreciated.

Regarding the previous poster's comment about Nikon having 51 (?) AF sensors, that does stand to reason that such a system would be more accurate. Does anybody know if the K10D/K20D AF system is Pentax's most technologically "up-to-date"? I swore I was done buying cameras, and the K20D has pretty much everything I need, BUT, better AF -would- be enticing!

*FHPhotographer* · 03-22-2009, 10:39 PM

Originally Posted by Sean Nelson

I did a similar experiment with my K100D Super, except I just put a small black dot on a blank white piece of paper and then hand-held the camera, half-pressing the shutter button and attempting to get a focus lock as I reframed the black dot at various positions. My results with the camera set to center AF only showed that the AF sensor would lock onto anything within the following portion of the image:

This area is a lot larger than the little red LED in the viewfinder would lead you to believe. I suspect that some folks who believe their camera is "mis-focusing" are really seeing the AF module quite legitimately lock onto something they're not expecting it to.

I get the same results, perhaps an even larger "spot" in fact. It would seem that an after-market screen with a model-specific AF center focus circle would be a money maker even in this economy,
Brian

blende8 · 03-23-2009, 01:25 AM

The cross sensors cover the whole AF rectangle.
The center sensor is covering the spot circle.
I don't know anymore where I got the image below, but it shows the covering:

Experimental:

Maffer · 03-23-2009, 04:23 AM

Okay...that's a wicked mess in the center with those points. Thanks for that image. Need to experiment with this.

*helios* · 03-23-2009, 07:30 AM

Originally Posted by blende8

The cross sensors cover the whole AF rectangle.
The center sensor is covering the spot circle.
I don't know anymore where I got the image below, but it shows the covering:

Experimental:

This thread has produced really useful information. Clearly if you want to know what you are focusing on it is necessary to set the camera to centre point focus and then be aware that it will focus on anything that lies within the centre circle in the viewfinder. It certainly helps to know this but as there is no way of reducing the size of the focus point you will still have a problem if your dog's eyes and nose are within the circle!

Archie

opiedog · 03-23-2009, 07:39 AM

Originally Posted by Maffer

Yea, interesting stuff indeed. I have mostly shots landscapes and macros using MF since I bought K200D and then K20D later on. AF has not been that important at all - not until I bought this Sigma 50-150 HSM II zoom. I have recently been pondering why the hell has it been so inconsistent with AF at times. For example trying to shoot our doggie by aiming her eyes with center AF point selected. This randomly results in hitting the nose with focus instead of eye (nose would be just on the edge of the center circle) when using wide aperture like 2.8 @ 100mm. Placing an even surface at the same distance like some contrasty book cover would lock the focus spot on every time.

So in theory some camera with, let's say, 50 AF points would hit the aimed spot more reliably? I guess the AF point areas would at least be a lot smaller and focusing on small subject (like dog's eye) more easier.

I have the same AF "inconsistency" with the Sigma 50-150 HSM II with my k2000. I was shooting my daughter's softball practice and the focus was on and off.

I did an experiment myself on the center AF and it wasn't exactly dead-center but more to the right and bottom.

There's nothing wrong w/ my k2000's AF nor my Sigma 50-150. I just have learn how the k2000's AF behaves with different subjects, different angles, etc.

regards,

jordan