[richardtugwell]

Finding some strange but inconsistent behaviour using 40mm Limited and K5



It was spot focused on the bricks to the right of the window. Shot wide open at f2.8. As you can see the right hand side of the image shows good sharpness, and actually an eerie amount of DOF considering the aperture, (from the back wall to the trampoline) and gets totally mush towards the left. It's almost as if the plane of focus has been shifted. I'm familiar with this using my view camera (intentionally) but .... what's going on. Seems to happen intermittently. Is the sensor tilting or something. You can see the camera POV is pretty well perpendicular to the wall. Initially it looks like a lens de-centering problem, but the 40mm has performed excellently and consistently on my K7. Any ideas?

Here's a <a href="http://zurich.dailyphoto.forthmedia.com/IMGP0837.jpg">bigger version</a>

[Monobod]

I have just returned my K-5 to Johnsons Photopia under warrantee for a similar but slightly different sensor alighment issue. On mine, the sensor and the electronic level were out of kilter with the camera and viewfinder. Setting the camera level in the viewfinder usind a spirit level as the subject, the image was 2 degrees out rotationally. Setting the camera level using the electronic level on the rear screen and the image was correct, but the view finder and the camera were two degrees off. Pentax said it could be recalabrated, so it has gone back.

Unrelated, but the pop up flash also has failed and will not recycle.

[Gimbal]

Maybe (in this case) it focused on the reflection in the window?

[richardtugwell]

Hi Gimbal - no, the autofocus was locked on the bricks to the right of the window, and then it was recomposed. In any case, would that explain the odd focus at the right hand side? There are objects at the right edge of the frame that are in front of the desired plane of focus, and look clearly sharper, and there are also objects behind that plane that are sharper.

[areidjr]

You say you focused on the bricks on the right. Was that rat right? If so, how many feet is that wall set back from the wall in the foreground. At f/2.8, the shallow DOF may well explain exactly what you are seeing.

[falconeye]

Hi Gimbal - no, the autofocus was locked on the bricks to the right of the window, and then it was recomposed

Whatever you think you focussed at, as a matter of fact, the reflection in the window (you ) is the subject which is in focus in this image. And that reflection is a lot behind the plane with the window.

[richardtugwell]

areidjr - the bricks immediately to the right of the window, between the window frame and the trellis.

falconeye - yes the reflection looks sharp, though it's hard to tell. however no one seems to be interested in why the right hand edge is more in focus than the rest of the photo. The trampoline legs are well in front of the intended focus plane, as are the chairs behind - they are near the ege of the image and yet they are at leasts as clearly focused as the reflection. Also, if we assume that the sensor plane is pretty much parallel to the wall, why such a discrepancy between left and right? As I said, there's nothing wrong with the lens - it's been on my K7 for 18 months with no issues.

(NB I have been focusing Pentaxes since 1964)

[Rupert]

My guess is that Falc is right...he usually is. I have no clue, but do know that shooting wide open often brings strange results, not just in the K5 but in my previous cameras, and my Nikon and Canon friends report similar results. Capturing a tiny DOF is a lot like trying to take a bottle full of light into a dark room.....and expecting to light it up.Not quite as impossible, but not always easy.
Regards!

[richardtugwell]

Hi Rupert - thanks for your thoughts. Yes, shooting a lens wide open challenges it's performance, but thats not really the point of this post. Anyway f2.8 at 40mm at this distance is hardly going to be tiny DOF.

Does nobody have any comments to make about the disparity between left and right?

[falconeye]

falconeye - yes the reflection looks sharp, though it's hard to tell. however no one seems to be interested in why the right hand edge is more in focus than the rest of the photo.

I wouldn't say no one is interested.

Phenomenons like yours are simply hard to analyze with a single photo as a base. If it wasn't, we all would know what all the photographed UFOs have been in reality

You can do a controlled experiment (*) with a test chart and some apparatus to make sure the optical axis goes perpendicularly through the center.

Without such a controlled experimen, we cannot help you. What you observed can be due to many things. Like the brick wall wasn't 100% parallel to the focal plane and some DoF effect plays a game with letting you believe the right side is sharper. I am pretty sure that the right half of the foreground brick wall is out of focus as well.

BTW, the sensor cannot tilt. It can only shift and rotate where both don't affect focus.

__
(*) - downlaod and print http://www.graphics.cornell.edu/~westin/misc/res-chart.html onto letter-sized/A4 paper (3 copies).
- hang the 3 copies onto a wall side by side, covering the entire Field of View of your lens (adjust the distance).
- use a wooden "right triangle" (or similiar device): place it at the center (twice: in horizontal and vertical orientation) and watch in magnified live view that one side points straight at you into the camera (an assistant helps here); this makes sure the perpendicular constraint is met. If not, change camera position.
- make a photo with trigger delay and upload a full size version to be discussed.

[Parallax]

Whatever you think you focussed at, as a matter of fact, the reflection in the window (you ) is the subject which is in focus in this image. And that reflection is a lot behind the plane with the window.

How would that explain why the trampoline to the right, which is closer to the sensor than the bricks on the far left, is in sharper focus than the bricks?
If the focus point were farther back than thought, the opposite would be true.

[falconeye]

How would that explain why the trampoline to the right, which is closer to the sensor than the bricks on the far left, is in sharper focus than the bricks?

I don't explain anything. I only noticed that the window reflections are in focus.

For an explaination, I recommended to do a controlled shot.

[richardtugwell]

I don't explain anything. I only noticed that the window reflections are in focus.

But you noticed nothing else?

Anyway, thanks for your comments, but we are obviously looking at this from different points of view.

NB Although the sensor is only supposed to move in a flat plane, with things like this manufacturing tolerances and engineering errors can come into play.

[falconeye]

But you noticed nothing else?

Anybody can speculate. I proposed to you a procedure how you can progress from here.

[conradj]

--Also, when trying to figure out issues abt image sharpness, it helps a lot to turn down the in-cam sharpening all the way. Set it to fine sharpness or extra fine sharpness and turn it all the way down. Make sure sharpening is also turned completely down or off in any PP software that's used. i'm seeing a lot of sharpening in this shot; it's hard to tell what's really what.

And of course: mount the camera on a tripod and turn off SR by turning on the 2s delay.

--It also doesn't look like your camera is pointed perpendicular to the wall; can't really tell from here since the visual cues looks like they're mostly constructed a bit lobsided. Since you say it is then it is.

Agree with falconeye in the focus plane seems well behind the brick wall. The picture frame and stuff behind the window glass, as well as the far brick wall and garden furniture on the right, are in better focus than the front brick wall.

O yeah, i just noticed this shot was at ISO400. For examining sharpness issues it should be lower and you should make sure all NR is turned off. Any image processing will always obscure these kinds of problems.