[Spotty Focus]

Mikesbike - yes, I realise the DOF is a lot smaller with the APS, but I don't think that's the problem. The problem is with edge to edge focus. The turbine picture was just a test and should have been at infinity - it was spot focused on a turbine and then turned to manual focus (on a tripod). It is possible that the AF didn't get the right distance but it still looks all wrong. The sheep picture was something I would have taken as a picture but was a good example - note again that the images is a left crop.


Here's the original sheep image (processed to JPG in camera)
Dropbox - IMGP0255.JPG - Simplify your life

It isn't perfectly sharp being hand-held but the field gets increasingly blurred on what I assume is roughly the focal plane. I realise looking at 100% is bad and some softness in corners is inevitable, but...

---------- Post added 01-03-20 at 02:10 PM ----------

Mikesbike - yes, I realise the DOF is a lot smaller with the APS, but I don't think that's the problem. The problem is with edge to edge focus. The turbine picture was just a test and should have been at infinity - it was spot focused on a turbine and then turned to manual focus (on a tripod). It is possible that the AF didn't get the right distance but it still looks all wrong. The sheep picture was something I would have taken as a picture but was a good example - note again that the images is a left crop.


Here's the original sheep image (processed to JPG in camera)
Dropbox - IMGP0255.JPG - Simplify your life

It isn't perfectly sharp being hand-held but the field gets increasingly blurred on what I assume is roughly the focal plane. I realise looking at 100% is bad and some softness in corners is inevitable, but...

Sorry, not allowed to edit yet. Meant to say that 'the example image above was a left crop of the full frame'

[BigDave]

When I look at the image enlarged/magnified, both right AND left get soft with the sharpest part of the image in the center. I would suspect the issue is the lens, but then again, if the sensor is not perfectly plumb to the camera system, it could be the sensor. If anyone has a way to test for this, please help out!

[Penview52]

Take a portrait photo of any target. If the blurred image is seen downward the lens is bad. If the upper part is blurred haze concentrated on one side of lens.

[btnapa]

As mentioned, it is generally worse on the left,

Any rattling from inside the lens. The other possibility besides decentering might be that an element moved/tilted on the left side and is causing this.

[rogerstg]

There are a few things you can do to test for decentering
How to check if your lens is decentered
A Simple Test for Decentering

Until you do some testing, we can only guess.

[mikesbike]

Absolutely, running more tests of a precision nature is imperative. T

I think you will find if shooting with a lens wide open, there will be less than optimum sharpness when you get to the edges. But this should be the same on either side. Then as you stop down to a smaller aperture, edges will improve, though usually the central area is sharpest.

Upon performing spot AF, and leaving the center of the frame where it is, that object should be at the sharpest focus point.

When running these tests, It seems doing the same testing with both lenses would be a good idea.

[Spotty Focus]

I found some suitable flooring to do some more accurate testing. It definitely looks to me like a decentred lens, with the focus on the left nearer to the camera and the focus on the right further away. I expect the 'real' images look worse on the left because the depth of field is lower for nearer subjects.


I tested the 16-85mm on the K70 and the KX, and my old 18-55mm on the K70. I used as closely matched a view as I could manage, and added a second K70 / 16-85mm image at a wider angle. I focused manually to avoid any AF errors (using live view), with the focus on the cross just left of the penny.


I'll upload the full images if I get a chance.

I'm not sure what the variation is meant to be in these lenses but it doesn't look great to me.

Incidentally, the view shifts slightly in the viewfinder when the autofocus changes direction - I don't know if this is normal or not. I believe it can be in some lenses, but I'm not sure about this one?

---------- Post added 01-06-20 at 07:21 PM ----------

Or am I seeing what I expect to see? Not sure about that 18-55mm image now. Perhaps it is only better because of the smaller aperture...

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[rogerstg]

I found some suitable flooring to do some more accurate testing. It definitely looks to me like a decentred lens

The tests you did were not suitable for determining the extent of de centering. You might want to test for de-centering before making that conclusion.

[Spotty Focus]

Perhaps I should have said 'tilt' rather than de-centred. I'm not normally one for lens testing, I'd rather just take pictures!


I don't have a star chart available but one of the first tests I tried (a few weeks ago) was to focus on a distant object and take a picture with it in each of 4 corners. The results were all over the shop. The wind turbine image at the beginning was one of these (meant to be at infinity focus). I've since realised that focusing it in the centre first isn't a very good test with a wide angle lens because the object will be at a very different distance from the plane of the sensor when it is placed in the corner of the image, although I suppose if it is sufficiently far away I don't suppose it matters. I'll try it again by focusing in one corner first, although finding a suitably busy linear subject a good distance from the camera that fills the width of the frame would surely do just as well.

[mikesbike]

The tests you did were not suitable for determining the extent of de centering. You might want to test for de-centering before making that conclusion.

Right. And all at closer distance. Results at closer range can be different than at normal-to-distant range.

However, from what I do see in this very non-definitive teat, the lines on the right appear to be sharper than those on the left. Not highly, but to a degree. But the slightest lack of precision as to the camera and lens being square to the grid being photographed can also produce a false impression.

By focusing on a solid, well-defined subject at some distance and wide open to reduce DOF, then moving the camera to place the object at different points of the frame, each shot is really focused at the same distance and plane, the camera simply being rotated from the same position, this eliminates that problem. So the focus, the distance, etc. will be identical for each shot, but when the shot is actually taken, it will come through a different area of the lens. Thus other variables are eliminated while the lens performance itself remains.

[rogerstg]

I don't have a star chart available

No access to a printer?

[Penview52]

The above first photo is a spot ( center ) focus and outside that focus is stating to blur. Since the plane of focus is not perpendicular to the lens axis and thin DoF hard to know if decentered. A decentered lens in any way will not give a good focus. I am suggesting a haze ( moisture build up in the lens ) have desiccant filled bag for storage . Try to shot a ruler with 45 degree angle from your position and the target is no. 6 or middle.

[Spotty Focus]

No access to a printer?

Not immediately, no, although normally I would have.


I should be able to find a suitably distant target in the vicinity to try the other method, but the rain, snow, wind and bad light haven't been helping.


The test on the floor was mainly because I thought the focal plane wasn't quite right and I wanted to see if it was any different between cameras and lenses. Also because it wasn't raining inside.

Bear with me...and thanks for the help.

[mikesbike]

You are right to wait for good lighting. Conditions should be good for quality photography, so any other factor is not in the way of clear analysis.

[Spotty Focus]

There was a bit of light this morning so I gave it a go. Probably not up to exacting standards, but hopefully good enough.

The obvious target was a large house about 1km away. Given I'm looking straight across a valley, the whole distant view should have been in focus, but I moved it to the corners anyway.

I'm not allowed to add any more attachments so I've added them to dropbox.


Dropbox - Test 16-85mm

First 5 are the 16-85mm, Second 5 are the 55-300mm

Last two are complete single images.

55-300 isn't 100% but is probably OK for a long zoom. Looks like the lens...