Strange banding in the sky

[Tringa]

Does anyone have an idea what has caused the banding in the sky on this shot?

It was taken about 1000 GMT this morning with a Pentax K5 and the 18-55mmkit lens set at 18mm at approximately 90 degrees to the sun.

F7.1@1/500s.

ISO 200

No filter on the lens.

The only PP was converting the original DNG file (on which the banding is also evident) to JPEG and resizing.

All other shots that included sky, including a shot virtually straight into the sun, show the same banding.

Thanks

Dave

[vladimir]

FWIW - it looks like a hardware malfunction.
If I were to guess - something like a stuck bit on the output of an A-to-D converter. But it doesn't matter, because it's not something that you can fix anyway.
I'd call Pentax tech support with this.
Good luck, and keep us posted about their handling of this.

[Tringa]

Went out at about the same time this morning and fortunately the weather conditions were fairly similar.

I took about 20 shots and there was no banding in the sky on any of the photos.

Very puzzled about what happened yesterday, but relieved that it does not seem to be a problem with the camera.

It is just very odd.

Dave

[Glenn NK]

Certainly is odd. I've seen banding but it was always uniform in width, and always perpendicular to the frame.

The banding in this image:

1) varies in width,

2) is not perpendicular to the frame,

3) at the upper right corner it changes direction (but still not parallel to the upper edge),

4) and at the lower right of the sky area, it is curved, and is concave upwards.

On my monitor I can't detect any banding in the water or trees.

To be truthful, it looked to me like a very unusual cloud formation.

Glenn

[vladimir]

Went out at about the same time this morning and fortunately the weather conditions were fairly similar.

I took about 20 shots and there was no banding in the sky on any of the photos.

Very puzzled about what happened yesterday, but relieved that it does not seem to be a problem with the camera.

It is just very odd.

Dave

Dave,
I am sorry, if I am going to ruin your mood a little, but I wouldn't be much relieved if I were you. What you are dealing with is "intermittent" hardware failure (as opposed to the "reproducible" hardware failure). Something is wrong, but we don't exactly know what, and you may have trouble (or not) proving to the manufacturer that something is wrong. I'd still call them and show them the picture, and try to get them to "fix" it... Because you don't want the problem to reappear at the critical moment...
Something to think about: What could've changed from the "problem" shooting to "no problem" shooting?
Did you recharge the battery?
Was the temperature different outside?
Was the camera itself warmer or colder?

Those things may affect the malfunction...

And now I'll try to explain my theory about the sticky A2D output bit.
Just an example:

Consider blue values in the sky going 0,1,2,3,4,5,...., 15
In binary it will look like this: 0000, 0001, 0010, 0011, 0100, 0101,..., 1111

Now, consider bit 2 (bits are numbered 0,1,2,3 from the right) is stuck at 0.

What you'll get is this (I am listing all 16 values)

Correct Values
-- Correct Decimal
---- Bit 2 stuck Values
------ Incorrect Decimal

0000 0 0000 0
0001 1 0001 1
0010 2 0010 2
0011 3 0011 3
0100 4 0000 0
0101 5 0001 1
0110 6 0010 2
0111 7 0011 3
1000 8 1000 8
1001 9 1001 9
1010 10 1010 10
1011 11 1011 11
1100 12 1000 8
1101 13 1001 9
1110 14 1010 10
1111 15 1011 11

I needed a table there, and what I get is... not quite, but hopefully it's still readable...
That's your banding (or something very similar). It bends (no pun intended ) with the darkness of the sky.

[Dave Humphries]

Hi Dave,

I'd agree with vladimir's diagnosis, the banding is obviously light value dependent and repeats itself at different values, this is, as vladimir says, a classic tell-tale of A-D (or D-A) converter problems, or anything in the D (digital) bit in the middle.

Because the image brightness increases from top left to bottom right (above the horizon), that's where the problem occurs, a bit like those contour lines on a map marking a certain height difference repeat; 50m, 100m, 150m, etc. or the isobars on a weather chart - and that's why, unlike some other digital processing artefacts, these bear no relationship to the top or sides of an image.

I too would be worried this might re-occur, se test some more and don't delete any of the images, you may need them as 'evidence' later, especially if it happens just after the warranty expires, tese will prove it was a pre-existing fault.

Hope you get a solution,

[Tringa]

Thanks for your comments Vladimir and Dave.

I will certainly keep checking for this happening again and will keep all the photos.

Unfortunately I dropped the camera very soon after getting it and I think Pentax's view will be that this has caused the problem (which I have to admit it may have done) even though until yesterday it worked perfectly. It could be costly.

Dave