5D Banding in Sky

This is the first time I've ever noticed "banding" if indeed this is what it is. Take a look at this pic.

ISO 100 F14 1/8th 24mm:



The sky in the left-hand side shows this:



And levels adjusted so you can see it better:



Note I may have had a polarising filter on (can't recall exactly), and there was lots of moisture in the air. However none of the other pics from this same location and subject exhibit the problem. The effect is visible in the RAW too, so it's not a processing thing.

Anyone else seen this? Anyone suggest a photoshop technique that might clean it up?

Circular bands=> it can not be interferences of the camera.
It is more likely to be a quantification artefact due to the JPEG algorithm.

Did you shoot JPEG or RAW?

I have had this problem before. I think it has something to do with the limited amount of steps a digital camera can render. When you increase contrast it becomes more visible so you have to be careful with skies and soft gradations when increasing contrast in PS. The best way to minimize this problem is to work in Raw and do the PP in 16 bit. Only after levels and contrast correcting you go to 8 bit.

Maybe I got it wrong but you should try to see the difference between 8 and 16 bit with the same subject. I can see the difference. good luck..

JPG Compression artifact.
--
WSLam
PIXOURCE Digital
http://www.pixource.com - Your source for Capture One (C1)
See Profile for more information

Solarization.. probably due to over-PP..
Take a look at the original image.

I believe this is called posterization.

There is a nice chap on this forum who has a webpage on basic PS things, including levels etc.

He has got quite a lot of info on posterization. You might want to check it out to see that's what you have.
http://www.ronbigelow.com/articles/posterization/posterization.htm
Kovax

I've seen this on a few shots from the 5D and on the 30D.

Easily fixed though, just do a noise filter and add a little surfact blur to the sky.
--
Visit me at

http://www.radiodenver.org/

For shots that you plan to PP pretty heavily I always develope from raw to a 16 bit tiff. Do all my PP then save later as an 8 bit jpeg. This will greatly reduce and in many case eliminate this posterization.

Bric

--
Photo Gallery @ http://www.therothenbergers.com

Shoot RAW and convert to 16 bit TIFF for editing.

That should fix you right up.

--

'Hmmm, I wonder what deep thought might convince these guys I am some kind of artist...'
http://www.jimroofcreative.com

Michael Boniwell said:

However none of the
other pics from this same location and subject exhibit the problem.
The effect is visible in the RAW too, so it's not a processing
thing.

Click to expand...

...If that is the case, all the answers provided here are, for the most part, WRONG.

There is a specific test provided by Norman Koren's ImatestPRO Vignetting/LightOff module that allows you to measure and determine, with a great deal of precision, any non-uniformities in the signal-to-noise ratio accross the image, assuming a global neutral color and unique exposure and light distribution (white frame).

Some of the answers you have on this thread partially point to a real issue, which is NON-UNIFORM distribution of signal-to-noise ratio that is attributed, directly, to the cam's Imaging Engine / Pipeline (a mix of RAW-to-RGB conversion algorithms, as well as the math. transformations / computations used to generate the .JPG output). All these non-uniformities WILL dissappear (as suggested) by simply shooting in RAW.

HOWEVER, you are specifically saying that they are also present in RAW conversions. This suggests that you may have to process this image in a WIDER COLOR SPACE. That is, once you generate the baseline RGB version from the RAW file, please, try to generate such conversion in wideRGB or ProPhotoRGB, from the get-go. If not, try to convert from the RAW-converted color space into ProPhotoRGB, right before editing. Then edit the image in that wider color-space, and try to keep the image there, even for storage purposes.

I found a SIMILAR problem in my RAW-converted images, on my older 1D MarkII, in pretty much the same (but opposite side corner), and the culprit was the limitations of the color space I was using. The camera captured such an intense, rich, deep blue that it did not "fit", from the get-go, in the color space I was trying to use for editing.

Try it out, and see what happens. It may or may not work, but that was my experience.

Happy shooting!

P.S.: I found particularly bothersome the RED chroma-channel noise you are getting, right from ISO100 sensitivity. Did you push, by any chance, the tonality of this image, with any levels or curves operation? Typically, I do not get this type of chrominance noise at these low ISO sensitivities.

All,

Thanks for tip about converting the RAW to 16 bit tiff rather than 8 bit. I tried that, and while the effect is still there, it's now much less noticeable, to the point where anyone viewing the print would not see it unless it was pointed out to them.

For the record, while viewing the RAW in "RAW Image Task", and "Neutural" picture style at 100%, the problem is still there, though very faint. As this is the first and only image I've ever noticed with this issue, I'm not too concerned, especially now that you've more or less solved it for me with the 16 bit method. Thanks again.

...Try DPP and use WideRGB as native output ColorSpace, in 16bit. Check results and let us know.

If you do not see it, you will then now what's the issue...