What's causing this patterned noise?

[Soma Jones]

I'm sure someone here will recognize this moire-like patterned noise and can tell me how to avoid it. I convert to DNG and do most of my initial PP in Lightroom and then with certain shots, import to PS. I suspect it has to do with sharpening in PS and saving as a TIff?
Strange that it's only started to occur with photos taken with my new 7D.

[IMG][/IMG]

[Snarkbyte]

At what point in your workflow does this pattern first appear? At first glance, I suspect this is an artifact from re-sampling for resizing. Have you checked the algorithm and parameters for downsizing?

[Soma Jones]

It appears after I open it back up in LR after having saved from PS as a TIFF. I'm almost certain you're correct but I'm having no success recreating this problem. It's happened to me only twice now in the past few weeks out of dozens of shots I've worked on.

[Snarkbyte]

I can only see the pattern in the sky. Did you by any chance process the sky in a separate layer in PS? Any special operations or filters on that layer?

[Glenn NK]

I can only see the pattern in the sky. Did you by any chance process the sky in a separate layer in PS? Any special operations or filters on that layer?

Al:

I think you're onto something - there seems to be a very sharp line at the sky/water interface - that doesn't look natural at all.

I've never seen a white line like that on the horizon - ever.

Glenn

[Colin Southern]

It appears after I open it back up in LR after having saved from PS as a TIFF. I'm almost certain you're correct but I'm having no success recreating this problem. It's happened to me only twice now in the past few weeks out of dozens of shots I've worked on.

Beginning to sound like database corruption. Do you save LR edits to the LR database, sidecar file, or DNG file?

[herbert]

Beginning to sound like database corruption. Do you save LR edits to the LR database, sidecar file, or DNG file?

This can be checked by deleting the image from the catalogue and reimporting the TIFF (or staring a new catalogue and importing to there). If the pattern is still there then it is a pixel level artifact.

It could well be the result of a frequency domain based filter. High pass or low pass filters will only allow through patterns of a certain frequency. You may have done some processing that amplified a previously unnoticeable spatial pattern.

It could be a compression artifact but I doubt it. If you are saving as TIFF then you should not be using lossy compression. Have a think about the type of filters and post processing you do in Photoshop and keep an eye out for it.

One other idea is that it is a TIFF format that Lightroom cannot read. Lightroom cannot interpret layers. So you have to tell Photoshop to save a flattened version of the entire image into the edited TIFF. You do this in the Photoshop options by selecting maximise compatibility in the TIFF save options.

Alex

[Snarkbyte]

Actually, the pattern does sort of resemble the "checkerboard" PS uses for a transparent layer. If the sky layer is partially opaque over a transparent background.... LR doesn't handle layers, so if the layers weren't merged for the TIFF file, the artifact might just be a partial transparency that would be covered by another layer when viewed in PS. Does that make sense?

[Shadowman]

It also looks like the pattern you get when you shoot through a screen door. "But surely you cannot have a doorway to the surf.

[Soma Jones]

Thanks all, I'm away at work at the moment but when I get home I'll do some more troubleshooting. No, I don't think I processed the sky separately and yes, I almost certain used a high pass filter layer for sharpening; a technique I rarely use anymore.
I do save the changes in DNG, not the LR database, nor sidecars per your excellant advice, Colin.

[Trailboss]

Soma, is this a scanned image? I had a few come out like that after I scanned some slides? Just asking, I'm certainly not skilled in post. Good luck.

[t6b9p]

One possibility is a G channel imbalance although I have never seen such a serious case for a colour image. I have not heard of it being reported for the 7D but it could be a QC problem with your camera.

There are two G sites for each R and B but as they occur on adjacent lines, it is possible that they can receive different signal amplification (by design or by fault). If your RAW processor is treating them as identical but your G channels have slightly different output then it can result in this checkerboard pattern. This problem is definately aggrevated by heavy processing and is most visible in continuous tone areas.

I recently reported on this concerning the Nikon D200
http://nikongear.com/live/index.php?...nel-imbalance/

I eliminated it by applying a correction to the G channel amplification and confirmed its origin by processing a RAW file through dcraw without demosaicing or interpolation. This processing method reveals the individual R, G and B photosites in the RAW image. The size of the photosites matched the size of the checkerboard artifacts perfectly.

I am not saying this is your problem but this is one of the lesser known (but wider spread than many realise) causes of the checkerboard pattern.

Sorry about the thumbnails, I haven't quite figured out the best way to load these. However, the pair of images show the checkerboard artifact before and after G channel amplification adjustment, and the lone image is a linear tiff, without demosaicing or interpolation showing the R, G and B sites.

[kirkt]

Is it at all possible to make the raw file (assuming you shot raw) available via FTP (like YouSendIt)?

kirk

[Soma Jones]

Sorry it took so long to get back to this. It's not in the .raw, it appears only after editing in PS; applying a high pass filter to the entire photo and then merging the layers/ saving as a .tiff
Solution: mask out the sky before merging. The high pass filter was used to enhance the detail only in the foreground anyhow.
I rarely PP this way anymore but it struck me odd that it didn't happen every time I did this to a photo.