Tales of the unexplained

[Spam]

You mean just after converting to a raster image, before saving it? And also visible when viewing at 100%?

George, I shoot in RAW only. Then I bring the files up in Adobe Bridge and open them with the RAW editor. After tinkering with them a little bit I open them in Photoshop (directly from the RAW editor) and tinker a bit more, often using Nik filters such as Viveza (which I used on this image). Then I save the PSD file, reduce the image size, then save it as a jpeg for posting online.

I downloaded it and tried to view with IView. It looks like Canon has an embedded full size jpg too, like Nikon. I don't see those waves. And when playing with brightness etc. in IView they don't turn up. Maybe time to compare converters?George

I can't see the striations in the RAW file either, but I can see them faintly when I open the file in Photoshop (even before running Viveza). The file I posted to Dropbox is the unmodified RAW file (ie without the .XMP file that contains my own modifications) .

[george013]

George, I shoot in RAW only. Then I bring the files up in Adobe Bridge and open them with the RAW editor. After tinkering with them a little bit I open them in Photoshop (directly from the RAW editor) and tinker a bit more, often using Nik filters such as Viveza (which I used on this image). Then I save the PSD file, reduce the image size, then save it as a jpeg for posting online.



I can't see the striations in the RAW file either, but I can see them faintly when I open the file in Photoshop (even before running Viveza). The file I posted to Dropbox is the unmodified RAW file (ie without the .XMP file that contains my own modifications) .

The point is where and when in the pp those waves occure. It'snot in the raw. In the Canon raw file a jpg is embedded created with the Canon converter and the camera settings for some in-camera pp. It looks like Nikon, I didn't know. For some first info http://photo.stackexchange.com/quest...mkii-raw-image.

Most likely a result of pp like Dem and Graham sayd. Analyze your workflow and try to find the point where it occures.
Photoshop might start with some pp.

George

[Spam]

P.S. just noticed that the file you downloaded was not the same as the one you posted, see bottom detail.

Another thing I can't explain, Grahame, but it's definitely the same pic - the time stamp in the exif is 21:00:52 for both, and I always keep the original IMG number on all the different formats.

[John 2]

Simon, just downloaded your file and converted it to a 16bit jpg after some adjustment for contrast/saturation etc. in ACR. Wacked up the Structure to 100% in Viveza and there is no sign of the striations that your image has. Interestingly however, there is some vertical banding visible. Not sure what that is.

[chauncey]

I could not quickly replicate the problem after downloading the RAW file...
could it be a facebook problem?

[bnnrcn]

Hi John May I ask what a 16 bit JPG is ? I have always thought and experienced that a 16 bit raw file becomes 8 bit when saved as JPG. So, if there is a 16 bit JPG possibility why photoshop doesn't let us do that ?

Simon, just downloaded your file and converted it to a 16bit jpg after some adjustment for contrast/saturation etc. in ACR. Wacked up the Structure to 100% in Viveza and there is no sign of the striations that your image has. Interestingly however, there is some vertical banding visible. Not sure what that is.

[Stagecoach]

I seem to recall we have had threads on 'banding' previously, both situations similar to this and also unevenness of gradient masks and it seems that some people see it and some don't.

Could it be anything to do with differences in computer graphics software and also monitors?

[John 2]

Hi John May I ask what a 16 bit JPG is ? I have always thought and experienced that a 16 bit raw file becomes 8 bit when saved as JPG. So, if there is a 16 bit JPG possibility why photoshop doesn't let us do that ?

Sorry Binnur, didn't make myself fully clear. I have suffered similar striations in the past and solved the problem (to me) on recomendation, by modifying my workflow. This involved saving the image from ACR as a16bit tiff. All subsequent pp is then carried out on16bit file before finally saving it as a jpg for posting. I believe it's the fact that all major pp activity is carried out on a16 bit image that makes the difference.

[Spam]

Sorry Binnur, didn't make myself fully clear. I have suffered similar striations in the past and solved the problem (to me) on recomendation, by modifying my workflow. This involved saving the image from ACR as a16bit tiff. All subsequent pp is then carried out on16bit file before finally saving it as a jpg for posting. I believe it's the fact that all major pp activity is carried out on a16 bit image that makes the difference.

But John, if I'm going straight from RAW editing into Photoshop, and saving as a PSD, wouldn't that be equivalent to TIFF, or actually TIFF (ie little or no compression)?

[george013]

But John, if I'm going straight from RAW editing into Photoshop, and saving as a PSD, wouldn't that be equivalent to TIFF, or actually TIFF (ie little or no compression)?

A picture I've shown more times, why not again.
All the editing is done on a RGB raster image. Always observe where it comes from. You never do editing on a tiff or jpg or whatever. They are disk files.
The converter can be an instance of the in-camera converter or an external converter. When you look at the embedded jpg in the raw-file, that's created with the in-camera converter. When you used your external converter, you get another raster-image which, if it's ok, is send to photoshop.
You can do it through a diskfile too. Tiff has the name not to change the raste-image, when done with no compression. Jpg is changing it by definition.


George

[proseak]

If yoiu look below the pic in ACR, you see the Workflow options; I use 16-bit Adobe RGB. If you use 8-bit, posterisation will occur, esp in blue gradients.

[John 2]

But John, if I'm going straight from RAW editing into Photoshop, and saving as a PSD, wouldn't that be equivalent to TIFF, or actually TIFF (ie little or no compression)?

Yes, it is just about the equivalent but your RAW image is an 8 bit file and you need to convert it to a 16 bit file before you work on it. You can either do that in ACR as Peter has described (left click on the sentence describing the current image properties in the middle below your ACR image and select 16 bits per Channel before you save it) or if you go straight through to Photoshop, select "Image/Mode/16 bits per Channel" before you work on it. Its' not necessary for all images but for consistency's sake, it has become a standard part of my workflow.

[george013]

Yes, it is just about the equivalent but your RAW image is an 8 bit file and you need to convert it to a 16 bit file before you work on it. You can either do that in ACR as Peter has described (left click on the sentence describing the current image properties in the middle below your ACR image and select 16 bits per Channel before you save it) or if you go straight through to Photoshop, select "Image/Mode/16 bits per Channel" before you work on it. Its' not necessary for all images but for consistency's sake, it has become a standard part of my workflow.

A raw file is 14 bit. The created raster image in the converter should be 14 or 16 bit, but never 8 bit. You're sure the choice between 8 or 16 bit is not for saving? So leaving the converter.
Once you've saved to 8 bit, you "enlarged" the info in the bits. You can't restore that with changing again to 16 bit. Of course you can work further with 16 bits.

George

[Stagecoach]

I can't see the striations in the RAW file either, but I can see them faintly when I open the file in Photoshop (even before running Viveza). The file I posted to Dropbox is the unmodified RAW file (ie without the .XMP file that contains my own modifications) .

Simon,

I have had a play with your RAW file again tonight in ACR then PS (inc adding Viveza structure) and whilst I can certainly achieve banding (and I'm pushing things pretty hard to achieve that) in all cases it is a simple stepped circular pattern radiating out from one area.

There was nothing I achieved that was any where near your elaborate patterns.

16bit or 8bit working and also saving in very low res (yours posted image was not) made no difference.

What I did find was that using noise reduction in ACR made it easier to achieve banding, but perhaps that's not surprising as some say use noise to mask it

Are you able to replicate it again?

[bnnrcn]

Thank you John, now it is clear I always use 16 bit in (ACR and/or PS). Many thanks to CinC for warning me about this matter when I subscribed to PS and started learning how to use it

Sorry Binnur, didn't make myself fully clear. I have suffered similar striations in the past and solved the problem (to me) on recomendation, by modifying my workflow. This involved saving the image from ACR as a16bit tiff. All subsequent pp is then carried out on16bit file before finally saving it as a jpg for posting. I believe it's the fact that all major pp activity is carried out on a16 bit image that makes the difference.

[Spam]

Simon,

I have had a play with your RAW file again tonight in ACR then PS (inc adding Viveza structure) and whilst I can certainly achieve banding (and I'm pushing things pretty hard to achieve that) in all cases it is a simple stepped circular pattern radiating out from one area.

There was nothing I achieved that was any where near your elaborate patterns.

16bit or 8bit working and also saving in very low res (yours posted image was not) made no difference.

What I did find was that using noise reduction in ACR made it easier to achieve banding, but perhaps that's not surprising as some say use noise to mask it

Are you able to replicate it again?

Grahame, yes, after some experimenting I was able to replicate the effect using Nik filters, and boosting the structure. Still can't figure how the swirls are not concentric, however.