50D Vertical Banding

[fike]

I installed the firmware update 1.0.6 that was supposed to correct vertical banding issues.  

I shot a sunset last weekend, and tried to boost the shadows, a lot!!  I pushed the fill light in ACR up to 100%.  I saw bad vertical banding in the shadows.  (I'll try to post the sample tonight.)

What causes vertical banding?  Is it a natural outgrowth of extreme ISOs showing high-levels of noise from excessive signal amplification?

[Panopeeper]

Are you talking about raw or sRaw? If raw: can you upload the raw file of such a banding image? The typical pattern noise is usually horizontal.

[fike]

Are you talking about raw or sRaw? If raw: can you upload the raw file of such a banding image? The typical pattern noise is usually horizontal.

It is RAW, not sRAW.  I'll try to upload a sample tonight.

[fike]

So here are the sample images.  The shot was taken at ISO 200, f/8, and 1/8 second.  On the left is the RAW image processed "as shot." On the right, the image was +1.20 exposure and +100 Fill Light.  I know those are crazy-high numbers, but I am a bit baffled by what appears to be banding. I could get some usable detail after some aggressive noise reduction, if it weren't for the very ugly banding artifacts.  



Any ideas what I am seeing?  I guess I could have shot HDR, but that never works as well as I feel it should.

[Panopeeper]

Fike,

I asked for the raw file, not for a JPEG. Anyway, as far as I can see this is the customary pattern noise, some call it banding.

I could have measured in the raw image, at which intensity of the dynamic range this is and compare that with other cameras. Anyway, I guess this is in line with other cameras. The 40Ds are doing that, the 5D2s are doing it, etc; it is probably close to the end of the tenth stop of the dynamic range, you can'r expect better from the 50D.

[fike]

Fike,

I asked for the raw file, not for a JPEG. Anyway, as far as I can see this is the customary pattern noise, some call it banding.

I could have measured in the raw image, at which intensity of the dynamic range this is and compare that with other cameras. Anyway, I guess this is in line with other cameras. The 40Ds are doing that, the 5D2s are doing it, etc; it is probably close to the end of the tenth stop of the dynamic range, you can'r expect better from the 50D.

so you did say....

here is a link to it.

http://www.trailpixie.net/temp-image/Fall-...WV-50D-3902.CR2

[Panopeeper]

here is a link to it.

Everything is all right. That area with the cross-hatchet pattern is even darker than I estimated, partly at the end of the 11th stop, i.e. clearly out of the range of the camera. This scenery had a huge dynamic range. The shot could have been made with 1 EV higher exposure, only a tiny bit of the sky would have clipped; that would have reduced the pattern. Look at the raw histograms below.

In such situations it is the best to shoot bracketed and select the most useful shot.

[fike]

Everything is all right. That area with the cross-hatchet pattern is even darker than I estimated, partly at the end of the 11th stop, i.e. clearly out of the range of the camera. This scenery had a huge dynamic range. The shot could have been made with 1 EV higher exposure, only a tiny bit of the sky would have clipped; that would have reduced the pattern. Look at the raw histograms below.

In such situations it is the best to shoot bracketed and select the most useful shot.

Thanks for the info.  I did shoot some other exposures, but I don't like sunsets when I clip one of the color channels in the sky.  It tends to make some really weird color effects that I don't find to be attractive.  

I was amazed to find that there was any image data in that shadow.  I really didn't have a high expectation of it being usable.

[Panopeeper]

I did shoot some other exposures, but I don't like sunsets when I clip one of the color channels in the sky

I too am very touchy regarding sky clippings. However, one needs to analyze the image carefully before picking one from the bracketed shots. The following capture shows, what would have clipped if the exposure had been a full stop higher. The green stripes indicate, that the red would have clipped in those areas; that's all.

[fike]

I too am very touchy regarding sky clippings. However, one needs to analyze the image carefully before picking one from the bracketed shots. The following capture shows, what would have clipped if the exposure had been a full stop higher. The green stripes indicate, that the red would have clipped in those areas; that's all.

Yep, those are the types of blown highlights that can make strange and unnatural colors...also can cause problems for printing.It is almost like the blown highlights create edges between color gradations, giving it a posterization look that is unpleasant.  

In the Raw Analyzer RGB histograms, what am I actually looking at.  I understand a histogram, but what are the scale marks in the X and Y axes?

[Panopeeper]

In the Raw Analyzer RGB histograms, what am I actually looking at.  I understand a histogram, but what are the scale marks in the X and Y axes?

Rawnalyze User's Guide, Histogram