Open

by damned truths

submit your photo


Hall of Fame
View past winners from this year

Please participate in Meta
and help us grow.

Take the 2-minute tour ×
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

Take a look at this image (shot with a Nikon D60). The shadows have been aggressively brightened to make the noise clearly visible. If you look closely, you'll see that the noise is not completely isotropic: there's a pattern of vertical lines.

enter image description here

  1. Why does this pattern appear? Is it related to the bucket-brigade-like libe-by-line readout process of the CCD?

  2. Do all cameras show this?

  3. Are there any tools to reduce specifically this type of noise pattern?

share|improve this question

1 Answer 1

up vote 5 down vote accepted

The Nikon D60 uses a CMOS sensor, not a CCD, so it has nothing to do with the read behaviour of CCD devices. (The CCD-type "star and line" noise is more acute, and I don't know of any current OTS tools that can clean it up adequately.)

What you are seeing is called banding, and all cameras using a Bayer or similar matrix will do it to some extent in poorly-exposed areas. It's not usually noticeable unless you are using very high ISOs or are trying to recover a severely underexposed image. Under more normal circumstances, it is well and truly drowned out by the image signal.

I know that Topaz DeNoise (an 8BF-type Photoshop-compatible plugin) has a facility specifically for dealing with banding noise using manually-settable parameters; I would assume that at least some of the other popular noise reduction tools would either offer something similar (but may automate the banding detection). (I've been very happy with DeNoise and have seen no reason to try anything else when the built-in NR in my RAW processor can't handle what's there cleanly.) Sometimes it may be necessary to apply overcleaned (and nearly detail-free) versions of an image in Luminance or Color modes (at various opacities) to the original in an image editor in order to get a decent balance of detail (which is often lost in the NR process) and clarity (which may appear to be entirely absent from the original).

Most of the time, though, it's better to accept and work within the limitations of the medium. As interesting as an image may be, there may be no good way to get it cleanly with your camera (because, say, there's no way to use a slow enough shutter speed to gather adequate light, or subject movement makes HDR impossible), so if there's no deep personal or historical significance to the scene, and it can't stand being simplified and abstracted in processing, it may be better just to pass it by. (That's a matter of maintaining mental well-being.) If you have to pass up too many images, it may be time to upgrade to a more capable camera (all of the current-generation APS-C and full-frame DSLRs are much better in low light than Nikon's 10.2MP CMOS sensor was).

share|improve this answer
    
Thanks for providing the keyword, now I can google for it. My question was more out of theoretical interest (I'm more interested in why it happens than how to avoid it). –  Szabolcs Jul 8 '13 at 11:16
    
(And yes, you're right, this particular image was very severely underexposed. The noise is actually much stronger in the image than the banding and I should have mentioned that some noise filtering has been applied to this image, which then made the banding stand out more. The banding is not really a practical problem.) –  Szabolcs Jul 8 '13 at 11:19
1  
Actually the D60 has a CCD sensor, but I guess that's besides the point then. –  Szabolcs Jul 8 '13 at 14:56

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.