8

Following is a crop from a RAW shot, developed in Lightroom 5.7, into a setting sun from a Sony A77II – ISO 100, f/5, 1/20s, with a TTL fill flash.

In the darker regions near the sun a series of horizontal lighter lines can be seen. What causes these?

Light banding near sun

Here is the same image with contrast cranked up to make it easier to see the bands. Starting from an arbitrary band, I counted the number of pixels separating each successive band and got the following sequence: 18, 24, 12, 24, 12, 18, 24, 12, 18, 18, 24, ...

My guess is that it might be the sensor equivalent of lens flare, in which case the question would be: what is the physical characteristic of the sensor that is being illuminated?

  • Looking for "horizontal banding on sensor in bright light" on Google, I found ma.juii.net/blog/reason-for-horizontal-lines-on-camera-images and dsp.stackexchange.com/questions/19853/… ... I wonder if they apply to a Sony? – Elenesski Nov 20 '15 at 7:40
  • Zooming in, they don't look linear. Please upload a crop of the middle region in true size, png. – TFuto Dec 3 '15 at 18:19
  • @TFuto - here is full size no crop. No lines are visible except near the sun, which is upper left of the frame. I haven't tried any forensic processing or analysis, but the original is RAW so I could post that or try other things if anyone has an interesting idea. – feetwet Dec 4 '15 at 19:05
  • @Elenesski - Unlikely. This is a mirrorless camera, but it has a mechanical shutter, and it would be glaringly obvious if that were malfunctioning. Also, it's receiving DC power from its battery, not AC as the second link notes could cause problems. – feetwet Dec 4 '15 at 19:08
  • Can you post a second photo, with arrows or other marks to indicate which part of the image you're asking about? I think the use of the word "banding" may be confusing to some readers, as it is often used to describe something else. – A K Dec 4 '15 at 20:10
1

There horizontal lines are some kind of noise problem setting/reading the horizontal addresses in the camera array. So if it is a 10 bit address then bit 4 in the address is always set or never set. This would show up more clearly on the raw image. The compressed image would average out the problem to some extent.

  • OK, maybe, but any hypothesis on why that would only occur on every tenth or twentieth row? The spacing between the bands isn't even consistent. And it only appears in the part of the image near the sun but where there is decent dynamic range. – feetwet Dec 23 '15 at 2:41
  • I'm too lazy to analyze it, but I'd guess that the pattern is something like every 16 pixels in image. problem also seems to be somewhat intermittent. So it would show up in multiples of 16 pixels, 16, 32, 48 and so on. I also expect that this is somewhat masked by image compression. – MaxW Dec 23 '15 at 2:54
  • I just updated the question to include sample band offsets from each other in pixels: 18, 24, 12, 24, 12, 18, 24, 12, 18, 18, 24, ... – feetwet Dec 23 '15 at 16:31
1

I noticed the same trouble with sony A77II. Now my camera is at Sony's repair center, so I can't do any test, but I think that the problem could be the e-curtain. Did you try turning it off? I'll try when my camera will be back from the repair center.

  • Interesting. I wasn't even aware of that option. Sony's manual says: "When you shoot at high shutter speeds with a large diameter lens attached, the ghosting of a blurred area may occur, depending on the subject or shooting conditions. In such cases, set [e-Front Curtain Shutter] function to [Off]. When a lens made by another manufacturer (including a Minolta/Konica-Minolta lens) is used, set this function to [Off]. If you set this function to [On], the correct exposure will not be set or the image brightness will be uneven." But this shot was 1/20s with the Sony 2.8/16-50mm lens. – feetwet Jan 15 '16 at 1:28
  • Yeah, i noticed the trouble with the Sony 16-50 too, at 1/25s. The problem described in Sony's manual is another one. I'll let you know when i'll do some tests. (sorry for my english ;) ) – Andrew Jan 15 '16 at 10:58
  • I did the test: the lines still appear using the first curtain :( – Andrew Jan 18 '16 at 15:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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