5

Is it normal or not to have such stripes in a photo taken with strong backlight? Photographed on a new Canon 700D, jpeg from CR2.

enter image description here

apazent
  • 51
  • 1

2 Answers2

5

No, it is not normal. These stripes look a lot like Newton's Rings, even though they aren't rings. Did you use some filter in front of the lens, which might have created those patterns?

ths
  • 7,161
  • 1
  • 11
  • 25
  • No, there was nothing between the lens and people on photo. Please open the higher resolution: the shape of stripes is quiet regular, from whole photo from left to right. – apazent Nov 01 '18 at 20:21
  • 1
    It is, and it isn't regular. The width of the stripes is highly variable (roughly sinusoidal). They might also be ever so slightly curved. I can't be certain. My money is on a bad filter capacitor inside the camera, because this looks like AC hum bleeding into the internal circuitry of the camera. – dgatwood Nov 01 '18 at 21:12
  • @dgatwood, yes, I meant this sinusoidal regularity. Curved - not sure, it seems they don't vary along the width.

    So can I state the camera is defective?

     – apazent Nov 01 '18 at 21:34
  • The only way to know for sure would be to repeat the experiment with a different lens, and try to rule out other things, like some sort of weird optical effect caused by the slightly out-of-focus window screen, inspect the lens for any damage/smudges/unexpected cellophane stuck on the rear element, unusual light sources just to either side of the photo, etc. Rule everything else out. – dgatwood Nov 02 '18 at 05:50
  • 1
    Also, this isn't shot in live view mode, is it? If so, try not doing that and see if the problem goes away. It might be interference caused by a defective backlight inverter. – dgatwood Nov 02 '18 at 05:52
  • Also, this isn't a DO lens, is it? – dgatwood Nov 02 '18 at 05:55
  • DO lens wasn't used, the same is true for live view mode. I'll try to repeat the experiment having checked the rear element, with another lens and maybe even with another 700D, presumably from the same batch. – apazent Nov 02 '18 at 12:23
  • Newton's rings don't necessarily have to be ring shaped. I am wondering if this could be interference from a gap between the antialiasing filter and the sensor surface. – jarnbjo Nov 02 '18 at 17:21
  • You mean antialiasing filter produces a kind of aliasing? It's funny. How can I check that? – apazent Nov 02 '18 at 20:04
  • you can probably not check on that directly, but if the same effects occur with a different lens, then it's a problem with the body and you should return it or let it be repaired on warranty. – ths Nov 02 '18 at 20:13
1

Based on your comment: your image is super-underexposed. From the metadata:

  • Exposure time: 1/125s
  • Aperture: f/5.0
  • ISO: 1600

and it looks like, from the link you gave, that you've pushed this up about three stops in post processing. (That is, to the equivalent of ISO 12,800.)

It is not at all unexpected to see artifacts like banding when taking an underexposed photo at high ISO, and even less unexpected when you push up the shadows like that.

In this particular example, the banding seems rather strong, so there may be some problem. But... I'm also reminded of the old joke: "Doctor, it hurts when I do this." / "So, don't do that!"

mattdm
  • 143,140
  • 52
  • 417
  • 741
  • Okay, I agree, it's super-underexposed and ISO value is relatively high (not so high, usually I face problems with ISO 6400 or at least 3200). But the stripes are seen on the highlighted parts of unedited image too. Is that OK? – apazent Nov 01 '18 at 21:00
  • Also, why do we see sinusoidal waves, not just white noise? – apazent Nov 01 '18 at 21:03
  • I don't know about "OK", but especially for a 2013 camera at ISO 1600 and underexposure (see this question), I'm not shocked. – mattdm Nov 01 '18 at 21:32
  • 3
    The patterns of the stripes indicate this is more than just noise from underexposing and then boosting in post. I've shot Canon digital cameras enough to know what the banding one gets from pushing exposure too far with Canon sensors looks like, and this isn't only that. Something else is also going on here. – Michael C Nov 02 '18 at 04:09