Novel concentric ring artifact with long exposure NR

[atigun]

A while ago while doing some moonlight/aurora exposures with my D200, two exposures following each other showed these concentric artifacts. They were exactly centered in the frame. The camera was moved and aperture was changed between the exposures (I was doing moonlight exposure at medium aperture and then the aurora showed up. I forgot to open up the lens on the first exposure). Both were in aperture priority which maxed out at 30s with long exposure noise reduction on. No images taken earlier or later under similar conditions have showed these.
Link:
http://www.photo.net/photodb/photo?photo_id=5806306

Go to "Larger" which should give you a 100% crop. Processed and cropped in Capture NX.

I doubt I will ever see this again, however because of another ongoing thread describing a "rainbow effect"
http://forums.dpreview.com/forums/read.asp?forum=1021&message=22716413
(middle image) I thought it was a point in posting it.

It is pretty clear that in my case it is not a pixel level artifact. In both cases the rings comes closer and closer together towards the periphery.

 

[BJN]

My best bet is these are "Newton's rings". An interference pattern that appears where the low-pass filter comes in close proximity to the CCD surface. I've seen this in a Fuji S2 and a Nikon D70, always on long exposures with lots of gain to the image. The pattern is very subtle and is invisible in normal exposures.

--
BJN

 

[London N10]

I am fascinated by them. I kind of like to think that nature has chosen my image to reveal a moon bow to me. But, it is probably that Newton Ring explanation or something!

 

[London N10]

Just had a look at your image, and it really does look like Newton Rings, more so in your case than in my example, that you linked to. The amount of times I have to rescan stuff because of them is unbelievable. I am not sure that I can get my head around them from a digital camera though.

 

[flat five]

i've seen this at least 10 times on these forums in the past couple of years.. always aurora shots and always with a UV filter on the lens.. one of the posters did a before/after (with/without the filter) and the problem went away... i'll wager $25 that your lens had a UV on when you shot this..
jeff

--
http://jeff.emptyfortunecookie.com/folio

 

[atigun]

i'll wager $25 that your lens
had a UV on when you shot this..
jeff

Thanks for your comment. Well, yes I had (Nikon L37) as always, shooting in snowy environment. However I do not see how a multicoated flat piece of glass should work differently from any multicoated lens element. If I had taken the UV filter off after the two shots it would have "proved" that it was due to the UV filter. However I "forgot" to do this test, and guess what, on the third shot in the series the rings were gone even with the UV filter in place...

And quite frankly I am not sure I buy the alternative "Newton ring" hypothesis with respect to the low pass filter. I have not seen them so perfectly centered and concentric on glass slides... It rather points to a digital phenomenon. It would be interesting if anyone have tried to contact Nikon about this phenomenon.

It also would be interesting to hear if all occurrences have been centered and perfectly concentric.

 

[atigun]

In your uncropped image, were they perfectly centered or off-center?

 

[atigun]

Oops, I now see your reply in the original thread that they were perfectly centered.

 

[flat five]

It would be interesting if
anyone have tried to contact Nikon about this phenomenon.

it's not a nikon thing.. here's the same deal with a canon..
http://forums.dpreview.com/forums/read.asp?forum=1031&message=10991394

and yes, they are always perfectly centered..
jeff

--
http://jeff.emptyfortunecookie.com/folio

 

[atigun]

Thanks a lot for that link. Do not feel so much alone with this anymore... However I have a hard time completely buying into the theories in the thread. Why would it only show up in two of my shots and not all the others? And why mostly with aurora shots? Also my two exposures were at two entirely different apertures. I would have thought that would have changed the chance of creating interference.

 

[atigun]

I should have done my homework better:
I searched google for 'aurora ring artifact image '
and found the answer in the second hit, at photo.net.

http://photo.net/bboard/q-and-a-fetch-msg?msg_id=00A5d5

The keyword here is the narrow spectral emission of the aurora:

"Darren Cokin, Nov 12, 2004; 08:40 p.m.
--snip--

Thanks to the University of Alaska forecaster, the explanation follows. "These are interference fringes due to the parallel faces of the filter and to the narrow spectral emission at 5577 Angstroms in the aurora. That green, atomic oxygen emission line is the strongest emission in the aurora near our film and eye peak sensitivity, so it shows up first when there is any device in the optical path which sorts out the spectral emissions." So, don't use filters!"

I was thinking about contacting them, they are just across the street from me here...

 

[atigun]

Update: After all, this could be a capture NX phenomenon.To recall, here is the crop of the center of the Aurora image:



Concentric rings also showed up in a test designed to test compressed NEF quality in highlights. Basically an unfocused white piece of paper (unedited original looks uniform light yellow-white) where levels were edited manually in an "auto-level" fashion to spread the very few tones over the histogram. The subtle rings that unexpectedly showed up must originate in NX since the steps of the original tones in the image are much larger after editing levels. (The overall tones look smooth, however that is because NX interpolates across the original steps; the histogram of the edited Nef file shows interrupted lines.)
Capture NX ring artifact, crop center: