Weird noise in long exposure (30+ min)

Nikon Rumors Forum » Nikon DSLR

Weird noise in long exposure (30+ min)

(19 posts) (8 voices)

Tags:

No tags yet.

BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

So, I got into a very quick discussion with a guy about star trail photography (multishot versus single long exposure). I championed the single, long exposure (like the film days) while he told me that that was impossible because of the full moon he was shooting in (nothing would show up was how I interpreted it). Anyway, I set out to prove a point by shooting under a 95% full moon and in a small city. I got my start trails, and a really weird red band down the side of the frame.

The red side was on the left. The temperature was about 55-60 F (the low for the night was 55 F). I was using a well used d300, ISO was 100, at f/22, with the 18-200mm set to 18mm, exposure was about 37 minutes.

Here is the RAW:

Here it is with the white balance corrected, the contrast dropped to 0, and brightened a bit:

I know there is a spot on the lens (trust me, it's the lens). Also, this red showed up in another 30+ exposure that I didn't like with the exact same pattern (moving the camera and rotating it a bit did not alter the angle of the pattern). It does not show up using normal exposures.

I've heard of hotspots appearing with long exposures but not a whole gradient.

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

RAW:

Edited:

Posted 1 year ago #
golf007sd

preferred member
Joined: Nov '10
Posts: 585
offline

Have you tried using a different lens to see if you get the same result?

Posted 1 year ago #
TaoTeJared

preferred member
Joined: Apr '10
Posts: 2,422
offline

Looks like sensor overheating to me. Most DSLRs do that with anything over 15 minutes. I have a D300 and see the same thing on the same side when I tried it. Some of it appears to be from another light source as well. Anything in or near the city will cause it.

Most good Star trails are compositions you see are of of multiple images of about 2-3 minutes each or with film.

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

I wasn't going for good trails. I was out to prove a point, more to myself than to the other person. I could only see a few stars and a couple planets last night (which was perfect conditions for what I was aiming to do). I've done trails on film, which is why I was for the single, long exposure.

I didn't think about the sensor over heating. But upon looking in other forums, the over heating appears to be only in the corners, or at least more of a point source than a bar. I suppose, with enough time, the entire left side could start to heat up.

I'm running a test in my basement right now. Another 30 minute exposure with the lens cap on. If it's repeatable, then it's probably an overheating issue.

Posted 1 year ago #
NSXType-R

preferred member
Joined: Mar '09
Posts: 2,803
offline

I think that's what they call amp noise. I don't think there's anything wrong with your sensor.

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

After hours of testing with a lens cap on(2-30 min. 1-55 min.), I can't figure it out. I have just standard noise, no red bar.

Maybe it's a reflection off the lens hood? I don't know. I assume a reflection would change when the angle and position changed when I moved the camera.

I'm not really worried, just curious as to what this is.

Posted 1 year ago #
PB PM

preferred member
Joined: Jan '10
Posts: 2,217
offline

Did you try shooting with long exposure noise reduction turned on?

Posted 1 year ago #
TaoTeJared

preferred member
Joined: Apr '10
Posts: 2,422
offline

Like I said above, it also looks like stray light coming into the image. If there is any, an I mean ANY light within 1/4 mile you will see it with that long of an exposure. My D300's right side is hotter than the rest due to amp noise - not that much though.

Just curious - what in the world are you trying to prove or wanting to do that you need a 30min exposure for?

Posted 1 year ago #
casperwb

preferred member
Joined: Jan '11
Posts: 558
offline

TaoTeJared said:
Like I said above, it also looks like stray light coming into the image. If there is any, an I mean ANY light within 1/4 mile you will see it with that long of an exposure. My D300's right side is hotter than the rest due to amp noise - not that much though.

Just curious - what in the world are you trying to prove or wanting to do that you need a 30min exposure for?

.
excatly

what is the point?

or more to the point, is what you are doing within the parameters of the cameras specifications?

me thinks that if you push equipment past what is was designed for, then we cannot complain, unless we are doing it for testing purposes to see at what point it fails.

torture test for the poor camera if you ask me.

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

The basic point was that one can shoot trails in poor conditions (full moon) with long exposures. He said it wasn't possible, the gauntlet had been thrown. Hasn't anybody else set out to do something because somebody said it was impossible?

Back story: Guy has pretty nice work compositionally speaking (better than mine), but appears to know very little technical knowledge. I asked a simple question that amount to "Why do you use this technique (multi-shot) over the single shot technique?" He fed me a line of BS ("there are many techniques") with no reason why he chose his method and then later told me that the conditions that he was shooting in (a field in the middle of Nebraska, one day before full moon) made it "impossible to shoot trails."

I figured that it was very possible Since he didn't give me a reason why it wouldn't, I set out to find out. My only concern was the aperture being set to f/22 (most trails are shot wider than f/5.6). So I went outside the following night (one day after a full moon, on the outskirts of a 40k person city) to do a test. I got my trails (point proven, at least to me), and that red bar and a new problem to solve.

I'm not looking to be a jerk to the guy. I did it more to prove to myself that it could be done, just with poor results (which I expected, I was in a city with about 10 stars visible).

I've done trails on film, so I'm familiar with stray light. There is a car lot to the left and below the image which shows up as a bright blue in the image I shot before this. The car lot light floods the bottom of the frame (light blue in color fading to the sky, it looks like stray light), not the right hand side. I assume that stray light from the car lot, or other point source, would change position or angle. Maybe my logic is wrong.

Anyway, I have no idea what the parameters of the camera are. I found examples of guys shooting 2-3 hours exposures with nothing remotely close to what happened to mine.

I was just surprised by the results. I'm not concerned because the camera still works and I don't really plan on doing trails.

As for torture, this body has convinced me that the d300 is nearly invincible unless you set out to break it or you have a really bad accident.

If it's clear tonight, I'll set up again. Maybe I can reproduce it with one lens and not the other.

Posted 1 year ago #
jerl

preferred member
Joined: Dec '10
Posts: 425
offline

If you try it again, I'd suggest turning on long exposure noise reduction. For me, that's pretty much the usual routine for any exposure longer than a minute or so. For a 30+ minute exposure, I wouldn't even think about not turning it on, so give it a shot and see if it helps.

I'd also tape up the focus marking window on the lens if you have one. You probably don't need to, but better double check.

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

jerl, I'll definitely try taping over the focus window.

With the long exposure noise reduction, don't you shoot two images? One of the subject and one with the lens cap on? If that's the case, shouldn't something have shown up last night during the 55min exposure shot with the lens cap on? I had maybe two dozen hot pixels after that shot.

Posted 1 year ago #
The Ridgeback

member
Joined: Oct '09
Posts: 46
offline

The result you have is amp glow, caused by the heating up of the electronics within the camera, the longer you leave it the more extreme the result.

You can shoot star trails with a full moon, but you will only ever get the brightest ones, as the fainter ones are swamped by the moonlight,

DONT turn on the noise reduction in camera, take your own set of darks, see my reply to the using a telescope with a dslr,

As for long exposures verses shorter, you are better off taking shorter subs, up to 5 mins and then stack the images to produce a better image.

Looking at your image, why have you used f22? the stars are at infinity, you dont need any depth of field, I personally would stop the lens down 1 stop and then take the photo.

When you see images that state 1 hr 5 hrs ect, that is the combined length of time using shorter images, not the length of a single image.

Mac

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

I appreciate everyone's help on this but I'm really only looking to identify the red stuff. I knew, in theory, then proved (again, more to myself) that star trails could be taken in a full moon. I knew that they would look terrible because of all the ambient light from the moon. I've done trails on film, I understand trail photography. I am only interested in what caused the red crap on the right.

Ridge, I used f22 because that's what my math told me to do. As I said, it was taken during a full moon, in a city, with car lots 1/8-1/4 mile to the west of me. a 3-4 hr exposure would have yielded a neutral toned sky due to the light wash from everything. Shooting at f/16 would have shortened the length of the exposure. My goal was one single long exposure to achieve the trails, so the longer the exposure the better for my testing purposes.

I like the amp glow solution; however, why did it not show up last night in any of the three shots I did (the only thing I changed on the camera was the lens)? It's at least 10 f degrees warmer inside than it was outside the night I took the photos so I should have more glow. The camera was capturing an images for about 60 minutes (1st photo was 28 min, 2nd photo was 37 min) the first night and 105 minutes last night (2 images at 30 min, 1 image at 55 min). I almost doubled the time with negligible pauses (just enough to check the screen and push the release again) between exposures and yet had zero, and I mean zero "glow."

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

I bumped my ISO to 6400 and started running more tests. Now, I'm really confused. It appears to be some sort of amp glow...but it's lens dependent. Here are the results:

18-200mm VR @ f3.5 for about 2.25 min

50mm f1.4 AFS @ f3.5 for about 1.85 min

Rokinon fisheye (9mm?) @ f3.5 for almost 3 min. (all manual lens)

Do the lenses themselves create glow from the electrical contacts?

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

Oh, I also have confirmation from another person that their 18-200 makes the exact same marks when they shoot long exposures.

Posted 1 year ago #
The Ridgeback

member
Joined: Oct '09
Posts: 46
offline

Its looking like two problems,
1st is amp glow, which you will always get unless you cool the dslr sensor. (thats why we take dark frames when doing astronomical photos and keep the exposures to about 5-10 mins)
2nd more worrying, if its different on every lens, then the only answer is light leak.

did you cover the viewfinder, as you do get some light leak there,

you can test for light leak, by taking the picture in a totaly dark enclosure and shining a very bright torch over the camera, obviously the image should stay dark.

Just had a thought,

The nikon D3 uses an IR sensor to measure the speed of the shutter to make sure that it is withing the tolerances, (thats why the D3 is not a good camera to fully convert to IR)
so if the D300 uses the same shutter module, then you will get IR leaking through the hot filter, as it does let some IR through and is not a total block, and looking at your lenses, you might be recording the IR reflections from the lenses.
would also explain why its more predominatly red.

have a look at this page for some more details on amp glow, ect.

http://www.markshelley.co.uk/Astronomy/Projects/canon350peltier.html

Mac

Posted 1 year ago #
BrownewellPhoto

member
Joined: Aug '11
Posts: 44
offline

Part of it is light leak, the lower left part, probably the viewfinder like you said. But there is still lots of noise with the 18-200mm.

I put the camera in a film bag and then put that in a box which was in a dark basement. I took a 2.5min exposure at ISO 6400. The results still point to some sort of lens dependent problem.

The fisheye and 50mm look alike, black with lots of normal noise. The 18-200mm still shows the red noise to the right. As I said, I found another guy who had the same problem with the same lens, he couldn't figure it out either.

I think I'll email Nikon and ask them what the deal is.

Posted 1 year ago #