Shooting stars and erratic noise

[flosseR]

Ok, I have been trying to shoot the night sky here and I am coming across a weird problem.
If I shoot at ISO 100 (not the default 200 in Nikon) I get a passable picture with an exposure of 30sec.
If I shoot at ISO 200 (the default) at 30secs I get these weird red and blue pixels, like erratic noise that just pops up.
I have noise reduction on in the camera so what is going on? And how can I get rid of it?

See attached images:
1 with ISO 100 and 2: ISO 200

REason why I am asking, I am about to visit my parents in south america and would like to use my fathers telescope with the D90 but if the results are like that its a waste of time.

It's not the sensor since it disappears with ISO 100

//F

 

[Nicolasdec]

What camera are you using? on my 5DII I have an option called "Long exp. noise reduction". When thats activated, I can do very long star trails (1 1/2 hour exposures) with very little noise.

 

[flosseR]

Nikon D90,
just skimming through the user manual now... I do have a HIGH ISO Noise Reduction...
EDIT: There is a Long Exposure Noise reduction and it has always been ON.....

 

[ArtandStructure]

Perhaps I am not seeing what you see but as far as I can tell looking at both images, the slight red and blue dots in the ISO 200 image are stars which are not visible or as visible in the ISO 100 image. This is as expected because your exposure time is the same for both images while your ISO is twice as sensitive to light in the second image and therefore twice as "exposed".

As far as the slight red and blue...it is slight. I suspect it would be hardly noticeable when printed out and it is certainly an easy fix if u desaturate them in post. Again, I believe these are stars and thus the long exposure noise reduction will not remove them from the image because they aren't noise. To test this theory take one image at ISO 100 for 30 seconds and the second at ISO 200 for 15 seconds. The exposure should then be roughly the same.

The rest of the image in both cases looks very clean.


All the best,


Jesse Widener
Art and Structure

 

[Matthew Yohe]

Google: "Hot Pixel"

 

[xStep]

I agree with ArtandStructure. You are seeing more stars and better colour definitions of the stars you were picking up before. It even looks like the second one was better focused.

As for the overall colour cast. Think Like pollution! The hint for that is the sky gets blacker to the top of the image, while the lower really brings out the extra light.

As Matthew pointed out, hot pixels can be an issue too. There are methods to help eliminate those.

 

[Ruahrc]

Perhaps I am not seeing what you see but as far as I can tell looking at both images, the slight red and blue dots in the ISO 200 image are stars which are not visible or as visible in the ISO 100 image. This is as expected because your exposure time is the same for both images while your ISO is twice as sensitive to light in the second image and therefore twice as "exposed".

As far as the slight red and blue...it is slight. I suspect it would be hardly noticeable when printed out and it is certainly an easy fix if u desaturate them in post. Again, I believe these are stars and thus the long exposure noise reduction will not remove them from the image because they aren't noise. To test this theory take one image at ISO 100 for 30 seconds and the second at ISO 200 for 15 seconds. The exposure should then be roughly the same.

The problem is that the mystery red and blue "stars" don't have trails like the other stars do, thus they are hot pixels and not faint stars. It is tricky though because some stars are redder or bluer than others (different stars burn at different temps) but these pixels look to be from the sensor, not the sky.

Perhaps the OP should check again to verify that "Long Exposure NR" is on and functioning. On my D80 when I take a long exposure image, it will take the exposure, then the top LCD will flash "Job nr" for a while (typically almost equal to the length of the exposure, so a 30s exposure will flash "Job nr" for about 30 seconds, for a one minute total time) which is the camera taking a second exposure with the shutter closed for noise reduction. Did your D90 do this when you took these pictures? That is a way to verify that Long exp NR is working properly.

Virtually every camera is going to have some hot pixels (think of dead pixels on LCD monitors) and the higher ISO you shoot at the more will be visible, but Long ExP NR is supposed to get rid of this because the hot pixels will be in the NR image as well and subtracted out

Ruahrc

 

[flosseR]

arghh

stupid is as stupid does.

for some reason I never really checked the job NR. on the display, just turned the cam off when it became unresponsive..

tried it last night WITH job NR on and the hot pixels are gone..

GReat stuff lol.. should have read the manual better.

thanks for the tip now its all good .

//F