Sensor noise in long exposures

[daveake]

Yesterday's "365" was a 10 minute exposure and I noticed rather more sensor noise than usual, taking me a while to remove in PP. Now, since I have 2 camera bodies of the same model, I wondered if this noise differed between them and if I happened to have used the body that was worse.

So, I tested by taking a shot on each camera, with lens caps and viewfinder covers on, and with the cameras inside a closed camera bag to make sure no stray light got in. Both cameras had exactly the same settings, using ISO 100 and 10 minutes exposure. These are the results (cropped sections from the full frame):

Body 1 (the one I noticed the noise on):



Body 2:



You don't have to count the dots to see the difference between them! Now I know why I was noticing extra noise, but I'm amazed at just how much difference there is. I shall now put a label on one body so I know to select that one for night shots!

Body #2 has a later serial number, so is newer. Dunno if that makes a difference or not. Also, I wonder if the noise is actually thermal noise picked up from dust on the sensor - certainly at least one of those bodies (I don't recall which one) is in need of a clean. Next step then I guess is to clean the sensor and see if the noise improves.

Your thoughts?

Dave

[JimNotJon]

Looks like sensor heating to me.. usually gives off them two colours. Do you have built in noise reduction de-activated? I turned mine off on my Pentax K10D to do star trails and forgot to turn NR back on having done a similar length exposure to you and getting the same "noise" marks.

Jim

[daveake]

JimNotJon said:

Looks like sensor heating to me.. usually gives off them two colours. Do you have built in noise reduction de-activated? I turned mine off on my Pentax K10D to do star trails and forgot to turn NR back on having done a similar length exposure to you and getting the same "noise" marks.

Jim

Yes, I have noise reduction turned off on both bodies. I'm just surprised that there's such a marked difference in performance between two nominally identical bodies. I want to do star trails next time we get a clear night sky, so I'll be choosing the less-noisy body!

[itsnotarace]

daveake said:

Dunno if that makes a difference or not. Also, I wonder if the noise is actually thermal noise picked up from dust on the sensor - certainly at least one of those bodies (I don't recall which one) is in need of a clean. Next step then I guess is to clean the sensor and see if the noise improves.

Won't make a difference in the slightest, those are "hot pixels". You can get them mapped out by Canon/Nikon/Fixation (but not Jessops etc)

[daveake]

Thanks

[RobDickinson]

Hm does dark field subtraction deal with hot pixels I wonder?


Anyhow if your doing single shot exposures of over 1min or so you should have long exp NR switched on.

The best way of doing star trails is multiple shorter shots though I think.

[daveake]

RobDickinson said:

Hm does dark field subtraction deal with hot pixels I wonder?


Anyhow if your doing single shot exposures of over 1min or so you should have long exp NR switched on.

The best way of doing star trails is multiple shorter shots though I think.

Well, the problem of course with having the camera do the NR is that if say I have a 10 minute exposure, I have to wait another 10 minutes for the camera to do the NR. Not ideal. That's why I was interested in whether there was a significant performance difference between the two bodies. Since one body is indeed a lot less noisy, that's the one I should use for such shots. It now carries a couple of small coloured labels so I know which is which

Yes, star trails are better with shorter exposures, say 30 seconds. I think the noisier body will still show some noise even then. I'll try.

Dave