Page 1 of 1

long exposure noise reduction.

Posted: Wed Dec 27, 2017 7:33 am
by andre paul
hi there.
i find this topic interesting.....
for those doing long exposure photography, how do you deal with noise reduction : in camera ou pos processing ? why ? any special tips ? 
lets say exposures  from 2 to 8 minutes .......

regards to all here.....
andre

Re: long exposure noise reduction.

Posted: Wed Dec 27, 2017 2:51 pm
by DChan
New to long exposure photography myself and so have been reading up on it. Here's an article about it:

COMPLETE GUIDE TO LONG EXPOSURE PHOTOGRAPHY – 2016 EDITION


Here's what he said re noise:
...modern day DSLR sensors have an amount of megapixels that were within the reach of medium format cameras only, until just recently. DSLR camera sensors can now easily have 35 MP or more and the quality of photos by DSLR’s when used in a regular way is fantastic. But when used in extreme and more rare conditions as in long exposure photography, then those sensors don’t seem to hold up very well and will cause a lot of noise. Even with shorter exposures...

First thing to do is to try to avoid noise by shooting at the lowest ISO (100) and if there’s still noise it’s easy to just get rid of it by using noise reduction software, in normal situations. The noise produced with the new generation of 35+ MP cameras however is so much that using noise reduction software won’t be an easy job anymore. It can be even so bad that you can’t remove the noise without accepting reduction of image quality and details at the same time.  At the moment of writing I don’t have an idea what camera manufacturers will do with the noise issue with their high MP cameras when used for long exposure photography but my suggestion would be to not use or purchase those cameras if you’re using it for long exposure photography predominantly. You’re better off then with the older generation of lower MP cameras...

...in-camera NR doubles the time of the exposure and since I usually shoot with exposure times of 5 minutes or longer this would increase exposure times for just one shot up to more than 30 minutes.
That's the most detailed guide on long exposure photography I've found so far.

Re: long exposure noise reduction.

Posted: Wed Dec 27, 2017 5:14 pm
by andre paul
DChan wrote:New to long exposure photography myself and so have been reading up on it. Here's an article about it:

COMPLETE GUIDE TO LONG EXPOSURE PHOTOGRAPHY – 2016 EDITION


Here's what he said re noise:
...modern day DSLR sensors have an amount of megapixels that were within the reach of medium format cameras only, until just recently. DSLR camera sensors can now easily have 35 MP or more and the quality of photos by DSLR’s when used in a regular way is fantastic. But when used in extreme and more rare conditions as in long exposure photography, then those sensors don’t seem to hold up very well and will cause a lot of noise. Even with shorter exposures...

First thing to do is to try to avoid noise by shooting at the lowest ISO (100) and if there’s still noise it’s easy to just get rid of it by using noise reduction software, in normal situations. The noise produced with the new generation of 35+ MP cameras however is so much that using noise reduction software won’t be an easy job anymore. It can be even so bad that you can’t remove the noise without accepting reduction of image quality and details at the same time.  At the moment of writing I don’t have an idea what camera manufacturers will do with the noise issue with their high MP cameras when used for long exposure photography but my suggestion would be to not use or purchase those cameras if you’re using it for long exposure photography predominantly. You’re better off then with the older generation of lower MP cameras...

...in-camera NR doubles the time of the exposure and since I usually shoot with exposure times of 5 minutes or longer this would increase exposure times for just one shot up to more than 30 minutes.
That's the most detailed guide on long exposure photography I've found so far.
thanks dchan!
i am a bit into long expore shots ;-)
i know very well joel's work for a longtime ....
i am actually looking for other photographes experiences dealing with long exposure noise, not for lessons/tutorials ... 
there are several ways in dealing with noise...

but thanks anyway for your answer! apreciate it ...

Posted: Wed Dec 27, 2017 6:41 pm
by E.J. Peiker
One tip, if you don't want to wait the 2-8 minutes between shots is to do your own dark frame subtraction. What you do is you take your shots and sometime during your sequence, take a single shot of the same approximate exposure length with the lens cap on and something dark draped over the lens opening, perhaps a neoprene sleeve or something to prevent any stray light that the lens cap might let in. Then handle the dark frame subtraction in post - a Google search will show you how to do that easily. That way you only lose the additional 2-8 minutes once.

Re:

Posted: Thu Dec 28, 2017 12:50 am
by DChan
E.J. Peiker wrote:One tip, if you don't want to wait the 2-8 minutes between shots is to do your own dark frame subtraction.  What you do is you take your shots and sometime during your sequence, take a single shot of the same approximate exposure length with the lens cap on and something dark draped over the lens opening, perhaps a neoprene sleeve or something to prevent any stray light that the lens cap might let in.  Then handle the dark frame subtraction in post - a Google search will show you how to do that easily.  That way you only lose the additional 2-8 minutes once.
In post?? Sounds like you could even take those dark frame shots at home afterwards. Or even have some taken and put away for future use if the exact exposure time is not required.

Re: Re:

Posted: Thu Dec 28, 2017 8:10 am
by E.J. Peiker
DChan wrote:
E.J. Peiker wrote:One tip, if you don't want to wait the 2-8 minutes between shots is to do your own dark frame subtraction.  What you do is you take your shots and sometime during your sequence, take a single shot of the same approximate exposure length with the lens cap on and something dark draped over the lens opening, perhaps a neoprene sleeve or something to prevent any stray light that the lens cap might let in.  Then handle the dark frame subtraction in post - a Google search will show you how to do that easily.  That way you only lose the additional 2-8 minutes once.
In post?? Sounds like you could even take those dark frame shots at home afterwards. Or even have some taken and put away for future use if the exact exposure time is not required.
No, that won't work well.  It has to be taken in the same temperature conditions and it can change with time.  Long exposure noise can change significantly based on temperature and can change significantly with time.  For the best results, take one every time you are out and for each exposure time.  If you are shooting 3 minute exposures and take a three minute dark frame and then decide the rest of your exposures are 6 minutes for an extra stop of light, take a 6 minute one.  Also if you are out all night and lets say that your first shots are done around 15C and it cools off through the night and drops to 5C, redo them at 5C.  Also don't do the dark frame first, do it after at least one exposure or maybe even 2 so that the sensor is not colder than it would be after a long shot.  It's all about temperature!

High end digital backs even give you a sensor temperature to help you make decisions on whether you should do a dark frame or not, if you turned auto dark frame off.

Re: Re:

Posted: Thu Dec 28, 2017 4:10 pm
by andre paul
E.J. Peiker wrote:
DChan wrote:
E.J. Peiker wrote:One tip, if you don't want to wait the 2-8 minutes between shots is to do your own dark frame subtraction.  What you do is you take your shots and sometime during your sequence, take a single shot of the same approximate exposure length with the lens cap on and something dark draped over the lens opening, perhaps a neoprene sleeve or something to prevent any stray light that the lens cap might let in.  Then handle the dark frame subtraction in post - a Google search will show you how to do that easily.  That way you only lose the additional 2-8 minutes once.
In post?? Sounds like you could even take those dark frame shots at home afterwards. Or even have some taken and put away for future use if the exact exposure time is not required.
No, that won't work well.  It has to be taken in the same temperature conditions and it can change with time.  Long exposure noise can change significantly based on temperature and can change significantly with time.  For the best results, take one every time you are out and for each exposure time.  If you are shooting 3 minute exposures and take a three minute dark frame and then decide the rest of your exposures are 6 minutes for an extra stop of light, take a 6 minute one.  Also if you are out all night and lets say that your first shots are done around 15C and it cools off through the night and drops to 5C, redo them at 5C.  Also don't do the dark frame first, do it after at least one exposure or maybe even 2 so that the sensor is not colder than it would be after a long shot.  It's all about temperature!

High end digital backs even give you a sensor temperature to help you make decisions on whether you should do a dark frame or not, if you turned auto dark frame off.
if yu were doing about three images in an afternoon , 5 minute shots, (in a session) , would yu go in camera or pos processing route ? ;-)  if i had to shoot 10 or 15 images in camera NR would be a problem, thats true...
how do you compare results ? 
do you usually have some hotspots left to clone out when working pos editing route ?

Re: Re:

Posted: Thu Dec 28, 2017 6:32 pm
by E.J. Peiker
andre paul wrote:
E.J. Peiker wrote:
DChan wrote:
E.J. Peiker wrote:One tip, if you don't want to wait the 2-8 minutes between shots is to do your own dark frame subtraction.  What you do is you take your shots and sometime during your sequence, take a single shot of the same approximate exposure length with the lens cap on and something dark draped over the lens opening, perhaps a neoprene sleeve or something to prevent any stray light that the lens cap might let in.  Then handle the dark frame subtraction in post - a Google search will show you how to do that easily.  That way you only lose the additional 2-8 minutes once.
In post?? Sounds like you could even take those dark frame shots at home afterwards. Or even have some taken and put away for future use if the exact exposure time is not required.
No, that won't work well.  It has to be taken in the same temperature conditions and it can change with time.  Long exposure noise can change significantly based on temperature and can change significantly with time.  For the best results, take one every time you are out and for each exposure time.  If you are shooting 3 minute exposures and take a three minute dark frame and then decide the rest of your exposures are 6 minutes for an extra stop of light, take a 6 minute one.  Also if you are out all night and lets say that your first shots are done around 15C and it cools off through the night and drops to 5C, redo them at 5C.  Also don't do the dark frame first, do it after at least one exposure or maybe even 2 so that the sensor is not colder than it would be after a long shot.  It's all about temperature!

High end digital backs even give you a sensor temperature to help you make decisions on whether you should do a dark frame or not, if you turned auto dark frame off.
if yu were doing about three images in an afternoon , 5 minute shots, (in a session) , would yu go in camera or pos processing route ? ;-)  if i had to shoot 10 or 15 images in camera NR would be a problem, thats true...
how do you compare results ? 
do you usually have some hotspots left to clone out when working pos editing route ?
If the photography isn't time sensitive at all then I would just let the camera do it.  It is always the best way from the standpoint of final long exposure noise elimination but it isn't always practical or possible - in those cases, the single shot and then doing it in post is much better than not doing it at all.  This is especially true in situations where the sun is heating the camera more and more over time.

Posted: Fri Dec 29, 2017 6:01 am
by andre paul
one of these days i might do some testing noisewise with my d4 instead of using the d810 ....

Re:

Posted: Fri Dec 29, 2017 8:05 am
by E.J. Peiker
andre paul wrote:one of these days i might do some testing noisewise with my d4 instead of using the d810 ....
Realize that high ISO noise and long exposure noise are two completely different things and have little to do with one another.  High ISO noise is due to having to boost the signal to get an image since raising the ISO results in fewer photons being collected for a given exposure time thereby reducing the signal to noise ratio (which also reduces the dynamic range).  Long Exposure Noise is due to the sensor heating up resulting in what appears like stuck pixels.

Re: Re:

Posted: Fri Dec 29, 2017 4:54 pm
by andre paul
E.J. Peiker wrote:
andre paul wrote:one of these days i might do some testing noisewise with my d4 instead of using the d810 ....
Realize that high ISO noise and long exposure noise are two completely different things and have little to do with one another.  High ISO noise is due to having to boost the signal to get an image since raising the ISO results in fewer photons being collected for a given exposure time thereby reducing the signal to noise ratio (which also reduces the dynamic range).  Long Exposure Noise is due to the sensor heating up resulting in what appears like stuck pixels.
Yes i am aware  about noise categories 
I have been reading that lower count megapixels in fullframe sensor would have larger photosites ( i think thats the word..) and would generate less thermal noise then higher megapixel sensors....

Re: Re:

Posted: Fri Dec 29, 2017 8:31 pm
by E.J. Peiker
andre paul wrote:
E.J. Peiker wrote:
andre paul wrote:one of these days i might do some testing noisewise with my d4 instead of using the d810 ....
Realize that high ISO noise and long exposure noise are two completely different things and have little to do with one another.  High ISO noise is due to having to boost the signal to get an image since raising the ISO results in fewer photons being collected for a given exposure time thereby reducing the signal to noise ratio (which also reduces the dynamic range).  Long Exposure Noise is due to the sensor heating up resulting in what appears like stuck pixels.
Yes i am aware  about noise categories 
I have been reading that lower count megapixels in fullframe sensor would have larger photosites ( i think thats the word..) and would generate less thermal noise then higher megapixel sensors....
For the same process generation that would be true but all bets are off when you mix processor generations and even add in things like BSE, stacked CMOS, etc.  Too many variables to make a generalization like that.

Re: Re:

Posted: Fri Dec 29, 2017 9:37 pm
by andre paul
QokE.J. Peiker wrote:
andre paul wrote:
E.J. Peiker wrote:
andre paul wrote:one of these days i might do some testing noisewise with my d4 instead of using the d810 ....
Realize that high ISO noise and long exposure noise are two completely different things and have little to do with one another.  High ISO noise is due to having to boost the signal to get an image since raising the ISO results in fewer photons being collected for a given exposure time thereby reducing the signal to noise ratio (which also reduces the dynamic range).  Long Exposure Noise is due to the sensor heating up resulting in what appears like stuck pixels.
Yes i am aware  about noise categories 
I have been reading that lower count megapixels in fullframe sensor would have larger photosites ( i think thats the word..) and would generate less thermal noise then higher megapixel sensors....
For the same process generation that would be true but all bets are off when you mix processor generations and even add in things like BSE, stacked CMOS, etc.  Too many variables to make a generalization like that.
Ok! Got it!
When i return home i will test both cameras for comparison, without long expore NR to see how noise compares.
I guess testing os the only way to see how these cameras behave ...in similar conditions...

Posted: Sat Dec 30, 2017 10:20 am
by E.J. Peiker
Yes you can test them in similar conditions but the results may be totally different if it was a lot colder or a lot warmer.