Stacking - need help

Started Jan 7, 2013 | Discussions thread
DuncanDovovan
Senior MemberPosts: 1,277
Like?
Re: Stacking - need help
In reply to DuncanDovovan, Jan 14, 2013

DuncanDovovan wrote:

BlueFightingCat wrote:

I recently discovered that I can stack several exposures of the same starscape to reduce the effective ISO. Essentially I just want to get a photo of the stars (not star trails) with as little noise as possible.

Scenario: I take 32 exposures at 30 seconds each with ISO 6400 and F2.8.

Questions:

1. Somebody said that by taking 32 exposures at ISO 6400 and stacking the images I would end up with an effective ISO of 400. Can somebody explain the math behind that. I don't quite understand the relationship. Likewise the same person said if I took 128 exposures I would get an effective ISO of 100.

2. How does stacking work taking into account the earth's spin. If I take 32 images with a static foreground won't the stars be in a slightly different location in each image compared to the foreground? Does the stacking software take care of this?

BFC

1) Math =

1 times 480 sec @ ISO 400 =
1*2=2 times 240 sec @ ISO 800 =
2*2=4 times 120 sec @ ISO 1600 =
4*2=8 times 60 sec @ ISO 3200 =
8*2=16 times 30 sec @ ISO 6400

But this would not be entirely true, because the higher the ISO, the more worse noise becomes (not linear). Therefore to get that rule of thumb, you double all:

1 times 480 sec @ ISO 400 =
4 times 240 sec @ ISO 800 =
8 times 120 sec @ ISO 1600 =
16 times 60 sec @ ISO 3200 =
32 times 30 sec @ ISO 6400

Or for your other one:

1 times 480 sec @ ISO 400 =
4 times 240 sec @ ISO 800 =
8 times 120 sec @ ISO 1600 =
16 times 60 sec @ ISO 3200 =
32 times 30 sec @ ISO 6400 =
64 times 60 sec @ ISO 3200 =
128 times 30 sec @ ISO 6400

I did some tests and indeed the results are similar for these.
So if you have a good tracking device, you could theoretically avoid the stacking and just take a very long exposure.
But:
a) You have more risks to capture garbage like clouds, birds, airplanes, satellites, camera shake, car lights etc. When you take more exposures, you can leave those out. In 1 exposure, you just have it on your exposure.b) The longer the exposure, the more precise your polar alignment must be. At some exposure lengths, this can only be done by using active tracking of a star on a tracking device. With multiple exposures, your tracking needs to be accurate enough for 1 exposure, because the stacking software can align the rest.c) For multiple exposures you can take 8 dark shots to subtract them in post processing. This means the total time you use for dark frames can be less than the total time for the exposures. If you take 1 big exposure, your dark frame will take as much as the exposure. So you can do less in 1 night.
2) Yes they do. But you will get the best results if you use a tracker like the Polarie to minimise differences between the shots. Depending on the focal length you use, the frame that all 32 or 128 frames will have in common (= stars visible on all frames) can be very small. With a tracker you avoid that.
-- hide signature --

I'm a peaceful person looking for clarity and harmony and willing to help. If you believe my post had the intention to harm you in some way, you are probably wrong. =;-) Please verify first before assuming something bad happened to you. I can be a bit too brief at times. =;-)

There is a mistake in the 2nd table here is the corrected version:

1) Math =

1 times 480 sec @ ISO 400 =
1*2=2 times 240 sec @ ISO 800 = 
2*2=4 times 120 sec @ ISO 1600 =
4*2=8 times 60 sec @ ISO 3200 =
8*2=16 times 30 sec @ ISO 6400

But this would not be entirely true, because the higher the ISO, the more worse noise becomes (not linear). Therefore to get that rule of thumb, you double all:

1 times 480 sec @ ISO 400 =
4 times 240 sec @ ISO 800 = 
8 times 120 sec @ ISO 1600 =
16 times 60 sec @ ISO 3200 =
32 times 30 sec @ ISO 6400

Or for your other one:

1 times 480 sec @ ISO 100 =
4 times 240 sec @ ISO 200 = 
8 times 120 sec @ ISO 400 =
16 times 60 sec @ ISO 800 =
32 times 30 sec @ ISO 1600 =
64 times 60 sec @ ISO 3200 = 
128 times 30 sec @ ISO 6400

I did some tests and indeed the results are similar for these.
So if you have a good tracking device, you could theoretically avoid the stacking and just take a very long exposure at low ISO.
But:
a) You have more risks to capture garbage like clouds, birds, airplanes, satellites, camera shake, car lights etc. When you take more exposures, you can leave those out. In 1 exposure, you just have it on your exposure.b) The longer the exposure, the more precise your polar alignment must be. At some exposure lengths, this can only be done by using active tracking of a star on a tracking device. With multiple exposures, your tracking needs to be accurate enough for 1 exposure, because the stacking software can align the rest.c) For multiple exposures you can take say 8 dark shots to subtract them in post processing. This means the total time you use for dark frames can be less than the total time for the exposures. If you take 1 big exposure, your dark frame will take as much as the exposure. So you can do less in 1 night.
2) Yes they do. But you will get the best results if you use a tracker like the Polarie to minimise differences between the shots. Depending on the focal length you use, the frame that all 32 or 128 frames will have in common (= stars visible on all frames) can be very small. With a tracker you avoid that.
-- hide signature --

I'm a peaceful person looking for clarity and harmony and willing to help. If you believe my post had the intention to harm you in some way, you are probably wrong. =;-) Please verify first before assuming something bad happened to you. I can be a bit too brief at times. =;-)

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow