Lunar Stacking...

Messages
398
Name
Andy
Edit My Images
Yes
I've got a question about lunar stacking. This might be totally obvious... But last night I had very little battery and just done a quick 10 shots of the moon which I plan on stacking using registax... I understand that these are merged to bring out more detail...

I also understand that for noise reduction focus stacking looks for the higher signal colours and makes these stronger so to speak and dampens noise..

If I am stacking for more detail. What's the difference between taking 2 identical photos (tripod and a burst shot) and copying the image and stacking it as a duplicate? (is the only difference the noise in the image appears in a different location?)

I know this is a very newbie like question, but I don't fully understand how it works.

Thanks.
 
  • Like
Reactions: JPL
Noise is random, so is different in each photo. Atmospheric distortions are random, so vary in each photo. So if you take 10 shots and sum the values of each pixel and divide by 10 to get the average then you'll be reducing the proportion of noise and atmospheric distortions and increasing the proportion of detail. There's more to it than that in the more sophisticated astronomical stackers because they can also take advantage of the sub-pixel dithering introduced by slight camera movements and slight atmospheric refraction movements to increase resolution, but thinking of the process simply in terms of averaging out the noise and atmospheric distortions is the big simple primary idea.

For example, if you take a very long exposure of a busy street scene, say ten minutes, then all the moving things such as pedestrians and vehicles will smear out and disappear leaving you with a photograph of an empty street. Same idea.
 
Great question and great answer. I have for a long time wondered why astrophotographers stack and this relatively simple answer puts it in a way I can easily understand, thanks.
 
Last edited:
Makes sense of it for me too. Thanks Chris
 
Less noise is potentially achievable as you will average out the noise. More details - no chance. Only less. You are not getting even close to the clever sensor shift image merging.

Also the moon moves rather quickly through the sky. Failing to align all images perfectly you won't even stack anything successfully. Finally, for the same reason you are limited to very fast exposure times, ideally well over 1/100s. More exposure = more blur in the moon unless you have a tracker device.
 
Less noise is potentially achievable as you will average out the noise. More details - no chance. Only less. You are not getting even close to the clever sensor shift image merging.

But reducing noise and transient atmospheric distortions will make more of the detail visible, even just with simple averaging.

Also the moon moves rather quickly through the sky. Failing to align all images perfectly you won't even stack anything successfully.

But you don't necessarily have to do the alignment. Some image stacking software will do it. In fact some cameras will take multiple short exposure images and then stack them with auto image alignment done in the camera as a technique for reducing noise and improving detail in hand held night shots.

Finally, for the same reason you are limited to very fast exposure times, ideally well over 1/100s. More exposure = more blur in the moon unless you have a tracker device.

I have a 500mm lens which on my crop sensor camera has a horizontal angle of view of 3 degrees. The moon subtends a visual angle of half a degree. That means in my 24MP (6000 x 4000 pixels) image the moon's diameter occupies 1000 pixels. The moon moves its own diameter in the sky in one hour. So if my arithmetic is correct it will move one pixel in about 1/3rd of a second. Let's say my lens etc. is so extremely good that I actually have one pixel sharp detail resolution. Let's make the further conservative assumption that the moon moving 1/4 of a pixel will noticeably dull the sharpness of my incredibly sharp moon image. That means I can safely photograph the moon (with my 500mm lens on my 24MP crop sensor) at any shutter speed above 1/12th of second.

Since the moon is as bright as a sunny day, that's not going to be a problem.

Of course my arithmetic could be wrong. I look forward to being corrected :)
 
Of course my arithmetic could be wrong. I look forward to being corrected :)
OK then. Since you asked.

The moon moves its own diameter in the sky in one hour.
Wrong. The moon moves its own diameter in the sky in a little over 2 minutes.

So if my arithmetic is correct it will move one pixel in about 1/3rd of a second.
Wrong. If the moon moved its own diameter in an hour - which it doesn't - then on your figures that would be 1/3rd of a pixel per second, not 1/3rd of a second per pixel.

The first error means you're out by a factor of 30. The second error is a factor of 10 - but it's in the opposite direction, so overall you're only out by a factor of 3, and your conclusion is till sound.


Here's what you would have written if you had got the arithmetic correct:

I have a 500mm lens which on my crop sensor camera has a horizontal angle of view of 3 degrees. The moon subtends a visual angle of half a degree. That means in my 24MP (6000 x 4000 pixels) image the moon's diameter occupies 1000 pixels. The moon moves its own diameter in the sky in two minutes. So it will move eight pixels in one second. Let's say my lens etc. is so extremely good that I actually have one pixel sharp detail resolution. Let's make the further conservative assumption that the moon moving 1/4 of a pixel will noticeably dull the sharpness of my incredibly sharp moon image. That means I can safely photograph the moon (with my 500mm lens on my 24MP crop sensor) at any shutter speed above 1/30th of a second.
 
Last edited:
[I asked if someone would correct my arithmetic about avoiding blur in moon shots due to the movement of the moon.]

OK then. Since you asked.

Wrong. The moon moves its own diameter in the sky in a little over 2 minutes.

You're right. It moves a tad less than 15 degrees in an hour.

Wrong. If the moon moved its own diameter in an hour - which it doesn't - then on your figures that would be 1/3rd of a pixel per second, not 1/3rd of a second per pixel.

The first error means you're out by a factor of 30. The second error is a factor of 10 - but it's in the opposite direction, so overall you're only out by a factor of 3, and your conclusion is till sound.

Here's what you would have written if you had got the arithmetic correct:

I have a 500mm lens which on my crop sensor camera has a horizontal angle of view of 3 degrees. The moon subtends a visual angle of half a degree. That means in my 24MP (6000 x 4000 pixels) image the moon's diameter occupies 1000 pixels. The moon moves its own diameter in the sky in two minutes. So it will move eight pixels in one second. Let's say my lens etc. is so extremely good that I actually have one pixel sharp detail resolution. Let's make the further conservative assumption that the moon moving 1/4 of a pixel will noticeably dull the sharpness of my incredibly sharp moon image. That means I can safely photograph the moon (with my 500mm lens on my 24MP crop sensor) at any shutter speed above 1/30th of a second.

Quite right! How embarrassing to have made such a mess of it!

At my age there is always the suspicion that sufficient marbles to derail proper cognitive functioning may have fallen out of my brain. My excuse is that having temporarily reduced mobility and in pain due to a bad back, I started composing the post in a noisy kitchen well supplied with champion interrupters. Sometimes it would take three of four goes spaced out with several minutes of irrelevant political fencing to finish one sentence. I started my arithmetical argument based on the moon moving nearly 15 degrees an hour, and later decided it would be more easily understood if I cast the argument in terms of the moon's movement along its own diameter, assuming as a simple starting point that the moon's image would be 1,000 pixels across.

In the noisy confusion I accomplished some of the necessary edits, thought I'd done some others, and ended up irritably losing patience and posting the thing without proof reading it. After all, the conclusion seemed to be in the right ball park!

We'll have to wait and see if I do this kind of thing again. I've noticed that age-related cognitive decline is not uniform. Some days, some hours, the old fool is quite sharp; at other times hopelessly confused.
 
Back
Top