AstroPhotography - stacked exposures

Messages
5,335
Name
matt
Edit My Images
Yes
Right, so I understand the principle of stacking multiple images of astro shots, but do I stack a number of correctly exposed shots. Or do I take a load of shots that are almost pitch black and they become a single correctly exposed shot. My simple mind seems to think it's the former and of course correctly exposed could mean stacking different exposure timed "subs" to get the whole thing correctly exposed.
Last night for instance I took a few shots at 1/60th, a few at 10 seconds and some more at 30 seconds. The ones at 1/60 were almost all black, the ones at 10 and 30 seconds contained quite a lot of detail. All of this was just a test of whether the tracking was sufficiently accurate to not need guiding, I appreciate that it wasnt a definitive test as I only tried a 50mm lens, but I was also testing my remote application would allow me to control the camera from the living room.
I have an EQ5 mount with RA and DEC motors, I didnt mount my scope just the camera and a lens for this test, but the results got me wondering what is the correct procedure regarding stacking the resultant images.

Matt
 
I was kind of hoping someone would respond to this as I too am interested.
 
I suspect you need correctly exposed images, there are different stack modes in photoshop, which give different effects (never tried them with astro though) I think a lot of people use star stacker though, not sure what different modes that has.
 
Usually you will stack correctly exposed shots in order to improve clarity in the whole image. It becomes a little more complex when you have a foreground that you want to appear in the shot.
In that case you will need to expose the foreground correctly, whilst not worrying about the sky as you have your sky shots separately.
 
I presume since the aim is to use the scope you're talking deep sky rather than wide field/Milky Way? When I do astro imaging I use either a 200mm scope or a camera lens (usually the 150-600mm). The scope isn't mine. Neither the scope's setup nor my own setup is guided (tracking yes, guiding no) so the maximum I can go to is about 1 min. Fainter objects all get 1 min subs (but usually 60 - 70 of them) and if I'm lucky I can see something, but not much, on a single sub. Brighter objects may get less exposure and may also get varying exposure lengths. The software I use has the facility to do an HDR type integration. I will just say that I use PixInsight, which is dedicated astrophotography software. I have used DSS but in general I was dissapointed in the results (but it's free so I'm not knocking it at all - it was probably the idiot with her hand on the mouse........). I've never tried integrating (stacking) with Photoshop (I don't use it). If you want to take a look at the Astronomy album on my Flickr the details are there, certainly on the more recent stuff. You may (or may not) find it useful. You don't say what you imaged but probably forget your 1/60 exposure, but bear in mind all my images are taken at 800 iso as the cameras I'm using are fairly old and low end so high iso noise becomes an issue.
 
Define "correctly exposed" in this context. Think about it.
In my view it would be where I have enough detail, which in the case of stars would be a white spot, in the case of the moon/planets would show more detail. In some of my star shots, the 1/60 exposure there is little to no view of the stars, would stacking bring out more detail?
 
I presume since the aim is to use the scope you're talking deep sky rather than wide field/Milky Way? When I do astro imaging I use either a 200mm scope or a camera lens (usually the 150-600mm). The scope isn't mine. Neither the scope's setup nor my own setup is guided (tracking yes, guiding no) so the maximum I can go to is about 1 min. Fainter objects all get 1 min subs (but usually 60 - 70 of them) and if I'm lucky I can see something, but not much, on a single sub. Brighter objects may get less exposure and may also get varying exposure lengths. The software I use has the facility to do an HDR type integration. I will just say that I use PixInsight, which is dedicated astrophotography software. I have used DSS but in general I was dissapointed in the results (but it's free so I'm not knocking it at all - it was probably the idiot with her hand on the mouse........). I've never tried integrating (stacking) with Photoshop (I don't use it). If you want to take a look at the Astronomy album on my Flickr the details are there, certainly on the more recent stuff. You may (or may not) find it useful. You don't say what you imaged but probably forget your 1/60 exposure, but bear in mind all my images are taken at 800 iso as the cameras I'm using are fairly old and low end so high iso noise becomes an issue.
I may or may not use my scope as the lens as I find the 5D3 a bit too heavy in that it doesn't feel as secure as I would like when attached to the scope (skywatcher ed80). I have a choice of lenses up to 400mm and TC's (1.4 & 2) so I could do DSO and/or some planetary stuff I guess. Thanks for your post above, very helpful, I will check out your Flickr account.
 
Last edited:
I find the 5D3 a bit too heavy in that it doesn't feel as secure as I would like when attached to the scope (skywatcher ed80).

That's interesting as that might have been the one I was eyeing up at Astrofest (and decided to think about it a bit/a lot longer). Admittedly my 350D is fairly light. My Sigma 150-600C works superbly with the crop sensor camera on my lightweight EQ3-2 mount for the larger DSOs. You'd be pushed to do planetary with what you have. You need huge magnification to bring out any details.
 
Define "correctly exposed" in this context. Think about it.
In my view it would be where I have enough detail, which in the case of stars would be a white spot....
In the case of which stars?

A common estimate for the number of stars in the observable universe is about 10^22 - 100,000,000,000,000,000,000,000 - though that could be out by one or two orders of magnitude in either direction. How many of them do you expect should show up as a white spot? Now what do you mean by "correctly exposed"?
 
In the case of which stars?

A common estimate for the number of stars in the observable universe is about 10^22 - 100,000,000,000,000,000,000,000 - though that could be out by one or two orders of magnitude in either direction. How many of them do you expect should show up as a white spot? Now what do you mean by "correctly exposed"?
Ok let me put it another way, I have several photos of an area of the sky, each frame has 20 very dim pin pricks of light, if I stack these shots will more pin pricks of light be visible than the 20 I could originaly see.
 
I have several photos of an area of the sky, each frame has 20 very dim pin pricks of light, if I stack these shots will more pin pricks of light be visible than the 20 I could originaly see.
Are they always the same 20 very dim pin pricks?
 
Are they always the same 20 very dim pin pricks?
Yes it's the same area of the sky in each photo, same camera settings in each shot, multiple exposure of same area, moved slightly by tracking tripod mount, so each pin prick is in exactly the same location in each shot, well, to within tripod mount/motor movement tolerances.
 
Yes it's the same area of the sky in each photo, same camera settings in each shot, multiple exposure of same area, moved slightly by tracking tripod mount, so each pin prick is in exactly the same location in each shot, well, to within tripod mount/motor movement tolerances.
I'd be surprised if it is the same 20 pin pricks in every shot. Here's why.

What the camera is basically doing is capturing and counting photons, right? The more photos that are captured by a given pixel on the sensor, the brighter that pixel appears on the final image. But the trouble with astro work is that there aren't many photons from any individual star. In 1/60th of a second you might not get any photons at all from a particular star, and so the location of that star will appear black on the sensor. However ... suppose the brightness of that star is such that your sensor captures, say, 10 photons per second on average. In a single exposure of 1/60th there's only a 1-in-6 chance that it will capture a photon, so the image will probably appear black. In a 1-second exposure it will capture about 10 photons. But if you take 60 exposures at 1/60th each, there will be 10 photons distributed across the 60 exposures - probably 50 frames with none and 10 frames with one, though there's a small chance that one exposure might capture two.

So I think the really faint stars will be just barely registering on some shots but not on all of them.

Obviously there's an issue distinguishing between faint starts and random noise, but that's a different conversation.
 
Last edited:
Hopefully not hijacking this thread. But how do you do panoramas at night when it's pitch black. Don't understand how you know how to overlap when you can't see anything on the back of the screen?
 
I'd be surprised if it is the same 20 pin pricks in every shot. Here's why.

What the camera is basically doing is capturing and counting photons, right? The more photos that are captured by a given pixel on the sensor, the brighter that pixel appears on the final image. But the trouble with astro work is that there aren't many photons from any individual star. In 1/60th of a second you might not get any photons at all from a particular star, and so the location of that star will appear black on the sensor. However ... suppose the brightness of that star is such that your sensor captures, say, 10 photons per second on average. In a single exposure of 1/60th there's only a 1-in-6 chance that it will capture a photon, so the image will probably appear black. In a 1-second exposure it will capture about 10 photons. But if you take 60 exposures at 1/60th each, there will be 10 photons distributed across the 60 exposures - probably 50 frames with none and 10 frames with one, though there's a small chance that one exposure might capture two.

So I think the really faint stars will be just barely registering on some shots but not on all of them.

Obviously there's an issue distinguishing between faint starts and random noise, but that's a different conversation.

So bottom line is either technique should work but in different ways?
1) 1/60 with say 500 shots will stack into possibly a single image showing lots of stars, some faint some bright depending on how many photons were captured for each star.
2) 10 or 30 second exposures stacked will show similar results if stacking less frames, say 20 of each, as each frame has already collected a "lot" of photons.
 
That's interesting as that might have been the one I was eyeing up at Astrofest (and decided to think about it a bit/a lot longer). Admittedly my 350D is fairly light. My Sigma 150-600C works superbly with the crop sensor camera on my lightweight EQ3-2 mount for the larger DSOs. You'd be pushed to do planetary with what you have. You need huge magnification to bring out any details.

I have the Skywatcher ED80 (about £450 now) and took a long time deciding what I needed within a reasonable budget without going right over the top, as in all photography the law of diminishing returns seems high. I didnt want a reflector, too much faff, I wanted something reasonably quick to setup and tbh with rough polar alignment it seems pretty good (just pointing the main leg North - using a compass and at the correct elevation angle seems to work ok). I could do a proper polar alignment but it's not fully dark until quite late now so a rough one seems to do ok for short exposures and tracking well enough to keep the Moon in the middle of the scope.
I've seen some quite acceptable results with an 80mm ED Refractor posted on the net for Mars, Jupiter and Saturn, certainly good enough for what I want without having to pay a King's ransom for kit and finding only 3 nights in the year provide good "seeing" conditions no doubt whilst I would be on holiday and away from my kit. I'm also keen on photographing the Moon which shouldnt need too much magnifications as its so flippin big. :)
 
So bottom line is either technique should work but in different ways?
1) 1/60 with say 500 shots will stack into possibly a single image showing lots of stars, some faint some bright depending on how many photons were captured for each star.
2) 10 or 30 second exposures stacked will show similar results if stacking less frames, say 20 of each, as each frame has already collected a "lot" of photons.
I think so, in theory, roughly. One minute of exposure is one minute of exposure, regardless of whether it's one continuous exposure or (say) 600 exposures of 1/10th of a second each. The same numbers of photons are counted.

However the big potential gremlin here is noise. I would be worried that the "signal" in short exposures might be swamped by the noise. Unless you were trying to use some sort of speckle imaging technique, I can't see any obvious advantage to using short exposures. Personally I think I would calculate the longest exposure time that I could use without the image exhibiting star trailing, and then use that time for each exposure.
 
Personally I think I would calculate the longest exposure time that I could use without the image exhibiting star trailing, and then use that time for each exposure.

Spot on. Longer exposures are better, even more so on DSOs like nebulae and galaxies which have a lot of faint detail. I'm limited to 1 min max but my partner who uses a guided 250mm reflector will use 3 min subs or longer. Shooting dark frames of the same exposure length but with the lens cap on (they should also be at the same temperature so do them before you go indoors) also helps reduce noise.
 
Thanks for the information.
 
I've yet to try this but I recently heard of an application called Sequator which apparently makes stacking astro shots much more straightforward. The website gives some guidance on how to effectively capture exposures to work well with the software. It suggests to take some dark frames with the lens cap in place and can automatically handle the noise subtraction with them.

https://sites.google.com/site/sequatorglobal/tutorial
 
Back
Top