Stacking data from multiple nights

This is a question that I see all the time on forums, so I thought it would be useful to share some thoughts. I am not an expert on this, so some of this information will be based on (lots) of reading of people who know far more than I do. I will update it as I continue to learn over time.

There are a few reasons (at least) that you might have multiple sets of data to integrate

• A multi-night project that you continue to add data too
• You are shooting a target like M42 or M31 which has a high range of brightness (high dynamic range)
• You are shooting with different filters

In the HDR case, you would probably have different length subs (from as short as a second, upwards toward minutes). In the longer subs, much of the bright data might be clipped, and you would use HDR techniques to blend this into a single image that contains the best of the data to generate an image that is not possible with a single set of subs.

In the case of different filters, you would integrate each filter into its own master light (i.e. all the Red, or all the Ha) and could use the techniques below to do this.

There are two main ways

1. Integrate subs for each session into one master. Integrate the masters
2. Calibrate all subs (per filter), then integrate into a single master (per filter)

My understanding is #2 usually gives the best signal and it is the one I am going to describe.

The other benefit(s) of #2

1. You can weight all your images together and discard bad subs globally. This might be a benefit if you had some less-than-great data in one of your first nights and wanted to remove that and replace with new subs.
2. Since you will likely be integrating a larger number of subs, you might be able to use better pixel rejection algorithms and get a cleaner stack – in DSS, it will recommend the best algorithms to use and one of the things that influences this is the number of light frames you are integrating.

Method #2 does require you to either keep calibrated subs around for long periods, or keep the original subs and all calibration frames. Not a problem for me as I am a hoarder, but that might not be the case for everyone

First, I will cover the basic process, then give some tips for Deep Sky Stacker on how to do this.

The basic technique is. For each session:

Acquisition:

• I recommend creating a completely separate directory per session (nest them below a top-level directory if you want to keep them all together)
• Shoot your light frames as normal
• Shoot your calibration frames as per your normal workflow (Darks, Flats, Flat Darks)
  • Some people like to re-use flats or darks over multiple sessions. I re-take them but both ways can work.
• If you shoot calibration frames each session, I add them to a “Calibration” directory in the same place as the lights. If you only shot one set, it might make sense to put them at the top level
• If you don’t have your Bias frames, shoot those and store them somewhere

Processing:

In DSS, this is now quite simple, but the user interface is not immediately obvious at first.

Here is where your files would usually show up (in the Main Group) – Lights, Bias, Darks, Flats etc.

What many people (myself included) miss is the “Main Group” tab – what could that mean ? Since most of us start off very simply, we put all our files in there and move on. This is a fine way to integrate one set of frames.

But now look what happens if I add an image. I am going to add a single bias frame to make this easy to see, and also there is a reason I picked a bias.

Here is how the UI looks now

Look inside the red box and we see Group1 – well that looks interesting. We can also see our single bias frame in the Main Group.

Lets select Group 1 and see what happens

With Group1 selected, we see we now have an empty file list that can have its own Lights, Darks and Bias frames. Easy, right ?

Well, there is a gotcha (which I did not know about the first time I did this). It’s documented but the UI doesn’t make this clear.

When you have multiple tabs, you do indeed put each sessions images (both lights and calibration images) in its own tab.

However, the Main Group is special! This is what I missed the first time. Any frames in the Main Group will be SHARED across all the other tabs. What use is that you might think? I think it is definitely confusing and the UI could warn you or give you guidance.

The primary reason is so that you don’t have to add shared calibration files in multiple tabs. Now, what kind of calibration files might be shared ?

For me it’s bias frames only (if I am using them). But if you have a darks library (very common for people with cooled astro-cams), then it can be darks as well. Also if you are comfortable not taking flats every night, then it might be flats also. Just bear a few things in mind

• Bias can change over time, so if your data spans multiple sets of Bias frames, you should add them to every tab (EXCEPT Main Group!)
• If your subs are different exposures, then your darks will also likely be different exposures. To be safe, you should add the darks into each group.
• The same goes for flats if they are different exposures (common if you take sky flats for instance)

So the process is

• Add any shared calibration frames to Main Group. As you do, the Group 1 tab will appear
• Add your session-specific calibration frames to Group 1, as well as your lights. The Group 2 tab will appear as soon as you have added your first file to Group 1
• Repeat until all your subs are added
• You will need to stack the data separately for each filter ! As far as I know, DSS doesn’t know about filters

Now you can integrate as usual. The one frustration here is that I don’t know of a place to see details such as image score for all subs in one place, so you need to click each tab.

Once you have added everything and integrated, you will get your master light. Do that once per filter (if appropriate) and then you can process as normal.

There is one caveat with the HDR mode (only a small number of DSO targets usually need this). If you do HDR then there are two ways to do it

• Stack each exposure separately (i.e. make a 1s, 10s and 180s stack) – these could still have multiple nights of data
• Use DSS’s HDR mode

The first way (multiple master lights) is best if you will use another tool such as Photoshop or PixInsight to HDR combine the images.

DSS can also do this for you (although you wont get any control over how it does, and I don’t know how good of a job it does). To do this, click the “Stacking parameters” when you go to register and stack your images

Then select the “Lights” Tab

You can then select the “Entropy Weighted Average” option. I like to make multiple masters in this case and combine them elsewhere but this is an option if you want a very simple process

DSS can also output the calibrated (or calibrated and star aligned) lights (i.e. one file for each sub). However, it’s not clear if DSS can stack some images that are calibrated with new images that are uncalibrated.

There is also an important caveat here (which DSS mostly hides, but other tools may not). If you stack calibrated frames from multiple nights by using Pre-calibrated AND REGISTERED frames, you MUST make sure they are all aligned to the same reference image or they wont align. For DSS, I recommend just not using the intermediate files and keeping it simple. For other tools, there are ways to address that.

I suspect if you simply add those files to a tab Group with no calibration files in the Main group or that group, then it might save you re-calibrating older data but I am not sure and I’d probably not take the risk. If you do want these files, just select the Intermediate files tab in Stacking Parameters and enable the options.

Hopefully this will help some of you as you begin your journey into astrophotography. As always, I am happy to correct any errors in my posts or add new content, so please let me know if you spot any errors or have suggestions.

This is definitely only an introduction and I am sure I will learn more as I get better. I have moved to Pixinsight for all of my stacking and processing, so at some point I will probably write up how this can work, although there are many people who are far more expert than me in Pixinsight.