Stacking old and new photos ?

[George Gearless]

Is it possible to stack photos from different sessions? Is it useful? I.e, does it make the photo better (more data)?

If yes; is there a specific technique involved or do you just stack like you normaly would and hope DSS (or whatever stacker you use) handles the discrepancies?

[vlaiv]

Yes, Yes/yes

That was the short answer. Here's the long one.

You can certainly stack subs from different sessions, and many people take multiple sessions to complete their data sets. You can also stack already stacked images, but this is not as good approach.

Ideally you want your subs to be of equal quality. Both in sharpness and in SNR. Stacking works the best if all frames are of equal SNR. "Regular stacking" that is. There are techniques that assign weights to subs based on achieved SNR. Even in course of the single evening SNR will change. There are multiple factors that determine SNR of each sub - sky transparency, LP levels (that change both in time and with position on the sky), atmospheric extinction (target moves across the sky and extinction is the least near zenith and much more lower in alt). For best results, stacking algorithm needs to account for this and assign weight to each sub.

Same goes for subs taken on multiple nights. Sometimes there will be more LP, sometimes less. Transparency will vary more between nights then on a single night, and of course, target position changes. There are also effects like presence of moon.

I'm not sure that DSS has algorithm that addresses all of that, but that does not mean it can't produce a stack of data from different sessions. It just means that under some circumstances results will not be as good as with software that does the above.

One of the reasons I mention that stacking stacks from each session is not good approach is because you lose ability to treat each sub that goes into final stack differently (assign different weight and anything else that is needed).

If you use simple average, then large stack of individual subs will be equal to stack of "partial stacks" from each session.

[fireballxl5]

good question as I collect data over many nights as and when the weather, Moon and time commitments allow. These I add to the stack for a given imaging project, usually with no issues. These projects usually make use of the same calibration file set.

A different situation I anticipate in the future is going back to a previous target with the same (or maybe different) kit but when a new set of calibration data is needed, e.g. after I have swapped cameras and re-taken flats. Best would be to re-stack all raws from the latest runs and the old ones, but these will need to be re-calibrated with 'legacy' flats/darks.

I use MaxImDL to stack my images and I don't think it can load and assign different flats/darks to different sub-sets of raws, or can it? Maybe I should start to save the calibrated subs to avoid this problem? Otherwise I would have to find a way to add the legacy stack to the new subs/stack, in order to maximise the benefit of using the old data. Advice here would be appreciated?

Hope this makes sense!

CS, Andy

[JRWASTRO]

2 hours ago, George Gearless said:

Is it possible to stack photos from different sessions? Is it useful? I.e, does it make the photo better (more data)?

If yes; is there a specific technique involved or do you just stack like you normaly would and hope DSS (or whatever stacker you use) handles the discrepancies?

Is it possible to stack photos from different sessions?

Answer: Yes provided that the image from one session to the next are perfectly aligned.

does it make the photo better (more data)?

Answer: It will make the image less noisy. Stacking relies on the property of random noise that says its average value is 0 (zero). So taking the average value of a number of frames will improve the image.  There is, however, a point of diminishing returns because the improvement is a function of the square root of (the number of frames averaged). Improvement for 16 frames is x4 where as the improvement for 25 frames is x5

[vlaiv]

2 minutes ago, fireballxl5 said:

good question as I collect data over many nights as and when the weather, Moon and time commitments allow. These I add to the stack for a given imaging project, usually with no issues. These projects usually make use of the same calibration file set.

A different situation I anticipate in the future is going back to a previous target with the same (or maybe different) kit but when a new set of calibration data is needed, e.g. after I have swapped cameras and re-taken flats. Best would be to re-stack all raws from the latest runs and the old ones, but these will need to be re-calibrated with 'legacy' flats/darks.

I use MaxImDL to stack my images and I don't think it can load and assign different flats/darks to different sub-sets of raws, or can it? Maybe I should start to save the calibrated subs to avoid this problem? Otherwise I would have to find a way to add the legacy stack to the new subs/stack, in order to maximise the benefit of using the old data. Advice here would be appreciated?

Hope this makes sense!

CS, Andy

Either save calibrated subs, or include masters with sets of subs. I do later, then I can calibrate raw subs when I want to have another go at stacking, or add more data, or whatever.

Not sure how Maxim does it, but if you have an option to only calibrate data without stacking - then save raws + calibration masters.

[Scooot]

When using a DSLR  and it’s live view to frame & focus with no laptop etc,  are there any techniques  that help to match the framing on a different night. I guess this is just trial & error and experience? 

[PeterCPC]

In BYE you can load an old image so that you can use it to align with what you are seeing. You need a laptop though.

Peter

[feilimb]

Dss allows the use of groups, for stacking data from different sessions. There are tabs below the main file list area for this I think. If you have calibration files (eg darks and bias, which are the same for different sessions then you can put them in 'Group 0' and leave them out of eg. Group 1 (at least I think that's how it works). 

[carastro]

I regularly use data from different imaging sessions and combine them.  even data from several years apart, and yes it improves the quality of the image. 

Quote

Yes provided that the image from one session to the next are perfectly aligned.

It's not necessary to be perfectly aligned, but the closer you get the better, otherwise the stacking software will rotate different sessions, and you will end up having to crop off the edges to produce a finished image.

Both APT and BYEOS have methods of aligning new images to previous ones in framing.  DSS has a method of stacking different sessions as described above, though in my experience I found I got a better result by stacking the stacks. 

I even combine data from different equipment (cameras and scopes), but you would have to stack separately, and register the images to each other before stacking the different stacks and yes I do lose a fair bit round the edges when I do this, but sometimes that doesn't matter too much as the main target is in the middle.

I use a dual rig with differing scopes and cameras, and my method is to align the cameras so they are looking at the same thing, and rotated the same.  Then stack the data separately and stack the stacks in Photoshop.  However as this is not quite what the original poster was asking I won't elaborate on this.

Carole 

 

 

[ollypenrice]

Registar is an excellent program for 'stacking stacks.' 

Vlaiv is right that running the entire set of subs of both sessions through in one go to make a new full stack is the best way but I find that stacking stacks is very little different in reality. 

It is not necessarily 'more detail' that you'll get from this operation. You'll get a better signal to noise ratio and so be able to stretch fainter signal into view and sharpen the bright details a little harder.

Olly

[George Gearless]

Thanks to all for some great responses. I'll check out my archives and see what I want to try and lump together.

I am still very much new in the realm of AP. But I am pretty sure that the pictures I have taken, with the equipment at my disposal, are as good as I am ever going to get them. It is in the processing that my skills are seriously underdeveloped. Oh, how I wish they'd invent a program for people like myself that has a big red button reading "click here to make it pretty". But where's the fun in that, right?

 

[carastro]

Half the work is in the processing.  You can actually spend more time processing than capturing.

Carole 

[ollypenrice]

27 minutes ago, George Gearless said:

Thanks to all for some great responses. I'll check out my archives and see what I want to try and lump together.

I am still very much new in the realm of AP. But I am pretty sure that the pictures I have taken, with the equipment at my disposal, are as good as I am ever going to get them. It is in the processing that my skills are seriously underdeveloped. Oh, how I wish they'd invent a program for people like myself that has a big red button reading "click here to make it pretty". But where's the fun in that, right?

 

 

25 minutes ago, carastro said:

Half the work is in the processing.  You can actually spend more time processing than capturing.

Carole 

In my view capture is a mechanical process with very limited scope for refinement. You can finesse with the sub length and the balance of time per filter, etc, but there comes a time when you have nowhere to go to improve.        Well,  OK you can also spend more on your mount and optics, of course!!!

This is never true of processing. The room for better processing knows no bounds. 

Olly

[carastro]

Very true Olly, I've been imaging for around 8 1/2 years now, and still learning new things in processing.  

Carole