Stacking multiple nights in one?

[WolfieGlos]

I've read a few times that you should stack multiple nights separately in groups in DSS. Although I use SiriL and I don't have that option, I haven't noticed anything untoward using frames from multiple nights in 1 stack. Aside from edges and framing !

The only reasons I can think of for stacking nights separately are:

1) To get matching temperatures if using dark frames

2) If sessions are months apart, and a new set of flats/biases are to be used.

3) If using different ISO / exposure times, i.e. due to the Moon phase or a HDR rendition of M42.

So if I'm not using darks, but imaging on 3 consecutive nights with the same ISO and exposure times, is there any reason not to just stack them all in 1 group?

Chris

[Fegato]

I will often integrate multiple night subs. In terms of your reasons for not doing so:

1. I have a cooled camera, so I always shoot at the same temperature (not sure I would worry too much if the temperature wasn't exactly the same)

2. I have certainly had to use different flats for groups of subs within the same stack. So I will calibrate with different flats, but then integrate the calibrated subs into one master. I use Pixinsight, so can't comment on how easy or not that is to do with other packages.

3. I have certainly integrated different exposure lengths. Pixinsight has a good quality algorithm, so will weight the subs accordingly. Normally the longer exposures will get a higher weighting (other things being equal)

[Elp]

I tend to stack as quickly as possible within reason prior to the next session, that way you can histogram stretch and see if there were any issues. Then I stack the stacks.

[WolfieGlos]

On 28/12/2022 at 20:18, Elp said:

I tend to stack as quickly as possible within reason prior to the next session, that way you can histogram stretch and see if there were any issues. Then I stack the stacks.

I do that as well, but also to check noise levels and have a first quick pass at editing and comparing the additional data that gets added. But I’d never thought of stacking the stacks! I might give that a go next time.

Can you stack stacks of different exposures/ISO? With the Moon phase now coming I’m going to try NB for the first time and am expecting variations of exposure times.

[Elp]

A lot of stacking software is written so when trying to stack data of different types it won't allow it, but I would have thought the above is totally possible otherwise what would be the point of being able to change camera settings. I've never felt the need to though, once I start a project the camera settings stay the same until complete. I've only had errors when pixel resolutions don't match, ie trying to stack a cropped image with another raw stack, it doesnt stop you from doing it manually in GIMP or PS.

[900SL]

On this subject, I'm about to start a second evening of OSC on the Veil. I've already stacked 40 frames from my first session a week ago. These were calibrated with DF&B frames in SiriL

Am I correct in assuming that I should stack the second evenings frames using new lights and the previous darks and flats. 

I would then have a 'master' stacked and calibrated light frame  from each nights work, which are then stacked to give a final master frame (ie just two lights stacked, no additional calibration frames)?

 

 

[Fegato]

1 minute ago, 900SL said:

On this subject, I'm about to start a second evening of OSC on the Veil. I've already stacked 40 frames from my first session a week ago. These were calibrated with DF&B frames in SiriL

Am I correct in assuming that I should stack the second evenings frames using new lights and the previous darks and flats. 

I would then have a 'master' stacked and calibrated light frame  from each nights work, which are then stacked to give a final master frame (ie just two lights stacked, no additional calibration frames)?

 

 

I use Pixinsight. I've always thought that the route to the best result is to integrate (stack) all the calibrated frames in one go. This means that they are all compared to each other for quality, the best overall frame is used as a reference, pixel rejection routines work across all frames etc etc.  Not sure about SNR improvement, I'll let someone else comment on that, but there seem to be enough reasons to do it this way anyway.

In terms of calibration - I will calibrate groups of files captured on different nights separately if I have different flat frames for the sessions (sometimes I do, sometimes not - depends what has changed in the meantime). My darks and bias are "permanent" masters, so these don't change.

 

[900SL]

1 hour ago, Fegato said:

I use Pixinsight. I've always thought that the route to the best result is to integrate (stack) all the calibrated frames in one go. This means that they are all compared to each other for quality, the best overall frame is used as a reference, pixel rejection routines work across all frames etc etc.  Not sure about SNR improvement, I'll let someone else comment on that, but there seem to be enough reasons to do it this way anyway.

In terms of calibration - I will calibrate groups of files captured on different nights separately if I have different flat frames for the sessions (sometimes I do, sometimes not - depends what has changed in the meantime). My darks and bias are "permanent" masters, so these don't change.

 

Thanks Fegato. I use Siril so it looks like I need to dip a bit deeper into separating calibration and stacking using scripts.

At the moment I usually run a script that automatically does everything in one go, with DBF calibration and background extraction. It's all 'under the hood' and out pops a calibrated stacked light frame for post processing

You can however use scripts that will halt after registration, calibration etc at various points. I need to get googling as I think I understand the principle you are referring to. Siril also allows frame by frame FWHM analysis in graphical form of a series of calibrated light frames so you can reject poor frames before stacking  

So the best way seems to be:

Create a series of calibrated light frames for each evenings run using the flat frames from that session and the master dark and bias frames

Check FWHM graph and reject poor frames

Stack the total, no additional calibration   

 

    

[Fegato]

Yes that sounds good.