Can I use layers to combine different stacks of the same target?

[patraip1]

Hello guys.
I have gathered quite a lot of data of M33 taken on 4 different nights.
As I was experimenting with ISO and exposure lenghth they are all different.
First set is 426 subs 10sec iso200
Second 456 subs 13 sec iso 400
Third 156 subs 15 sec iso 400
Fourth 245 subs 15 sec iso 200
I have flats dark flats and bias for each set (I dont use darks) I have an unmodded NIkon D5500 at prime focus on a 130p skywatcher, on an alt az AZGTI.
I stack in DSS with ST recommended settings
I have processed each set of data in ST and have 4 different tiff files.
Is there a way to combine them in ST?
How will I manage to match size and alignment?
I tried putting all the data in DSS to stack together in groups and it will take 88hours....

Thanks for any tips.
Tryfon

[almcl]

Short answer, is it is possible but you will need to stack in DSS using a reference frame from just one of the data sets for each set. You then need to bin and crop each set in ST by exactly the same amount; one pixel difference in size will prevent the layer module accepting the data, and for 4 datasets this can be extremely fiddly.

Question though is what do you hope to achieve? M33 is a largish rather faint target and short exposures at low ISO may not capture much detail. Generally mixing ISOs and exposures is not recommended, better to find the best exposure combination and stick to it.

[patraip1]

thanks for the answer!
I was hopping to get some more detail if at all possible.
Do you mean that I should use one frame from one stack in a different group in DSS, use that as a reference frame and stack the other subs in a different group,if I was to use just 2 sets of data?
This is the result of 400 frames 13sec iso 400

[Mike in Rancho]

It's alt-az. This adds an extra wrinkle.

It it were me, so that more of all that time acquiring data was (easily) usable, I would do short tests at the beginning, then settle on keeping the same ISO and exposure for all sessions.

Your hard limiting factor is star trailing/rotation, so you need to take some high iso test shots in order to find the longest exposure you can take but still have round stars. Might be 10 seconds, 15, ? Once you know how long you can go, then keep that and move on to ISO by checking your histogram and balancing how many oversaturated stars, if any, you are willing to accept, while keeping your peaks off the left edge. 200 and 400 are the "best" iso for the D5300, and the D5500 is likely similar. So either of those.

Salvaging the current data can also be tricky or difficult. DSS has an "entropy" stacking setting that is designed to mix different exposure times and iso/gain. I don't know how well it works or how ST-permitted it is, but is probably the best option. Mixed exposures in normal stacking, even in DSS groups, is not likely to give the results you want.

In order to use layer blending of separate stacks, as noted the final files need to be not only registered but match cropped to the same x and y resolutions. The better route per ST is to use the final output of one stack as the reference for the next stacks (reference only, not checked for stacking). This can also be trouble, as a final stack is often too dark to be used as such (32 bit and dark or bias subtracted) without some extra work by temporarily upping the brightness in the FITS settings and then putting it back. Also, intersection mode becomes a hindrance.

A non-recommended route, which I've used many times when receiving other's data, such as unregistered LRGB filter stacks, is to use DSS to register them to a reference. Match cropping must be done first (strategically if possible) as DSS requires matching x and y sizes. Standard mode must be used, not intersection. A reference is chosen and registration is run. Then you "stack" each one all by itself. Of course, it's only one file so no stacking or rejection algorithm actually takes place, the file it just re-written but shifted so as to be registered. This is done for each file. Finally, the overlaps (you can think of them as stacking artifacts) need to be cropped. There can be a good bit of back and forth between DSS and ST, or Gimp, to accomplish this. You will then have your registered linear stacks.

Then of course, after each file is processed in ST, Layer only blends two images at a time, so that will take a few steps. Something like Gimp or PS can do multiple layers.

Soooo...unless you are really really beholden to this data, I'm not sure you want to jump through all those hoops? Especially since they might not blend all that well anyway. Racking up more integration time of the same exposure and iso seems the better route to more detail in your M33. And if you know one of your combos is already good (maybe the 13s iso400, or one of the others), you've at least got a head start to add to it.

[hixx]

Hi, the biggest bang for the buck you would probably get will be by using darks. Even if Ampglow is not likely to occur given your exposure times, darks will also remove the Bias and mitigate the shot noise pattern. the only way to get around darks would be heavy dithering during capturing. there is quite some visible noise in the sky background in your picture.
clear skies,
Jochen

[patraip1]

Thanks for the rplies guys!!
I will try to follow Mike 's suggestion. I think I am settling at iso 400 and 13 sec, in order not to have star trails. Darks have an opposite effect in my stacks and increaee noise (cannot tell why), a fact that I have verified with all my data (I still have darks for all my sessions), so I am not using them anymore.
I was just wondering if there was a plausible way of salvaging older data in different iso and exposures of the same target.
I will play around with DSS to see if anything is possible and report back...

[patraip1]

OK!
I had a stack already running from yesterday in DSS .... I placed one light in main group and did not check it and then created 4 groups with my different data of m33. I placed the corresponding flats and biases in each group. I chose 1 light from one group with a good score and framing and set it as a reference frame and then I used an average stacking method. 36 hours later DSS provided a fts file . Total numner of stacked images 1182...
Here is the fts file of all the data for anyone who would like to take a look. (Bare in mind that with my older iso200 data I had a flats issue and experienced vingetting )
https://drive.google.com/file/d/14gV6bH ... sp=sharing

Here is a quick process in ST. I think there is more data than the single stacks but that might be just wishfull thinking

M33ALL.jpeg (103.24 KiB) Viewed 5924 times

And here is my second try

M33ALL_1.jpeg (214.71 KiB) Viewed 5920 times

and here is a 3rd attempt

M33ALL_2-01.jpeg (325.97 KiB) Viewed 5883 times

I am basically playing around with wipe and then autodev after binning and cropping, which produces massive differences in the images. I then use almost all modules in order with pretty much default settings: contrast, hdr (reveal core), sharpen, color, entropy, superstructure, denoise and repair. I am quite happy with the last one and I think it was worth the 46 hours of stacking everything together....
If anyone has any suggestions feel free to dive in at the data....


PS:I also have the stack of one of my groups (400iso 13 sec , 426subs) for reference
https://drive.google.com/file/d/1IRtT01 ... sp=sharing

[happy-kat]

With several nights and mixed data I would be tempted still to throw them into DSS.
Each night with it's own lights into a separate file group each (not the default main group but added groups).
Decide on the one reference frame for all and mark as reference and tick (file group1).
Add this frame to the other groups but do not tick including the default group.
Using this info (taken from DSS web help) decide where to put your calibration frame
Dark, Flat and Offset/Bias Frames from the Main Group may be associated to Light Frames of any group.
Dark, Flat and Offset/Bias Frames from others groups may be associated only with Light Frames of the same group.
To read further help for when to put your calibration files here's the link
http://deepskystacker.free.fr/english/technical.htm