Multi session image stacking

[Bobby1970]

I am trying to add two sessions of data together to create one final image. 

 

What I have done is:

 

Lights+darks from 1st session added to "main group" in DSS

 

Lights + darks from 2nd session added to "group 1" in DSS

 

 

Check all images in both groups and let DSS do it's stuff???

 

Does this sound about right? 

 

The final image is no better than just a single session to my eyes. I was expecting more.

 

The temp was a couple of degrees less on my second session, would this be a problem even if I use darks for each session which were captured at the same time as the lights for each session??

Edited March 2, 2021 by Bobby1970

[GazAstro]

Whenever I've done this I just pile them all in one group ... I don't think you'll see any difference until you stretch the saved Tiff

[Padraic M]

I've just done this a few times, and your process seems sound but a few questions. 

As you're talking about temperature, are you using an uncooled camera? If the lights are at very different temperatures then yes you should have separate groups with matching darks. Similarly are your lights at different exposure lengths, gain/ISO, etc.? Are you using flats/dark flats? Again, different groups for each set of lights/darks/flats. iirc bias frames are common for the camera so don't go in different groups but I don't use bias frames. How much exposure have you got in each group - is there enough to make a real difference in image quality?

As Gaz says, you can just drop them all into one group with one set of darks, it should work well enough, but might be noisier if there's a big temperature difference. This would work less well if you're using flat frames, and have changed focus or camera orientation between sessions.

[Bobby1970]

1 hour ago, Padraic M said:

I've just done this a few times, and your process seems sound but a few questions. 

As you're talking about temperature, are you using an uncooled camera? If the lights are at very different temperatures then yes you should have separate groups with matching darks. Similarly are your lights at different exposure lengths, gain/ISO, etc.? Are you using flats/dark flats? Again, different groups for each set of lights/darks/flats. iirc bias frames are common for the camera so don't go in different groups but I don't use bias frames. How much exposure have you got in each group - is there enough to make a real difference in image quality?

As Gaz says, you can just drop them all into one group with one set of darks, it should work well enough, but might be noisier if there's a big temperature difference. This would work less well if you're using flat frames, and have changed focus or camera orientation between sessions.

My camera is Uncooled.

 

I did two different nights, same target M45.

 

Night 1 : 40x 120s lights @ 130 gain, 30 darks.

Night 2: 38x 120s lights @130 gains, 30 darks.

 

Not bothered with any flats yet. 

 

I watched a youtube video which pointed something out about having files in "main group" and "group 1" which seemed to be what my issue may have been.

 

The solution is to put a different light frame in "main group" but do not check it. Then use "group1" for night 1 lights+darks and "group 2" for night 2 lights and darks. This seemed a bit better, maybe i am expecting too much.

 

 

This is the 2 session image:

 

 

This is the single session one:

 

maybe its my rubbish processing lol

 

 

[Padraic M]

Well processing certainly is a quagmire that many of us take years to wade through. I can see differences between the two images, and overall I'm surprised that I prefer the single session version, given the additional data in the other version. The best way to work your way to better processing is to try lots of different approaches and see which works best.

 

First off, I have to say that I've never gotten on with Startools. Others can get great results, but it definitely needs top-quality input data. I'd recommend that you download Gimp, it's free and it's excellent. All you need to do to start with, post-stacking, is to load your stacked TIFF into Gimp and do some simple histogram stretching (levels & curves) and colour balancing.

 

Wrt DSS and multi-session, what you say makes sense when I read the manual again - if you put darks, flats or darkflats in the Main Group, they "may be associated to Light Frames of any group". This is a little ambiguous, but I read this as if you want to apply separate darks for two sessions, you need to put them in Group 1 and Group 2. You will need to add a 'dummy' image to the Main Group to open up a Group 1 tab. 

 

I'd just try putting all of your lights into Main Group with one set of darks from either night. Start there, it's simple and should give perfectly good results. Once you're happy you've got as much out of that as you can, you can play with multiple groups and matching Darks.

 

In the single-session image, the background sky colour is a nice almost-black. It's better than in the two-session, which has a strong colour cast. That could be down to your processing, but I wonder did any of your optical elements fog up on the second night out? There seems to be a lot of fogging in the central area that doesn't correspond to nebulosity. There's also an apparent foggy star trail heading to the top-left corner on some of the larger stars that I don't see in the single-session image. You might need to review each of your lights, and decide if they are of a good enough quality to use.

 

On the stars, check the individual lights and see if the stars are sharp, or are they saturated, blurred or bloated. This will determine if they're over-exposed (unlikely at 120s?) at capture or whether you're over-stretching them in processing. Again, just do some simple stretching in Gimp - less is more, until you know what the data can take. Most people will generate a star mask to protect the stars while stretching. Star haloes can also come from condensation on the optics or sensor, or from interior reflections from filters or correctors. 

 

What I really like is all of that nebulosity that you've captured in the blue channel. The outer layers of red/brown/magenta are unfortunately distracting from it. The best M45 images have a near-black sky, crisp stars, and just that blue reflection nebulosity around the main stars.

 

hth!

[Bobby1970]

12 hours ago, Padraic M said:

Well processing certainly is a quagmire that many of us take years to wade through. I can see differences between the two images, and overall I'm surprised that I prefer the single session version, given the additional data in the other version. The best way to work your way to better processing is to try lots of different approaches and see which works best.

 

First off, I have to say that I've never gotten on with Startools. Others can get great results, but it definitely needs top-quality input data. I'd recommend that you download Gimp, it's free and it's excellent. All you need to do to start with, post-stacking, is to load your stacked TIFF into Gimp and do some simple histogram stretching (levels & curves) and colour balancing.

 

Wrt DSS and multi-session, what you say makes sense when I read the manual again - if you put darks, flats or darkflats in the Main Group, they "may be associated to Light Frames of any group". This is a little ambiguous, but I read this as if you want to apply separate darks for two sessions, you need to put them in Group 1 and Group 2. You will need to add a 'dummy' image to the Main Group to open up a Group 1 tab. 

 

I'd just try putting all of your lights into Main Group with one set of darks from either night. Start there, it's simple and should give perfectly good results. Once you're happy you've got as much out of that as you can, you can play with multiple groups and matching Darks.

 

In the single-session image, the background sky colour is a nice almost-black. It's better than in the two-session, which has a strong colour cast. That could be down to your processing, but I wonder did any of your optical elements fog up on the second night out? There seems to be a lot of fogging in the central area that doesn't correspond to nebulosity. There's also an apparent foggy star trail heading to the top-left corner on some of the larger stars that I don't see in the single-session image. You might need to review each of your lights, and decide if they are of a good enough quality to use.

 

On the stars, check the individual lights and see if the stars are sharp, or are they saturated, blurred or bloated. This will determine if they're over-exposed (unlikely at 120s?) at capture or whether you're over-stretching them in processing. Again, just do some simple stretching in Gimp - less is more, until you know what the data can take. Most people will generate a star mask to protect the stars while stretching. Star haloes can also come from condensation on the optics or sensor, or from interior reflections from filters or correctors. 

 

What I really like is all of that nebulosity that you've captured in the blue channel. The outer layers of red/brown/magenta are unfortunately distracting from it. The best M45 images have a near-black sky, crisp stars, and just that blue reflection nebulosity around the main stars.

 

hth!

Thanks so much for the response. 

 

I have tried to persevere with Startools for years and tbh I have never really been happy with the results. It feels more like luck sometimes rather than what I input making much difference. 

 

I think I am going to try using GImp. Got to be worth a go. 

 

Regarding stacking and possibly just using one of the set of darks from either session, as the temperature was different by a couple of degrees 10 v 12 Deg C between the sessions. Which would be the best darks to use? The ones taken at a higher temperature or the ones at a lower temp?

 

Many thanks

 

 

[TerryMcK]

Another tool to consider is the very Photoshopesque (is that a word?) Affinity Photo from Serif. It is really cheap, has lots of the functionality that PS has and can also run some PS plugins. I believe that some of the astrophoto tools have been ported over to it to such as  Astronomy Tools from ProDigital Software although I've not investigated more.

Certainly Topaz Denoise works from it.

[Padraic M]

59 minutes ago, Bobby1970 said:

10 v 12 Deg C

I'm not a big expert on this but I'm not sure if a difference of 2 degrees would be material. Pick one and try it - it will help you work on other issues and iron them out. If you get a noisy image but everything else is ok, then you can move on to experiment with the different darks.

[Bobby1970]

So i took the same two-session stack and had a go in GIMP, i cant for the life of me figure out star masks and layer masks yet. lol

 

This was about as good as i could do tbh. 

 

 

Is it better than my efforts in Star tools, i don't know, seems less details showing here compared to my single session effort processed with startools. 

 

[Bobby1970]

Then, i had a go at re-stacking.

This time i dropped all of the lights, and one set of darks into the same group. I had a go with Startools, again. And finished in GIMP 

 

 

I think this is probably the best of a bad bunch tbh. As they say, you can't polish a turd, but you can roll it in glitter lol.

 

 

[Carastro]

Presume you are using DSS.  I haven't used it for years, but to be honest, I used to stack each session separately and then stack the stacks.  I found it worked better anyway.

 

It will also have the benefit of you having the individual stacks so you can process each and this should show you the benefit of having the multi stack when you compare the results.

 

One thing that will improve your images no end is the taking of flats which you say you haven't bothered with so far.  They will make a lot of difference to your final result, so I would advise learning to do these asap.   N.B. if you have moved the camera within the scope or removed it between sessions, you will need separate flats for each session or the dust won't line up.  

 

I found not having flats made it really difficult to process an image with the vignetting problems.

 

HTH

 

Carole 

 

 

[Bobby1970]

1 hour ago, Carastro said:

Presume you are using DSS.  I haven't used it for years, but to be honest, I used to stack each session separately and then stack the stacks.  I found it worked better anyway.

 

It will also have the benefit of you having the individual stacks so you can process each and this should show you the benefit of having the multi stack when you compare the results.

 

One thing that will improve your images no end is the taking of flats which you say you haven't bothered with so far.  They will make a lot of difference to your final result, so I would advise learning to do these asap.   N.B. if you have moved the camera within the scope or removed it between sessions, you will need separate flats for each session or the dust won't line up.  

 

I found not having flats made it really difficult to process an image with the vignetting problems.

 

HTH

 

Carole 

 

 

Thanks Carole. I appreciate the input. 

How exactly do I take flats with my OSC ZWO 178mc? 

 

I understand and have done it when I've used a DSLR. But not really with the OSC

[Padraic M]

Of the two new images, I have to say I prefer the first one! Startools is doing its usual trick of smudging the background. The other image is cleaner, and while I know what you mean about detail, the nebulosity is showing, and the stars are clearer. There is some slight star trailing, and some of the bigger stars are either over-exposed or over-stretched. Looking at the individual lights will tell you which. You might also want to check if your focus is as good as it can be.

 

It's a very good start, and a lot better than my first attempt at M45. There's a lot good in there, and analysing a bit more will tell you if you can improve the final outcome through more careful processing, or if there are lessons to learn for your next imaging session.

[Bobby1970]

I've sort of been at this "level" of processing for ever. lol. I just can't seem to get any better. 

 

I noticed some slight trailing of stars too, this is deffo from the second session. The guiding threw a wobbler at one point and i lost a couple of subs, could have been a cloud passing over maybe? 

 

I was reasonably happy (for me) at the first, single session only effort. When using Startools, even when using a mask, it seems to make no difference to the stars becoming bloated.

 

This is a single light frame from my first session (saved as a JPG for here), some of the stars look a bit odd shaped in places but my polar alignment was good. Are the stars overexposed? Perhaps i need a lower gain setting?

 

 

 

[AstronomyUkraine]

24 minutes ago, Bobby1970 said:

I've sort of been at this "level" of processing for ever. lol. I just can't seem to get any better. 

 

I noticed some slight trailing of stars too, this is deffo from the second session. The guiding threw a wobbler at one point and i lost a couple of subs, could have been a cloud passing over maybe? 

 

I was reasonably happy (for me) at the first, single session only effort. When using Startools, even when using a mask, it seems to make no difference to the stars becoming bloated.

 

This is a single light frame from my first session (saved as a JPG for here), some of the stars look a bit odd shaped in places but my polar alignment was good. Are the stars overexposed? Perhaps i need a lower gain setting?

The histogram on your processing software will tell you if the image is overexposed. If the curve is all the way over to the right, then you are clipping white pixels. What scope are you using to capture your images?

[Bobby1970]

1 minute ago, AstronomyUkraine said:

The histogram on your processing software will tell you if the image is overexposed. If the curve is all the way over to the right, then you are clipping white pixels. What scope are you using to capture your images?

The histogram curve is certainly not over to the right, its over the left as expected i believe.

Scope is an Altair Astro 72mm Semi Apo thing. I use my ZWO 178MC with it.

 

 

[AstronomyUkraine]

19 minutes ago, Bobby1970 said:

The histogram curve is certainly not over to the right, its over the left as expected i believe.

Scope is an Altair Astro 72mm Semi Apo thing. I use my ZWO 178MC with it.

 

 

The stars are clipped in the image. If you look at the histogram, you will see that all channels are clipped. You will notice the small bump on the extreme right of the histogram, that is clipping. You can try lowering the gain, or reducing exposure time. If the stars are clipped coming off the camera, not much you can do with them.

 

 

Edited March 5, 2021 by AstronomyUkraine

[Bobby1970]

3 minutes ago, AstronomyUkraine said:

The stars are clipped in the image. If you look at the histogram, you will see that all channels are clipped. You will notice the small bump on the extreme right of the histogram, that is clipping. You can try lowering the gain, or reducing exposure time. If the stars are clipped coming off the camera, not much you can do with them.

 

 

Wow, ok, many thanks. I was only really looking at the huge peak over on the left. I never realised that the small bit on the right would have made so much difference. Will certainly try a lower gain/shorter exposures.

 

 

[AstronomyUkraine]

10 minutes ago, Bobby1970 said:

Wow, ok, many thanks. I was only really looking at the huge peak over on the left. I never realised that the small bit on the right would have made so much difference. Will certainly try a lower gain/shorter exposures.

 

 

These are the default driver settings for your camera. I suggest starting at the highest dynamic range, or unity gain, see which works best. Keep images to 1 minute to begin with, see what the results look like.

 

Highest Dynamic Range ==>   Gain:  0    Offset:  25

Unity Gain ==>    Gain: 180   Offset: 25
Lowest Read Noise ==>   Gain: 270   Offset:  340

[Bobby1970]

26 minutes ago, AstronomyUkraine said:

These are the default driver settings for your camera. I suggest starting at the highest dynamic range, or unity gain, see which works best. Keep images to 1 minute to begin with, see what the results look like.

 

Highest Dynamic Range ==>   Gain:  0    Offset:  25

Unity Gain ==>    Gain: 180   Offset: 25
Lowest Read Noise ==>   Gain: 270   Offset:  340

Thanks again for the advice it is super appreciated, i have never been able to get to grips with the processing and/or settings of the camera tbh. 

 

i am not even sure what "unity Gain" means, but i think i have read somewhere that my camera doesn't actually have one.

 

I think i am certainly going to try lower gain settings, i have generally used between mid and high gains so its got to be part of the problem i feel.

 

 

I did have another play with Gimp, mainly just using levels and curves to try and understand how they work. Managed to get this out of my single session stack:

 

 

 

I know there are still plenty of issues with it, some of which are down to my camera settings etc, however i am quite happy with the balance of nebulosity and darkness of the sky here. This is the sort of look i am after.

 

Thanks to all of you again. I am sure i will be back with more questions following my next session lol. 

[AstronomyUkraine]

16 minutes ago, Bobby1970 said:

Thanks again for the advice it is super appreciated, i have never been able to get to grips with the processing and/or settings of the camera tbh. 

 

i am not even sure what "unity Gain" means, but i think i have read somewhere that my camera doesn't actually have one.

 

I think i am certainly going to try lower gain settings, i have generally used between mid and high gains so its got to be part of the problem i feel.

 

 

I did have another play with Gimp, mainly just using levels and curves to try and understand how they work. Managed to get this out of my single session stack:

 

 

 

I know there are still plenty of issues with it, some of which are down to my camera settings etc, however i am quite happy with the balance of nebulosity and darkness of the sky here. This is the sort of look i am after.

 

Thanks to all of you again. I am sure i will be back with more questions following my next session lol. 

Search for Nebula Photos on Youtube, his name is Nico Carver, he has many videos showing processing astro images with Gimp.

[Carastro]

7 hours ago, Bobby1970 said:

How exactly do I take flats with my OSC ZWO 178mc? 

The principle is the same for all cameras, but often easier with a DSLr as sometimes using AV and ISO100 will take it automatically for you.

 

In your capture software, is there anyway of reading the ADU?  If so this should be around 22,000 (1/3 well depth).  If not the histogram needs to be 1/3 across from the left.   If the reading is too bright you need to eitehr shorten the exposure, or dim the light and vice versa. 

 

If you have taken them before with a DSLR then the procedure is the same needing a diffuse light source.  Either the dim daytime sky or daytime sky with a T shirt/pillowcase or typing paper over the aperture to dim it down.  You will just have to experiment as to how bright this needs to be until you get the 22,000 ADU/1/3 well depth.  

 

N.B. You should try doing it with a mono camera and filters, you need different flats for each filter all needing different amounts of exposure to get the right ADU.

 

Carole