Adding RGB Stars to SHO image

[Startrek]

So I’ve captured Ha , Oiii and Sii data on an emission nebula to process a typical narrowband data set
I’ve also captured 10 x 15sec ( R , G and B data sets for stars )
How do I include the RGB stars to my SHO image
All data sets will be stacked and aligned correctly in ASTAP

Simplest step by step procedure please …….

Thanks in advance
Martin

[Stefan B]

Hello Martin,

I can't provide a step by step process at the moment but maybe this thread helps?

viewtopic.php?f=10&t=2375

Regards
Stefan

[Carles]

Process both stacks

Then open RGB dataset processed --> make star mask>>> to Layer Module keeping mask >>>open Narrowband dataset you can leave at Blend, then play with mask and Kernel (lower right, sorry not in fron of computer now ) to blend nicely.

Kust posted an example of this process here
viewtopic.php?f=11&t=2891&start=10

Regards
Carles

[Startrek]

Thanks but still getting my head around it

I found this post reply from Ivo ( attached ) back in 2021

Martin

[Mike in Rancho]

I'll just throw in that 10 by 15 seconds might be an itsy bitsy tiny wee bit short. Of course I could be wrong.

But you'll want clean enough SNR for good star colors, and not just the cores but the fainter diffraction regions surrounding them. Especially big ones. Remember you'll want the stretched RGB star layer to be able to fully overlay the finished NB stars, and not end up too small. Otherwise you'll end up with RGB cores surrounded by NB diffraction.

I believe Ron just ran into this in a recent posting, and remedied it in part by bloating up the RGB star layer. I think he had more than just a few minutes total RGB integration though.

At my B8 and f/3.8, for this I would probably do at least an hour - 20m per filter. And I'd keep the blown pixels to a minimum, so quite possible pretty short subs.

[Startrek]

Thanks Mike
I’ll make sure I capture at least 20min each RGB
A few questions in regard to Layer module use …..
What file type should I save the completed NB image and the completed RGB image prior loading into Layer module ( is jpeg or tiff ok ?? )
How do I load the completed NB image into Background Bg window and load completed RGB image into Foreground Fg window in Layer module ?
Ivo advises to create a Mask ( Star mask obviously) but which image ? NB image or RGB image
Then he says set Layer Mode to Color of Foreground Fg which should make the NB image inherit the stars color of the Background Bg ?
He then says use Mask Fuzz and Kernel radius to adjust to taste
Apologies I’m a bit lost with Layering, never used it ( Guys User notes don’t give an explanation of how to deal with NB and RGB stars )

Thanks in advance

Martin

[Mike in Rancho]

Don't worry Martin, Layer won't bite.

Like much of this, just play around a bit. No need to take notes. Push buttons and turn knobs until something breaks.

Ivo's steps are pretty complete I think, so you can just follow them along. Though I would process the NB to taste in the normal fashion rather than trying to match up with visual.

Now, what does everything mean and how does it work? Layer is a triple window module: Background, Foreground, and Combined (your result), in that order. I think. Actually I always forget, even though which is which flashes on the screen upon module entry. But no worries, there are all sorts of buttons on the top that do all sorts of this and that.

When you enter Layer, the current image goes into one slot (foreground?). You can open a matching dimension (and preferably registered) file, which goes into the other slot. You can swap which is Fg and which is Bg. You can go into Mask. Or you can create the mask before entering Layer. I believe the mask, however, is always tied to the original loaded file that you entered with. Or something like that.

The mask, custom (here, stars) or full, will be what gets blended, based on the type of blending chosen, into the combined result.

Fuzz helps blend the masked edges, and kernel radius imparts a bit of a softening blur. I have no idea why it's called kernel radius rather than blur. Probably some math thing. As an aside, Layer with a kernel radius blur is also very handy for singularities and blown pixels, say you saturated all your star cores to pure white, or M31's core is putting your eyes out. That sort of thing. A quick and easy targeted artifact repair.

Try it both ways, why not, but I would likely first give a shot to making the mask on the target (narrowband) image, sufficiently covering those stars and their surrounds (diffraction/glow) where you would like to get the color from the RGB file. As before, you will want those stretched sizes to be comparable in order for that to work.

Experiment.

There's a logic to combining RGB stars to NB, but there's some skill and artistry involved in pulling it off correctly. In a way it's NB Accenting but in reverse. You are RGB Accenting your SHO or HOO.

Fair warning, I've never been very good at it, especially against nebulous backgrounds. So I tend to just embrace my narrowband stars, particularly in SHO. HOO seems to need the help more.

Hope that's moderately useful. Ron and others who have done it more will probably have better and more concrete tips.

[Startrek]

Don't worry Martin, Layer won't bite.

Like much of this, just play around a bit. No need to take notes. Push buttons and turn knobs until something breaks.

Ivo's steps are pretty complete I think, so you can just follow them along. Though I would process the NB to taste in the normal fashion rather than trying to match up with visual.

Now, what does everything mean and how does it work? Layer is a triple window module: Background, Foreground, and Combined (your result), in that order. I think. Actually I always forget, even though which is which flashes on the screen upon module entry. But no worries, there are all sorts of buttons on the top that do all sorts of this and that.

When you enter Layer, the current image goes into one slot (foreground?). You can open a matching dimension (and preferably registered) file, which goes into the other slot. You can swap which is Fg and which is Bg. You can go into Mask. Or you can create the mask before entering Layer. I believe the mask, however, is always tied to the original loaded file that you entered with. Or something like that.

The mask, custom (here, stars) or full, will be what gets blended, based on the type of blending chosen, into the combined result.

Fuzz helps blend the masked edges, and kernel radius imparts a bit of a softening blur. I have no idea why it's called kernel radius rather than blur. Probably some math thing. As an aside, Layer with a kernel radius blur is also very handy for singularities and blown pixels, say you saturated all your star cores to pure white, or M31's core is putting your eyes out. That sort of thing. A quick and easy targeted artifact repair.

Try it both ways, why not, but I would likely first give a shot to making the mask on the target (narrowband) image, sufficiently covering those stars and their surrounds (diffraction/glow) where you would like to get the color from the RGB file. As before, you will want those stretched sizes to be comparable in order for that to work.

Experiment.

There's a logic to combining RGB stars to NB, but there's some skill and artistry involved in pulling it off correctly. In a way it's NB Accenting but in reverse. You are RGB Accenting your SHO or HOO.

Fair warning, I've never been very good at it, especially against nebulous backgrounds. So I tend to just embrace my narrowband stars, particularly in SHO. HOO seems to need the help more.

Hope that's moderately useful. Ron and others who have done it more will probably have better and more concrete tips.

Thanks Mike,
Yeh I guess the only way to learn is jump in the deep end with ST , it’s a pleasure to use. I’m one of only a small handful of folk on our local forum who use ST. Most opt for Pi and PS which fine if you want spend half your life at the workstation.
Next time I remember to capture some RGB stars , I’ll give Layer a whirl and see how we go
Clear Skies
Martin

[fmeireso]

I am experimenting with this also...i still don't know what the best method is. I noticed that making a starmask to separate the stars produces not fine stars imho. Unless i overlook something.

I tried yesterday a new approach , made a starless image of the nebulae duoband stack, substracted thatone from a OSC stack leaving me with the stars. They looked better and then layered actually the starless nebulae back in the starfield.

Stars looked better but the coloring in some places was a bit weird. Below the result, just a sample, you still see the aligment artifact

NewImage2 (Medium).jpg (331.25 KiB) Viewed 29224 times

Anyhow many things to test out to find out what gives the best results...