Adding RGB stars to Dual Narrowband Data

I only did this once as I'm new to this but I was quite satisfied with the result.

I took 2 hours of RGB subs with 180s each (depending on your sky).
For the RGB stack, I did the following: Gradients removed (GraXpert), Background neutral., SPCC, BlurXT + NoiseXT (but just a little bit), GHS stretch, Saturation increased + minor adjustments (e.g. CA around stars removed).

Then it's important to StarAlign the RGB image with the DNB image before extracting the stars. And then just combine the starless DNB with the RGB stars using PixelMath or any kind of script like ScreenStars.

30 min. to 1 hour should be fine.  I do 30 sec. subs X 30 each RGB.  However I shoot at F3.6.  Higher F ratios may require longer.  At F3.6 the stars were blown out at 1 min subs.

Remember, stars will be high S/N and the noisy background will be removed/eliminated. 

Lynn K.

Just a hint: You will get much better results by adding the stars by to the DNB image by "Screen blending" … and not just adding (+) them.
Best of course in that case would be to generate the stars from the RGB by using an "unscreen" method, which is e.g. provided by StarXTerminator … do not know if Starnet++ provides this feature as well …
Alexander implicitly suggested this above by using the ScreenStars script …

Chris

That's pretty much what I do too. But make sure you align the dnb stack with the uv-ir cut stack. I use Deep Sky Stacker for this.
This is my recent attempt of combining dnb data with rgb: https://astrob.in/sqfmsi/B/

If you don't want to shoot the target with a uv-ir filter, you can just use the HOO stars from the dual-band data processed through SPCC choosing the appropriate filters. The colours come out pretty close; not perfect but if you blend them back in, you're probably going to bring them in with much less stretch than for the main event, so you get star reduction at the same time that makes them much less noticeable anyway. You might try this and see if you find it acceptable. Of course, if you do want to get RGB stars, then the advice above is fine. Shorter exposures in rgb should be better than longer to avoid clipping…

Jeff Horn:
If you just want to use the narrowband data,j Alberto Ibanez has a nice explanation and tutorial here....RGB Stars from Narroband Data – AIASTRO (wordpress.com)

It just used the H and O channel in a combination he describes as very close to RGB.
I've used my spin on this for a long time and I really appreciate him sharing this, I think it works quite well.
If your narrowband subs are really long and the stars are blown out, then this may not work for you, but just something to think about.

Thanks for that article. I have actually been looking for a detailed explanation like this. I have been lately generating stars using the Foraxx script in PI and they look decent, but never actually did a apples to apples comparison with an RGB aquisition to see how they stack up.

For my image,

I made the stacking, the alignment and the pollution removal with APP
After with Pixinsight, I make a RGB layer, I treat it in classical way: BXT + NXT + HT
In non-linear mode, I add the luminance mask after adjusting the color level 
I finish by adding the LRGB mask to the SHO mask without stars.

At the beginning, I made 60x60s by level

Now, I made 30x60s for RGB and 60x60s for L

the time has been fixed with the 3 sigma rule.

Just coming back to this with a couple more questions if I may. I've been collecting data where I can and now have the following data sets:

300 second subs with L-enhance filter
300 second subs with UV/IR (Astronomik L3) filter
20 second subs with UV/IR (Astronomik L3) filter for the stars.

I've created two separate stacks so far using WBPP in Pixinsight:

Stack 1: 300 sec L-enhance data
Stack 2: 20 sec subs to be used just for the stars. (I'll extract the stars from this data and delete the starless so I can re-screen these stars with the main data set at the end - My hope is by using short exposures for the stars, I will maintain the star colour)

For the main stack (I'll call it stack 3), should I just use the 300 second subs of broadband data or should I also include the 20 seconds subs that I used to create the stars and the 300 second subs with the L-enhance filer?

my understanding is that when combining Ha data with an RGB image, I separate the colour channels, delete the green and blue and just use the red. As I'm using a OSC camera, is this just a waste of data and should I therefore include this data in the main stack?

Does that all make sense?

Stack dual narrowband data, DBE, BlurX, etc, remove stars

 
I always add RGB stars to my narrowband images and using StarXTerminator too late in the process can leave some marks. Especially if by "etc" you mean that'll you'll perform some color calibration. The author states that StarXTerminator should be ran right after stacking for the cleanest removal. Try this:

• Stack dual narrowband data
• DBE
• DynamicPSF. Keep the average FWHM in note as this will be your manual PSF value for BlurXTerminator.
• StarXTerminator
• BlurXTerminator (non-stellar only with manual PSF)
• Etc...