This is my first attempt at narrowband imaging. I love the fact that the backgrounds are cleaner and the stars are smaller, but maintaining the broadband color of the stars is a challenge.
I've gone thru many incarnations of the Fish Head Nebula. As I learn new thing I implement what I've learned into my PixInsight processing.

11-10-21
Sixth try -
I figured out a way to get the small narrowband stars with color, this method is not my final solution but for now it will have to do.

I processed the Ha,Oiii and Sii into a fully processed RGB_image
StarNet - remove the stars from the RGB_image

Start with the three cropped Ha_C_ABE, Oiii_C_ABE, and Sii_C_ABE
EZ Soft stretch on all three
StarNet on all three to get 3 StarMask images - Ha_star_mask, Oiii_star_mask, Sii_star_mask

PixelMath - combine the 3 star_mask images into one Master_star_mask
    UnCheck 'Use a single RGB/K expression'
    'Color space:' - RGB
    Fish Head Nebule -
        RGB/K: Ha_star_mask * 0.8 + Sii_star_mask * 0.5
        G: Oiii_star_mask * 0.5 + Sii_star_mask* 0.5
        B: Oiii_star_mask * 1.2

HistogramTransformation
    Select Master_star_mask
    Set 'High range' to 1.05 - will reduce blown out colors, mainly red
    Apply to Master_star_mask

Convolution
     Soften the star in Master_star_mask

Could use Master_star_mask as a mask for the RGB_image - protect the background

PixelMath - combine the Master_star_mask with starless RGB_image
    Check 'Use a single RGB/K expression'
    'Color space:' - RGB
    RGB/K: Master_star_mask + RGB_image

This produced a final image with smaller stars with pretty good color, even though some of you may say that they have false colors. Maybe they do but for now I don't have a better system. If you do then please let me know !!


11-9-21
Fifth try -
This time I combined my R,G and B images, removed the stars and put them into my starless narrowband image. The stars aren't as well behaved as the narrowband stars but they do have some color - tradeoff !! My processing will have to improve before I'm satisfied.
I also backed off on the Denoise, so the image isn't quite as smooth and looks a bit more natural.  I backed off on the saturation also, I think I like the previous image better because the stars were smaller and had less junk around them. No one said it would be easy !!

11/8/21
Fourth try -
Combination of SHO and ForaxX palettes. This is my first narrowband attempt and it took a lot of time and effort to get the final result. The stars are white but my next revision will be with stars from an RGB image.  The hardest part seems to be the stars and their halos and color.
I tried real hard NOT to over-process it, but you can be the judge of that. Let me know what you think, both good and not-so-good.
Monty

11/2/21
Third try -
I added more data but I'm still low on Ha and Oiii. The colors are starting to come out but there is still too much noise. I still processed it with a mix of SHO and ForaxX palettes.

10/28/21
Second try - 
I created a SHO image and a ForaxX palette image and combined them with PixelMath. Still using the really bad data from cloudy nights - it looks like clouds for the foreseeable future.

The Fish Head Nebula -
This is my first attempt at narrowband imaging. The night had  a lot of high clouds so I was only able to get about 1.5 hours of horrible data. It may be a week before I have a clear night, in the meantime I will practice with what I have and watch videos on SHO, HOO etc.
First try - HOO
Red - Ha
Blue - Oiii
Green - Oiii