This is a crop from one of the panels of the SHO Cygnus Dust mosaic I did recently, reprocessed with Ha stars. I think the colors and contrast came out a bit better here. I do like the Ha stars more as well, but do hope to work RGB stars into my narrowband after learning how.

One thing I did differently in processing is removing the faintest of stars, as the RASA 8 requires larger 12nm narrowband filters and they let in a lot of smaller stars. Unfortunately, when viewing on the web these smaller stars are barely visible and instead only work to wash-out the nebula. However, using Starnett++ and PI's Binarize I was able to target only the faintest of stars for removal, with this technique:

In PI,

1. Create a duplicate of the image, name it "starless", and run Starnet++ on it. Set it aside.

2. Create another duplicate of the original, naming it "starmask".

3. Run Starnet++ with "create mask" checked, and extract lum or convert to grayscale if it's in color.

4. Next, Binarize the star mask so as to exclude the faintest of stars (higher values will exclude more brighter and brighter stars), I did not want to be too aggressive here.

5. Now run MorphologicalTransformation set to "Dilation", size'd appropriately to cover the entire star, 13 worked for me. Make sure to click the "circle" icon for the shape.

6. This should have expanded the stars for masking, now run Convolution to smooth the mask (3.0 StdDev worked for me).

7. Apply the "starmask" mask to the original image, and invert the mask. The medium to large stars should now be protected.

8. Open PixelMath and put in "starless", apply to the original image while it's masked.

9. All faint stars not protected should be wiped out, but the more prominent stars should be protected.

10. If any dark circles around stars is seen, then step back the "starmask" to before step 5 (Dilation) and increase the size.

Thanks for viewing!