Chris has been kind enough to let me have a play with his data, the colour balance that I've ended up with is probably closer to a SHO than the HOO I was expecting where the Ha has turned out orange as opposed to red. That's most likely my colour balancing so far but this has just been a quickish play.
Attached is a crop around the centre and the link [url=https://www.astrobin.com/full/ueuqxn/0/]HERE[url/] is to the full frame with a slightly different stretch.
End to end processing was around 25-30 minutes with a bit of playing around in between. I could have just a much better job on the stars as they look a bit funky but that takes more effort than I could be bothered with at the time
From what I can tell the filter appears to work quite well and isn't showing any weird halos which is nice to know. It's another tick for me in the direction of purchasing one to beat some of the light pollution.
Processing is as follows:
- Make a preview in an area that appears to be a neutral black spot
- Background Neutralisation on that area
- Make a preview in the central bright region (around the pillars)
- Color Calibration using a white balance on the pillars preview and turning off Star Detection. Set the black reference as the black preview
- RGBWorkingSpace to set all to 1.
- Extract CIE*L twice. Once for a L and another for stars
- On the Stars image I played around with Morphological Transformation with a Morphological Medium Operator to correct for the star shapes (tracking errors). Takes a bit of trial and error and fattens the stars slightly but does make them rounder.
- DynamicPSF on both L & Stars to have two difference PSFs for deconvolution.
- Run Deconvolution on Stars to fix some of the error caused by Morphological Transformation (Using the PSF from Stars)
- PixelMath ($T^.1) on both images
- StarNet on L and then StarNet with "Create Star Mask" on Stars
- PixelMath ($T^10) on L
- Deconvolution with 50 passes (with PSF from L)
- PixelMath ($T^.1)
- PixelMath (L+Stars)
- PixelMath ($T^10) on L
- Histogram Transformation both the RGB and L to get the black point at the bottom of the histogram and the mid point 1/4 of the histogram
- LRGBCombine to blend the L and RGB together
- Use L as a Mask, open Color Saturation, set at 0.25 and saturate to your hearts content
I could have created a RGB stars image and set the colour balance to get nice RGB stars to add in, could have also used masks to protect the stars during this process or just removed the RGB stars entirely and then added the L in to a starless RGB (makes subtle looking stars) but it was all more work than I wanted
Hope this gives some of you some potential processing ideas
