Again, relatively simple processing in PixInsight and heavily relying on RC-Tools ([Star|Blur|Noise]XTerminator).
- WBPP pre-processing, set to maximum quality
- BlurXTerminator (correction only) on each calibrated group of subs (i.e. Ha, Sii, Oiii, R, G, B)
- SHO combination
- PixelMath combine SHO as per;
- R=.4*Ha + .6*Sii
- G=.4*Oiii + .3*Ha + .3*Sii
- B=Oiii
- StarXTerminator to remove stars
- NoiseXTerminator to remove some noise - really needs more data
- BlurXTerminator to sharpen
- Histogram transformation for a bit of an initial stretch, using unlinked Screen Transfer Function as a setting)
- Curves Transformation to stretch - this time messing with RGB/K, c and s.
- Stars are a quick bodge. Source is the orig SHO data.
- Start off with the stars left over from StarXTerminator above
- Crop - as above
- FastRotation - as above
- HistogramTransformation to eliminate the background
- Curves Transformation to stretch
- CorrectMagentaStars - sounded like a good idea, the colours are a bit borked anyway...
- Add in stars into SHO image with PixelMath
- combine(starless, stars, op_screen())
- And following up with
- FastRotation - wanted the result to be more landscape
- Oh, there was a crop in there somewhere...
NOTE: First version didn't have RGB stars and had 20 less Ha frames
Done. Second target I've done SHO narrowband/false colour with. Not sure I like it, but this one pleases me a lot, and it was fun doing it. I found most of the data from a few months back, but added a few hours recently.
As with before, I chose to use blurX first in correct mode across calibrated subs (once per Ha, Oiii, Sii, R, G, B channel) to correct some optical imperfections and possibly some guiding issues. The resulting stars at that early stage were noticeably improved. Combining and sharpening were all quite straight forward from there.
Second revision has 1.67Hrs of Ha and 0.25Hr of RGB more than the first. Stars are definitely improved here.
Comments