HoSS palette and how to easily "re-palette" a mono image (Pixinsight)
Last year, during a long spell of cloudy nights, I decided to take a look at my recent mono data to reprocess it and see if I could improve some images. While working through some of it, I remembered recently seeing other asrtophotographers post HSO palette images. So I decided to experiment a bit with various combinations of palettes. One combination I came up with worked really well for a region within IC 1805. I call it hoSS palette - which might be a little confusing. Here is the pixelmath combination:
R = H-alpha*0.5 + OIII*0.5
G = SII
B = SII
Hence the term "hoSS". Perhaps it should be [ho]SS? Or just HOSS.
Anyway, I thought the palette was very cool - lots of reds/pinks and darker colors. Here was the result (star color was from a different palette):
For this post I want to walk you through the palette combination process in Pixelmath, and show you a very easy way to "re-palette" an image using the luminance channel.
There are several methods of combining your data sets in Pixinsight. For this demonstration, I'm going to skip any linear editing. I find myself doing less of it these days anyway as I refine my processing workflow. So, I'm going to open my three data sets and apply a screen transfer function (STF) stretch them (using the histogram transformation process). If you don't know how to do this, I'll show you in the video.
Once all three sets are non-linear, I combine them into hoSS palette using Pixelmath. Here's the equation again: R = H-alpha*0.5 + OIII*0.5
G = SII
B = SII
Here is the result of the combination:
Now, the problem with this palette is that it only uses the H-alpha signal in half of a channel (red). H-alpha is almost always the strongest and most detailed signal (with the best SNR). So the result of this hoSS combination is quite noisy. Had I been imaging specifically for hoSS palette, I would have doubled the SII integration time.
Regardless, I could continue my processing routine on this image and still end up with a decent result. But, I like to make things easy. Lucky for me, I've already completed a SHO palette edit using this data. So I'm going to import that completed image into Pixinsight, quickly register my hoSS to it, extract the luminance from it, and apply that luminance to my hoSS image. This takes care of my noise problem instantly, and I get the added benefit of better details from the H-alpha signal that is more prominent in the SHO image. Besides, I just want the interesting colors of the palette!
Here are the steps for creating and applying the luminance from a fully-processed SHO image to the hoSS image:
Import fully-processed SHO image into Pixinsight
Open the Star Alignment process and select "view" at the top, then click the arrow next to that and select your SHO image as the reference image
Apply the Star Alignment process to your hoSS image (it will create a new "registered" version). Alignment was needed in this case because I previously cropped the edges of my SHO image during processing.
Optional but recommended: remove the stars from both the hoSS and SHO images using StarXTerminator
Select the SHO image, and extract the luminance channel (top left in Pixinsight)
Open the LRGB combination process. DESELECT the R, G, and B buttons, and select "chrominance noise reduction". Everything else can remain default. Click on the icon next to the L space and select your (starless) SHO image.
Drag the triangle from the LRGB combination process to the REGISTERED hoSS image (also starless).
Here's my starless SHO edit:
Here's the extracted luminance:
And here's the hoSS image after applying the SHO luminance:
And there you have it - a hoSS image with a processed luminance. The brightness of the new luminance desaturates the image a bit. A simple Curves adjustment of saturation and RGB adjustments will begin to bring the colors back to life (see video). This is the fun part - the part where you get to be a bit of an artist. Here's a quick adjustment using only curves:
If you notice some color noise creeping in as you adjust curves, simply apply an instance of ACDNR chrominance noise reduction.
Finally, the image after some more Curves tweaking and recombining with the hoSS stars. Thoughts on the star color?:
Essentially what we've done here is "re-paletted" the SHO image. Now, go back to Pixelmath and create your own "re-palettes". All you have to do is align your new palette image (Star Alignment tool), remove the stars, then apply the LRGB combination. It's a quick and easy way to experiment.
Thanks for reading - as always comments are welcomed and encouraged. I hope you found this helpful.