Log in

View Full Version here: : Advice with NB Workflow & HST Palette / PixInsight


AstroTanja
28-07-2014, 07:02 AM
(as posted on the PI forum)
Hi everyone.
I recently migrated to narrowband imaging, and struggling to get the hang of my channel combinations. My 1st attempt of the Carina nebula was acceptable https://www.flickr.com/photos/astro-tanja/12910032393/, but my recent image of NGC 6188 doesn’t resemble anything I was expecting.

1.) What is the proper combination in PixInsight, pre or post stretching? I have seen tutorials both ways

The image attached here (or high resolution here: http://www.astrotanja.com/wp-content/uploads/2014/07/ngc6188_preview1_2.jpg) was stretched and then combined in PixelMath using the HST Palette. This is a quick edit as I’ve been playing with linear and non-linear combinations and various other edits - so excuse imperfections & noise.
I used the following processes to get to this result
Calibrated&registered / DBE on individual channels / MT on SII
Histogram stretch
PixelMath to combine stretched channels R=SII, G=HA, B+OII
ABE
Curves / Saturation
LocalHistogramEqualisation
Histogram
MMT & UnsharpMask
I know there’s a lot more to this in processing, but colour is my concern now. Do I treat combining the NB in the same fashion as LRGB? And do the same regulars, like CC and SCNR apply?

2.) But the important question: Very many posts I have found online, that were done in HST palette looks almost the reverse colour of mine. The blues and orange/gold tones were initially what I was after. Example this APOD: http://apod.nasa.gov/apod/ap100716.html
It doesn’t matter which combinations I try within the HST pallet, I can not get to this result.

Can anyone advise where I am going wrong with channel combination on this NB image?
Much appreciated.

RickS
28-07-2014, 09:00 AM
Hi Tanja,

Getting attractive colours from the SHO palette isn't easy (at least, IMHO). To get accurate RGB colours you need to do colour combination and calibration while your data is still linear. Narrowband colour is much more (even more?) subjective so there's no reason why you can't do colour combination after stretching if you want, and that's probably why you see people doing it both ways.

What works for me is a linear RGB combine followed by an unlinked STF stretch. That gives me a starting point with reasonable balance between the colour channels. Then I do a lot of colour tweaking. SCNR is useful to get rid of an overabundance of green. CurvesTransformation with the "H" curve lets you adjust colours selectively, e.g. change cyan to blue. Operations on the "a" and "b" CIELAB curves are my favourite way of adjusting colours, especially selectively using masks (use PixelMath and CIELAB a and b values to create a mask for specific colour ranges). ColorSaturation is also a useful tool.

It takes a while to figure out and it is useful to have an understanding of the CIELAB colour space.

Cheers,
Rick.