Horizons: Narrow Band Rosette Nebula: Demonstration of Technique - MOD images didn't turn out

George Yendrey

Adam,
I tried the technique for PixelMath w/ho_mod and hs_mod files and didn't get the results in the mod files.  Of course, I was using it on my LSHO data (a small set, not very 'deep' image wise) of  IC 5070, the Pelican Nebula.

Only 7 hrs integration time total, so not a large data set for monochrome narrow band.  All pre/post processing was in PixInsight.  The Sii (25x240s) was very very sparse, almost non existent compared to the L, Ha, and Oiii images.

So, I created the ho and hs mod images by multiplying the Ha starless, O3 starless, and S2 starless respectively per the video.  The resulting HO_MOD and HS_MOD images were so dim I was tempted to stretch them again.   The lack of signficant S2 data made the PixelMath setup almost moot; the only spot where S2 really showed up was along the edge of the Pelican head and that is where some orange/red gradiation was evident after following the PixelMath setup for using the _mod images.  Given the sparcity of S2, I suspect a similarly based HOO processing may give better results for The Pelican Nebula.  Suggestions?

I learned a lot from this, but I would like to see your take on NB Post processing when maybe only Ha or Oiii are the majority of the available narrow band image signal.

BlockHead

I agree- you need significant signal in the bands you want to control. In addition, although fainter- sometimes the SII and HII map to exactly the same structures and features- so there isn't much "new" information or wayt o contrast the colors and HOO is probably the way to go. 

I hope to be releasing more NB examples. My goal is to make another "FastTrack Training" like video for NB. However, the example will likely be not a super challenging one like you describe.

-the Blockhead

George Yendrey

Ultimately, I went with a pixel math algorithm for dynamic HOO (foraxX) to create my "base" Pelican image then used color masks to focus on where I wanted the color to be.  The SHO version I tried was very interesting, but not very realistic as the colors became oversaturated/exaggerated (IMO).  The final revision was an almost straight HOO palette with a little bit of enhancement to color and dark structure while working to maintain the cloud detail in the dimmer regions (that show up as black in a lot images).

I used an NB2RGB pixel math algorithm to get some believable RGB stars before adding them back in via Pixel Math (initial stars/starless versions were created with Star Xterminator).

My result(s) can be found here: https://astrob.in/5skzqs/H/