DEEP SKY WEST ASTROPHOTOGRAPHY FORUMS

You need to log in to create posts and topics.

Orion Widefield - 9 Panel Mosaic Rokinon 135 f2

My version of the 9 panel Orion widefield with the Rokinon 135 f2 setup at DSW.  Rather intense colors, but my wife likes this version and she's the artist so I'm going with it. ?

My first time with a mosaic like this.  I've done a 2 panel but that was nothing.  I now know 99 ways NOT to do a mosaic and one that works, sort of.  I started with these tutorials: http://www.lightvortexastronomy.com/tutorial-preparing-a-mosaic.html  and http://trappedphotons.com/blog/?p=994

Shout out to David Ault's dnaLinearFit script.

I started with assembling just the monochrome versions of the mosaic.  BAD mistake.  Those of you who have done this before may know, but I didn't.  Each channel mosaic looked great, but when I combined RGB it looked terrible.  Bad color gradients that ABE, DBE didn't take out because of the sudden transitions and effects on mid and high tones too.  I tried doing it from the master luminance mosaic and partial frame adaptation but no joy.

I realized I had to process each panel in RGB before assembling.  So, I did simple registration, LinearFit to green, combine, crop, careful DBE (no ABE it trashed it) and watched the color range with STF and putting each panel next to each other.  A long process, but definitely worth the trouble.

So, voila, here mine is.  I will add Ha data if/when the 8, 9 panels get in...should be interesting.

Uploaded files:
  • Orion_10172018_finalish.jpg

Another version with a bit less funky colors.  Maybe slightly more natural, if you can call it that after messing with the arcsinh stretch and dust/scratch filter.  Too many stars in the field, I mean what gives?  Not sure how many, but definitely more than 20

 

 

Uploaded files:
  • Orion_10182018_lesscrunchy.jpg

Since 7 Ha panels were available, I used the bottom six to make a comparison here...started with the same base linear RGB, then registered and put the Ha data on with the NBRGB script, using 5 nm and 1.2 boost.  Processed both identically through the MaskedStretch, throwing the luminance on top after stretched.  Assembled both panels, no Ha on top then with Ha on bottom, used some color adjustment and non very strong processing in Photoshop like zapping stars.  Definite increase in the dust and hydrogen-alpha (no, really?) on the image.

Uploaded files:
  • HaLRGBexp_10182018.jpg

I think all your versions look good.  As you noted we didn't collect Ha on all panels and in fact, I didn't use it at all in my processing.  This target is bright in Ha and for my purposes I found the contribution via luminance was enough.  Plus I did not want to add yet another layer to the processing.  Looks like you found a good way to manage the blending.  Well done.

-Lloyd

A spectacular result, after seeing your version I have entered the desire to reprocess my result, is a great job, I for this was also my first mosaic in LRGB I followed the way to assemble each monochrome panel independently, the result was that I I have been convinced with the result.

Instead for the Swan Mosaic I changed the technique, without any doubt for more than 4 panels is the best way.

Best regards.

Alberto.

Thanks for the compliments!  I agree the Ha came up well in the luminance and red channels but may I counter with this point:

https://apod.nasa.gov/apod/ap181014.html

I realize this is not to everyone's taste, but look how the Ha data tore through like fire in this.  Adding some of those deep exposures won't get there, but it will get a great pull-up on the panels and I think will add a good deal of dynamic range, detail to accent the final images.

 

I sought to reproduce the APOD somewhat, though I added a little customization to my desired effect.  To that end, I used the data with the Ha added into the RGB, MaskedStretch both the Ha narrowband and HaRGB.  I then combined the channels using the Ha narrowband for luminance as well.  The effect was rather...intense.  Proving somebody needs to take my copies of Lightroom and Photoshop away, I transformed the colors into something, er, weird.  I needed to increase the luminance in the blue hue with HSL to bring back the belt stars since the Ha luminance cut the stars so much and what's Orion without the belt stars?

Uploaded files:
  • HaRGB_withHaLuminance_3.jpg

Latest version.  Incorporated the Ha data for all panels, changed it per my current tastes.

Uploaded files:
  • OrionMosaic_12072018.jpg

Awesome results!

I wanted to give a big thanks to Lloyd for rounding out the data for the rest of the panels early this year!!!  I was amazed to see the updates there, thank you, Lloyd.

Great write ups and references and results, RFtinkerer.

Scott Stirling
Graham, NC