Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

NGC 7380 in SHO Mosaic of Varying SCNR - C&C Please

astrophotography ccd dso imaging refractor
  • Please log in to reply
17 replies to this topic

#1 BenKolt

BenKolt

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1965
  • Joined: 13 Mar 2013

Posted 16 September 2019 - 11:30 PM

Greetings!

 

I would appreciate your inputs on which of the following applications of SCNR have yielded the most pleasing results of NGC 7380 (the Wizard Nebula in Cepheus) in SHO.  I varied the protection method and amount of SCNR applied to the green.  In the following mosaic I've annotated these values with "Average Neutral" on the top row, "Minimum Neutral" on the bottom row.  The amounts vary from 1.00 on the left to 0.40 on the right in steps of 0.20.  I find the images and varying colors difficult to see on this post, so I encourage those interested to please take a look at the higher resolution AstroBin mosaic image.

 

NGC7380A_SHO1_103_1x1_0900s_20190825_-20C.v010J_SHO_MOSAIC-1.jpg

 

I'm having trouble deciding which is the more interesting or palatable palette.  It is my tendency to retain some green color cast in my SHO images, however this practice tends to be one that is met with some opposition by many as not being so pleasing.  I tend to agree that with the spread I show here, the ones on the right-most side retain too much green even to my taste.  To my eye, I'm drawn more towards the redder cast of "Minimum Neutral", particularly around an amount of 0.6, although maybe 0.7 may be better still.  This retains a slight but not overwhelming green (in my opinion), the blue is a more contrasting shade, and the red adds interest as well instead of the browner/yellower tint along the top row.  However, I'm interested in hearing others' choices.

 

I realize I am asking for what is for the most part a subjective evaluation, but I'd still like to see the responses, particularly if enough people chime in with comments to form some kind of unscientific consensus, to hear if any of my choices are pleasing or at least land in some kind of norm or acceptable SHO color scheme as to not induce convulsive reactions.

 

Thank you for your inputs, and I look forward to seeing what choices you make.

 

Best Regards,

Ben

 

 

 


  • elmiko and Hondo like this

#2 H-Alfa

H-Alfa

    Ranger 4

  • -----
  • Posts: 386
  • Joined: 21 Sep 2006
  • Loc: Spain

Posted 17 September 2019 - 04:19 AM

My choose would be Minimum Neutral 0.60. I really don't understant why people are so reticent to green. I understand that it doesn't... let me say "exists", but if an image has structures in yellow/orange mixed with blue emissions... the result has to be green, and in my personal opinion, to try to completely kill green leads to muted images. As always, it's a matter of taste...

Enviado desde mi MI 5X mediante Tapatalk
  • schmeah and rockstarbill like this

#3 gunny01

gunny01

    Surveyor 1

  • *****
  • Posts: 1736
  • Joined: 02 Jun 2014

Posted 17 September 2019 - 07:15 AM

0.8



#4 Salacious B Crumb

Salacious B Crumb

    Viking 1

  • *****
  • Posts: 624
  • Joined: 13 Apr 2017
  • Loc: Greater Atlanta Area, GA

Posted 17 September 2019 - 08:40 AM

Easy - the full-blown 1 for me cause it looks closest to a true Hubble palette (this from my non calibrated office monitor)...

 

 

- Mikko



#5 lucam

lucam

    Viking 1

  • -----
  • Posts: 593
  • Joined: 09 Oct 2017
  • Loc: Upstate NY, USA

Posted 17 September 2019 - 08:55 AM

Ben,

 

I don't like to use SCNR as a primary tool to reduce green in the SHO palette. It's quick but it kills one channel of information. What I try to do is move some of the data in the green channel to red and blue to create more pleasant colors. Two tools that are useful for this purpose are ColorMask in PI (smoother) with curves transformations and Selective Color in Photoshop (more flexible). 

 

By applying SCNR aggressively, you are reducing the image effectively to two channels (often very similar to HHO) and crushing color diversity. Out of the palettes you presented, I would say average neutral at 0.8 is close to what I like. However, I would start by pushing data out of green by other means and at the very end apply a light touch of SCNR to smooth out residual unsightly green, if necessary. 



#6 BenKolt

BenKolt

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1965
  • Joined: 13 Mar 2013

Posted 17 September 2019 - 03:50 PM

My choose would be Minimum Neutral 0.60. I really don't understant why people are so reticent to green. I understand that it doesn't... let me say "exists", but if an image has structures in yellow/orange mixed with blue emissions... the result has to be green, and in my personal opinion, to try to completely kill green leads to muted images. As always, it's a matter of taste...

Enviado desde mi MI 5X mediante Tapatalk

Thank you, Alberto.  As I said, that one is closest to my liking and I may end up going with that in the end.  But these inputs from others are interesting to me.

 

 

0.8

Gunny, do you have a preference of the "Average Neutral" or "Minimum Neutral" at 0.8 value?  To me the Minimum Neutral is awfully red, perhaps distractedly so, although not as profound as the 1.00 value.

 

 

Easy - the full-blown 1 for me cause it looks closest to a true Hubble palette (this from my non calibrated office monitor)...

 

 

- Mikko

Thanks for the input, Mikko.  I'm aware that this is a common choice for SHO combination.

 

Ben



#7 BenKolt

BenKolt

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1965
  • Joined: 13 Mar 2013

Posted 17 September 2019 - 04:03 PM

Ben,

 

I don't like to use SCNR as a primary tool to reduce green in the SHO palette. It's quick but it kills one channel of information. What I try to do is move some of the data in the green channel to red and blue to create more pleasant colors. Two tools that are useful for this purpose are ColorMask in PI (smoother) with curves transformations and Selective Color in Photoshop (more flexible). 

 

By applying SCNR aggressively, you are reducing the image effectively to two channels (often very similar to HHO) and crushing color diversity. Out of the palettes you presented, I would say average neutral at 0.8 is close to what I like. However, I would start by pushing data out of green by other means and at the very end apply a light touch of SCNR to smooth out residual unsightly green, if necessary. 

Thanks for your inputs about how best to distribute the data into RGB and the use of SCNR vs. other means such as CurvesTransformation with ColorMask.

 

It has been my desire to stick with an initial strict distribution of the NB data into their respective channels as SHO = RGB, however I am aware of other methods in which the data is balanced out differently using PixelMath.

 

I'll give the ColorMask and CurvesTransformation a try on my data and see how this method differs from application of SCNR.  At first blush I would think that careful application of SCNR with reduced value ought to arrive at similar results as you are suggesting, but I need to play around with it myself and draw my own conclusions.

 

Let me make sure that I am getting your preferred method straight before I proceed.  Should I make and apply a ColorMask of green (that is, protecting non-green) and then adjust (reduce) Saturation in CurvesTransformation, where the green will be the most reduced?  (I suppose the inverse could also be done if the image tolerates it: protect the green using the mask and increase the saturation of all else.)  Is this the right interpretation of your suggestion?

 

Thanks for your clarification.

 

Best Regards,

Ben



#8 lucam

lucam

    Viking 1

  • -----
  • Posts: 593
  • Joined: 09 Oct 2017
  • Loc: Upstate NY, USA

Posted 17 September 2019 - 04:53 PM

Thanks for your inputs about how best to distribute the data into RGB and the use of SCNR vs. other means such as CurvesTransformation with ColorMask.

 

It has been my desire to stick with an initial strict distribution of the NB data into their respective channels as SHO = RGB, however I am aware of other methods in which the data is balanced out differently using PixelMath.

 

I'll give the ColorMask and CurvesTransformation a try on my data and see how this method differs from application of SCNR.  At first blush I would think that careful application of SCNR with reduced value ought to arrive at similar results as you are suggesting, but I need to play around with it myself and draw my own conclusions.

 

Let me make sure that I am getting your preferred method straight before I proceed.  Should I make and apply a ColorMask of green (that is, protecting non-green) and then adjust (reduce) Saturation in CurvesTransformation, where the green will be the most reduced?  (I suppose the inverse could also be done if the image tolerates it: protect the green using the mask and increase the saturation of all else.)  Is this the right interpretation of your suggestion?

 

Thanks for your clarification.

 

Best Regards,

Ben

Ben,

 

Take a look at Rick Stevenson's original thread on the Pixinsight forum when he introduced the ColorMask script for suggestion for the SHO palette:

 

https://pixinsight.c...hp?topic=7751.0

 

Chris Gomez recorded a nice tutorial on SHO processing based on some data I shared with him. He uses the ColorMask script more liberally than in Rick's original tutorial:

 

https://www.youtube....h?v=m68R_eDDBfU

 

Cheers,

 

Luca



#9 Jcwillis4

Jcwillis4

    Vostok 1

  • *****
  • Posts: 100
  • Joined: 16 Jul 2018
  • Loc: Plano, TX

Posted 18 September 2019 - 11:25 AM

I also use colormask for my narrowband images.  I don't like SCNR on narrowband as green super exists especially in SHO/Hubble Palette.  For true color, SCNR is really helpful for light pollution fixes.  Here's my Wizard from last month in SHO.

 

get.jpg?insecure

 

Here's my process:

I do a star removal using Starnet++ (optional)

Stretch the channels separately to give me a good range of colors so I avoid something that looks too flat.

Combine the channels, then use colormask script for the color I'd like to adjust (i.e. green, blue, cyan, etc.)

Apply the mask as-is

Use curves transformation, but focus on the panels for Blue-Yellow or Red-Green.

 

By tweaking the direction of the Blue-Yellow or Red-Green balance, you can change the hue of the color, either minimally or drastically, without sacrificing the color intensity.  This can allow for much richer/deeper colors than what you get with SCNR, which is actually reducing the spread of particular colors, and can leave a "blue" area feeling a bit faint or washed out.

 

When you see the SHO pallette with very Orange tones, that's normally due to tweaking the Green more toward the red, and the baby blue is normally done by tweaking yellow out of the cyan areas.

 

SCNR I sue in very small levels at the end of my process (.15 or .25) to see if it helps for a final pop.   

 

Hope that helps.  No right or wrong way to do this stuff.  It's just my personal preference at the moment.



#10 gunny01

gunny01

    Surveyor 1

  • *****
  • Posts: 1736
  • Joined: 02 Jun 2014

Posted 18 September 2019 - 12:07 PM

Thank you, Alberto.  As I said, that one is closest to my liking and I may end up going with that in the end.  But these inputs from others are interesting to me.

 

 

Gunny, do you have a preference of the "Average Neutral" or "Minimum Neutral" at 0.8 value?  To me the Minimum Neutral is awfully red, perhaps distractedly so, although not as profound as the 1.00 value.

 

 

Thanks for the input, Mikko.  I'm aware that this is a common choice for SHO combination.

 

Ben

No preference, just what is appealing to my eyes.  Many times, I'll just correct in PS or LR.



#11 BenKolt

BenKolt

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1965
  • Joined: 13 Mar 2013

Posted 18 September 2019 - 12:54 PM

I also use colormask for my narrowband images.  I don't like SCNR on narrowband as green super exists especially in SHO/Hubble Palette.  For true color, SCNR is really helpful for light pollution fixes.  Here's my Wizard from last month in SHO.

 

get.jpg?insecure

 

Here's my process:

I do a star removal using Starnet++ (optional)

Stretch the channels separately to give me a good range of colors so I avoid something that looks too flat.

Combine the channels, then use colormask script for the color I'd like to adjust (i.e. green, blue, cyan, etc.)

Apply the mask as-is

Use curves transformation, but focus on the panels for Blue-Yellow or Red-Green.

 

By tweaking the direction of the Blue-Yellow or Red-Green balance, you can change the hue of the color, either minimally or drastically, without sacrificing the color intensity.  This can allow for much richer/deeper colors than what you get with SCNR, which is actually reducing the spread of particular colors, and can leave a "blue" area feeling a bit faint or washed out.

 

When you see the SHO pallette with very Orange tones, that's normally due to tweaking the Green more toward the red, and the baby blue is normally done by tweaking yellow out of the cyan areas.

 

SCNR I sue in very small levels at the end of my process (.15 or .25) to see if it helps for a final pop.   

 

Hope that helps.  No right or wrong way to do this stuff.  It's just my personal preference at the moment.

Thanks for your description John.

 

Many of these responses on the thread have prompted me to start looking into applying ColorMask for adjusting the colors, and I've been doing just that for the last couple of nights.  Your recommendations are timely as just last night I was realizing it might be easier to make the initial channel combination with each channel already stretched, and I was planning to practice that very thing this evening.  I've been combining solely in the linear stage prior to stretching, and I've been coming up with the flattened colors you mentioned.

 

Let me ask you and others this.  When you create the ColorMask, do you recommend a "Chrominance" vs. "Lightness" mask?  I'm wishing to learn the subtleties of each.  I observe that they are similar, but I am sure one is preferred over the other depending on for what puepse I intend to use the mask.

 

Secondly, when adjusting in CurvesTransformation in PI, should I be concentrating only on combined RGB adjustments, single out one channel at a time such as Green, or concentrate on Saturation only?  I've been going through the tutorials linked above, but what I'm looking for is a better understanding of which selection is best to be done for which circumstances, which goal in mind, etc.  (I'd like to develop skills with this that go beyond trial and error if possible.)

 

I very much appreciate the help as this kind of interchange is what drives us to improve and do better.

 

Best Regards,

Ben



#12 FiremanDan

FiremanDan

    Aurora

  • *****
  • Posts: 4771
  • Joined: 11 Apr 2014
  • Loc: Virginia

Posted 18 September 2019 - 02:30 PM

MN 0.60 followed by AN 0.80

I like some green.



#13 bmhjr

bmhjr

    Apollo

  • ****-
  • Posts: 1254
  • Joined: 02 Oct 2015
  • Loc: Colorado

Posted 18 September 2019 - 02:36 PM

Avg 0.4 or Min 0.4.  I like the green.

 

Bill



#14 2ghouls

2ghouls

    Viking 1

  • *****
  • Posts: 789
  • Joined: 19 Sep 2016

Posted 18 September 2019 - 02:45 PM

My personal preference is Minimum Neutral 0.6.  I think it hits a nice balance of showing the interaction of SHO, and it also has some very nice blue tones.

 

Min. Neutral 0.4 is also nice if you like a bit more green.



#15 elmiko

elmiko

    Gemini

  • -----
  • Posts: 3303
  • Joined: 27 Jul 2009
  • Loc: Arizona

Posted 18 September 2019 - 02:59 PM

Average neutral .80 . Really great image and post!

Mike



#16 PhotonHunter1

PhotonHunter1

    Viking 1

  • *****
  • Posts: 628
  • Joined: 19 Mar 2015
  • Loc: Northwest IL (for now)

Posted 18 September 2019 - 07:14 PM

My votes:

average neutral 80

minimum neutral 60



#17 BenKolt

BenKolt

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1965
  • Joined: 13 Mar 2013

Posted 19 September 2019 - 11:46 AM

Thank you all for your responses to my "survey" about which result from SCNR was your favorite.  I appreciate the number of responses and hope you'll understand that I won't be able to respond to each of you individually.  The data sampling of responses is quite small, of course, but enough to suggest a general average consensus right around the middle of my span of images.  Furthermore, if I were to "finalize" my image and post with my personal choice, "Minimum Neutral" with an amount of 0.60, I would not offend most viewers.

 

Having examined workflows and images from those who prefer to make dedicated color masks and adjust colors from there, I am seeing some real advantages to that process in producing interesting and contrasting color shades that don't seem possible using SCNR by itself.  I'm in the middle of applying this technique to this Wizard Nebula data as well as a two-panel mosaic of the Gamma Cygnus (Sadr) region.  With the latter I'm pleased with the unexpected variation of data that can be highlighted in SHO.

 

Thanks again, and I'll continue to look for additional votes, comments and critiques here while I work on expanding my techniques.

 

Best Regards,

Ben



#18 drmikevt

drmikevt

    Viking 1

  • *****
  • Posts: 869
  • Joined: 22 Jun 2015
  • Loc: Burlington, VT

Posted 19 September 2019 - 12:24 PM

Secondly, when adjusting in CurvesTransformation in PI, should I be concentrating only on combined RGB adjustments, single out one channel at a time such as Green, or concentrate on Saturation only? 

Ben - you'll grow to love ColorMask if you don't already.  If you haven't already started to do this, it's really important to monitor the transition areas, or areas at the edges of the created mask for whatever color you're working on.  If you push things too far, you will get a very obvious separation from the colors 'next door' (the area just outside the mask). It's easy to not notice this until later in the processing if you are not zooming in as you make the adjustments to color.  One way to help avoid this is to make sure to set the Mask Blur to 2 or 3 - this helps a lot.  I always use the Chrominance setting, fwiw.  

 

I can only say what I do, but I would not normally think about touching the RGB/K curve when a color mask is applied.  So, for green, I know I'll want to pull down the green a little.  But maybe I also need to pull up the blue, maybe not.  Etc.  Of course, each image is different and needs different things.  Over time, you can learn to predict what's needed.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: astrophotography, ccd, dso, imaging, refractor



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics