Author Topic: Are narrowband and RGB workflows mutually exclusive?  (Read 8466 times)

Offline joelshort

  • PixInsight Addict
  • ***
  • Posts: 260
    • Buckeyestargazer.net
Are narrowband and RGB workflows mutually exclusive?
« on: 2015 October 22 07:07:23 »
I have asked similar questions before, but unfortunately I'm just not getting narrowband processing in PI. 

What I'm trying to do is enhance traditional RGB data with Ha and OIII in the red and blue/green channels.  Here are my goals:
1.  Retain the contrast and whispy faint nebulosity of the narrowband data
2.  Retain the same RGB color palette with the narrowband data added in

But what I always end up with are cotton candy like colors that bear little relation to the color calibrated RGB data.  Whenever I try to add in the narrowband it completely changes the hue of the resulting image.  I've experimented with creating a linear RHa, GOIII, BOIII separate channel masters and then combining.  I've experimented with treating the Ha and OIII like a luminance.  I've experimented with linear and non-linear images etc.  But I just can't seem to come up with a workflow that meets the above goals. 

I've experimented with all of the obvious scripts that relate to this (SHO-AIP, NBRGBCombination, etc) and again I'm just not happy with the results.  I have also tried following Vicent Peris' tutorial on a "New Method for Narrowband Combination" (or something like that).

Are my goals above attainable, or are these two goals mutually exclusive?  If anyone can point me to a NB/RGB processin workflow that meets the above goals I would greatly appreciate it.  I've spent hours on this and haven't gotten anywhere. 

Attached is an example of what I am getting at, an image of NGC281, RGB color calibrated on the left, narrowband enhanced on the right.  The right image shows the extended faint nebulosity but the red hues of the RGB are completely changed. To my eyes it doesn't look like the reds are just "washed out" but that the hue has actually changed.  If I increase the color saturation it still looks like the pink instead of the deeper red.
Joel Short
www.buckeyestargazer.net
CFF135 f6.7, SV80ST, G3-16200M, QHY163M, QHY183M

Offline jkmorse

  • PixInsight Padawan
  • ****
  • Posts: 931
  • Two questions, Mitch . .
    • Jim Morse Astronomy
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #1 on: 2015 October 22 07:36:10 »
Joel,

Just a thought, but have you tried using the ColorMask script, combined with using the hue adjustment in CurvesTransformation?  The combination of those two tools should tackle exactly what you are trying to get at.

Best,

Jim
Really, are clear skies, low wind and no moon that much to ask for? 

New Mexico Skies Observatory
Apogee Aspen 16803
Planewave CDK17 - Paramount MEII
Planewave IFR90 - Astrodon LRGB & NB filters
SkyX - MaximDL - ACP

http://www.jimmorse-astronomy.com
http://www.astrobin.com/users/JimMorse

Offline joelshort

  • PixInsight Addict
  • ***
  • Posts: 260
    • Buckeyestargazer.net
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #2 on: 2015 October 22 08:35:15 »
I messed around with the color mask script in the past, but I confess I had forgotten about it.  I'll have to give that a shot again.  Thanks for the suggestion.

Recognizing that there isn't a cookie cutter approach, I'd still like to see a basic workflow that people use.
Joel Short
www.buckeyestargazer.net
CFF135 f6.7, SV80ST, G3-16200M, QHY163M, QHY183M

Offline jkmorse

  • PixInsight Padawan
  • ****
  • Posts: 931
  • Two questions, Mitch . .
    • Jim Morse Astronomy
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #3 on: 2015 October 22 08:59:57 »
Joel,

Best I can offer is the workflow in my cribsheet which I think you already have  :-\

Best,

Jim
Really, are clear skies, low wind and no moon that much to ask for? 

New Mexico Skies Observatory
Apogee Aspen 16803
Planewave CDK17 - Paramount MEII
Planewave IFR90 - Astrodon LRGB & NB filters
SkyX - MaximDL - ACP

http://www.jimmorse-astronomy.com
http://www.astrobin.com/users/JimMorse

Offline joelshort

  • PixInsight Addict
  • ***
  • Posts: 260
    • Buckeyestargazer.net
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #4 on: 2015 October 22 10:42:47 »
Yeah Jim I do have your crib sheet and it has been very helpful. 
Joel Short
www.buckeyestargazer.net
CFF135 f6.7, SV80ST, G3-16200M, QHY163M, QHY183M

Offline lucchett

  • PixInsight Old Hand
  • ****
  • Posts: 449
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #5 on: 2015 October 22 13:43:55 »
Hi,
Is the problem tied also the the blend of ha only?
How do you blend r and ha, g and Oiii ?
Thank you,
Andrea

Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1298
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #6 on: 2015 October 22 16:55:13 »
I tend to use different approaches depending on the data.  I think I have finally transcended workflows  :D

For me, the important thing is to process Luminance and colour separately.  Then I can focus on getting the detail I want in the Lum and separately worry about getting the colours right.  I've never had much luck with the combination scripts and I just use PixelMath and the other basic tools.  I use ColorMask a lot, of course  ;)

Cheers,
Rick.

Offline joelshort

  • PixInsight Addict
  • ***
  • Posts: 260
    • Buckeyestargazer.net
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #7 on: 2015 October 22 17:03:49 »
For me, the important thing is to process Luminance and colour separately.  Then I can focus on getting the detail I want in the Lum and separately worry about getting the colours right. 

This is precisely my problem Rick.  I've got a great looking luminance.  It's the color image that I can't get right.  I've got a great looking straight RGB image, but I want to enhance the red with Ha and green/blue with OIII. But no matter what I try adding the narrowband in makes the color look like cotton candy.

So, how do you combine Ha with R, and OIII with green/blue?  You say "pixelmath" and "colormask"...I need the details.    ;D
Joel Short
www.buckeyestargazer.net
CFF135 f6.7, SV80ST, G3-16200M, QHY163M, QHY183M

Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1298
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #8 on: 2015 October 22 18:09:46 »
For me, the important thing is to process Luminance and colour separately.  Then I can focus on getting the detail I want in the Lum and separately worry about getting the colours right. 

This is precisely my problem Rick.  I've got a great looking luminance.  It's the color image that I can't get right.  I've got a great looking straight RGB image, but I want to enhance the red with Ha and green/blue with OIII. But no matter what I try adding the narrowband in makes the color look like cotton candy.

So, how do you combine Ha with R, and OIII with green/blue?  You say "pixelmath" and "colormask"...I need the details.    ;D

Hi Joel,

Does that mean you've already got the detail you want from the NB data incorporated into your Luminance?  The whispy stuff you mentioned in your original post?  Assuming you have and that you're just trying to enhance the colour one simple technique is to use PixelMath to blend NB and RGB colours.

First it is helpful to use LinearFit to make sure all the data is roughly compatible (I'm assuming you're starting with linear data).  Pick one of the masters and LinearFit the others to match it.  Then play around with some different mixes in PixelMath, e.g. Red=0.75*R+0.25*Ha, Green=0.7*G+0.3*OIII, Blue=0.5*B+0.4*OIII+0.1*Ha.  Try a few and see if you can find something you like.  You can use BackgroundNeutralization if needed (it's even worth trying ColorCalibration to see if you like the result.)  Don't aim to get a perfect result - just something which is heading towards what you want.

Now do a linear noise reduction, if needed, and stretch.  After that you can use masks and curves to adjust the colours and saturation to taste.  Maybe you want to make the cyans more blue.  You can do that with a simple Hue curve or you could make a cyan ColorMask and use curves to reduce the green, add more blue, increase lightness, boost saturation, etc.  With a few rounds of adjustment you can make some substantial changes for better (or worse!)

Once you're happy with the colour then combine it with the luminance.  I usually extract the CIE L* from the colour (after ensuring it has a 1:1:1 RGBWS) and LinearFit the luminance to match it before using LRGBCombination.  This step will stop the colours getting washed out if the Lum is too strong.  You can also play a little with the Lightness and Saturation sliders.

As an alternative to the above, a really simple way to add some NB colour to RGB data is to apply the NB as a mask and then increase the appropriate colour with a curve, e.g. apply the Ha as a mask to your RGB image and boost the R curve a little and maybe even a tiny boost in B for the H beta.

Hope that helps.  I don't have a single magic bullet answer.

Cheers,
Rick.

Offline joelshort

  • PixInsight Addict
  • ***
  • Posts: 260
    • Buckeyestargazer.net
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #9 on: 2015 October 22 18:30:01 »
Thanks Rick, I appreciate you taking the time to explain things a bit further.  And I certainly do understand that there is no magic bullet workflow/process. 

I do already have the LUM complete as I want it, including the whispy nebulosity present in Ha that is not present in Red.  I"ll try experimenting with PixelMath rather than relying on the scripts.  Like you said, I also do set the RGBWorkingSpace to 1, extract the CIE L* and linear fit the L with the LUM.  However when I apply the LUM using LRGBCombination, I still get washed out colors.  And simply increasing the saturation does not restore the colors but rather boosts the new candy colors. 

I'll keep experimenting.  I'm just starting to go cross-eyed with all the experimenting that I've done over the course of many hours with this. 
Joel Short
www.buckeyestargazer.net
CFF135 f6.7, SV80ST, G3-16200M, QHY163M, QHY183M

Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1298
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #10 on: 2015 October 22 18:52:18 »
Joel,

LinearFit won't give a perfect result if the Lum and the colour have diverged significantly.  You might also try ColorMask to fix up the "candy" colours after the combine.  Say you're getting the dreaded salmon pinks then apply a red chrominance mask and you'll be able to make major colour and saturation shifts with curves.  Anything is fair game.  You can boost red, add a touch of blue, reduce green, mess with the lightness and saturation, tweak the a* and b* curves, even apply a hue curve.

Cheers,
Rick.

Offline jkmorse

  • PixInsight Padawan
  • ****
  • Posts: 931
  • Two questions, Mitch . .
    • Jim Morse Astronomy
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #11 on: 2015 October 23 07:25:43 »
Joel,

You are absolutely right about the cross-eyed part, but that is also part of the fun.  I can honestly say that, other than early efforts with poor data that I have just put to one side, I have yet to "finish" an image.  But, as Rick makes clear, its all about experiment, experiment, experiment  :smiley:

What I love the most about this software, though is exactly the discussion we are having in this thread.  There are so many different ways to tackle the same problem.  Have to agree with Rick on another thing.  I have pretty much abandoned using a set workflow any more.  Too many options for any given image set.

Best,

Jim
Really, are clear skies, low wind and no moon that much to ask for? 

New Mexico Skies Observatory
Apogee Aspen 16803
Planewave CDK17 - Paramount MEII
Planewave IFR90 - Astrodon LRGB & NB filters
SkyX - MaximDL - ACP

http://www.jimmorse-astronomy.com
http://www.astrobin.com/users/JimMorse

Offline lucchett

  • PixInsight Old Hand
  • ****
  • Posts: 449
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #12 on: 2015 October 23 10:16:54 »
Rick,
when you apply RGBWS 1:1:1
do you normalize the three values to 0.33 only or

do you change also gamma and cromaticity coordinates?

Thank you
Andrea

Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1298
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #13 on: 2015 October 24 15:45:38 »
Hi Andrea,

I normally just set the 3 slider to 1 which is the same as setting them all to 1/3 but easier.  If I'm working with linear data I set the gamma to 1 but otherwise I don't touch anything else.

Cheers,
Rick.

Offline jerryyyyy

  • PixInsight Old Hand
  • ****
  • Posts: 425
    • Astrobin Images
Re: Are narrowband and RGB workflows mutually exclusive?
« Reply #14 on: 2015 October 28 12:05:44 »
Have you ever tried the HaRGB script by AIP... you can very most inputs precisely. 
Takahashi 180ED
Astrophysics Mach1
SBIG STT-8300M and Nikon D800
PixInsight Maxim DL 6 CCDComander TheSkyX FocusMax