Author Topic: Differences between HARVB-AIP, SHO-AIP, NBRGBCombination and how to use?  (Read 12971 times)

Offline akulapanam

  • Member
  • *
  • Posts: 60
I'm curious about the differences between how the three different methods of combining narrowband data work, specifically with an interest in HA + OSC RGB, and if anyone had some good work flows to use.  In particular do you use the tools to combine with linear or non linear data?  I noticed it looks like the SHO author prefers non linear but I could be mistaken.

Offline jerryyyyy

  • PixInsight Old Hand
  • ****
  • Posts: 425
    • Astrobin Images
I think you are correct.  I have used the scripts a lot.  When I first stated using the scripts I looked for better documentation and could not find it... but I just located it. 

The French Astro Image Processing group is I believe the source.

http://www.astro-images-processing.fr/aip/

I am a member and speak French so I have searched for information.  There are tutorials and a lot of materials for members.  Now looking there is a tutorial on these scripts and the author says (in French) that it can be done with linear and non-linear but he prefers non-linear.  Makes a big point of equating background darkness and star size (otherwise halos).  Then shows how to make a master L, combine with the color, lots of options...etc. etc.... pretty complex.

http://www.astro-images-processing.fr/technique/traitement/sho-hst/pixinsight-script-sho-aip.html

I have new data on the Pelican and I'll see if I can process it with this technique and report back.... assuming I understand what goes on.

BTW  The AIP group is I think 25 Euro a year and they have lots of PI learning materials and canned scripts etc etc... but French language skills help. 

« Last Edit: 2015 July 15 07:29:14 by jerryyyyy »
Takahashi 180ED
Astrophysics Mach1
SBIG STT-8300M and Nikon D800
PixInsight Maxim DL 6 CCDComander TheSkyX FocusMax

Offline akulapanam

  • Member
  • *
  • Posts: 60
Thanks!  Unfortunately I don't speak French.  I ended up getting ok results with the HARVG script but I'm not convinced it is the most scientifically accurate approach.

Offline jerryyyyy

  • PixInsight Old Hand
  • ****
  • Posts: 425
    • Astrobin Images
Thanks!  Unfortunately I don't speak French.  I ended up getting ok results with the HARVG script but I'm not convinced it is the most scientifically accurate approach.

I am working on the Pelican in HOS and tediously gathering 2-3 1800s images a night if weather permits.  I'll go through the video and see what can be garnered. 

What I get is that you have a lot of controls in their system.  I think the basic idea is like with lRGB images.  You put together an RGB image with good color but dubious detail.  The get the best L you can and pop it in under (or over) the color so you get the fine detail colored by way you want from the RGB. 

So I had some time this AM and these are their steps (quickly) from the 15 min video:

1.  Try to equate the three images for background darkness and star size.
2.  Open the script and create the L.  Their suggestion is 100% H and 60% S.  Try different "method" of mixing and look in areas where you want nice gradients.... you can try many different methods.  They use "Light or Linear LIght" in the video.  Hit the button to make the image and you can close the script. 
3.  Call the created image "L" and mess with stretching etc to get out the detail you want to get. 
4.  Rename the images L S H O.
5.  Reopen the script and it finds the renamed images automatically, then you go to the bottom to the color Layer to Mixing.  R=S at 100% G=H at 80% and B = O at 100%... in the video. 
6.  Then go up to the L-SHO controls and push the L button (Ever use PhD guiding... Push Here Dummy). 
7.  Et, comme on dit en francais:  Voila! 
8.  Now there are lengthy French commentaries worthy of the Impressionists about getting color balance right, but not too much of one and not too much of the other... but the important point is that you can mix values from all three narrow band images into each of the three colors... mix and match (this is way cool) and I finally realized he is working on the Pelican also!  Each time you push the button you see the new color mix. 
9.  Then, of course, you can push the R button to add the L.  They have their own method for doing this using the check off box.  The alternative unchecked seems similar to regular LRGB combo. Here I had trouble with the linear image and I think I have to get the L stretched earlier in the process. 
10.  The end of the video talks about how to reduce star color... need to listen again.... and states that if you do this linear you have to check all the boxes at the top to get the autoSTF... the video was carried out in stretched images and I think that says a lot about the optimal choice for this method.  You can save you settings and load them for similar SHO images in the future. 

Anyway, Vive La France!
Takahashi 180ED
Astrophysics Mach1
SBIG STT-8300M and Nikon D800
PixInsight Maxim DL 6 CCDComander TheSkyX FocusMax

Offline Warhen

  • PTeam Member
  • PixInsight Old Hand
  • ****
  • Posts: 490
    • Billions and Billions
Hi akulapanam and Jerrry, I've used them quite a bit and when I produced the narrowband miniseries for the PI Part-3 tutorials, I corresponded with Laurent and Phillipe and I also 'decoded' some of the French (and the mis-translated English for that matter. ;>) FWIW, here are my two cents.

Ioannis's NBRGB is terrific for blending NB with an assembled RGB or LRGB so I would stick with that for this purpose. Linear or nonlinear if you prefer, are both fine. 

I would leave the AIP HaRVB and LRVB alone.

The SHO AIP script, though graphically awkward is very nice for straight NB combination (SHO, etc.). Again linear or nonlinear are fine but don't check Optimize STF ever! Although you can also introduce broadband, I would not. Use it also for making nice SynthL's of your NB masters. A second pass can also be done to marry the SynthL with the NB chrominance.

Of course there are some fancy PixelMath methods for mixing Ha and (L)RGB which come in really handy in more complex situations such as HII knots in galaxies. Look to Vicent's methods for these!
« Last Edit: 2015 July 19 19:34:45 by Warhen »
Best always, Warren

Warren A. Keller
www.ip4ap.com

Offline Philippe B.

  • PixInsight Old Hand
  • ****
  • Posts: 399
    • CIEL AUSTRAL
Hello

As I am at the origin of this script with Laurent who has written it, I use "SHO-AIP" very often even for HaRGB images.
I enter the 4 channels images, Ha, R, G, B
I make my mix for the "L" between Ha and R and create a new "L" image
Then I make my mix for the R channel by adjusting coeficients with R and Ha (like 40% of Ha and 60% of R).
Then I use the L+RGB rendering with no such option (no use AIP mix method, stay unchecked).
I can see in real time the result.

NBRGB is also a nice script but the result is so different !!! You must try both depending of what you need.


Offline jerryyyyy

  • PixInsight Old Hand
  • ****
  • Posts: 425
    • Astrobin Images
Indeed the results are quite different!

I have been shooting a number of Sharpless2 objects.  This is Sh2 115 constructed with the same LRGB image and the same H-alpha image but with the two different scripts.  I found the NBRGBC a bit too red for my taste even when scaling turned down to 20%... the default is 120%.... it does grow on you.  The AIP is at default settings... a more pastel... perhaps Monet like effect... 

IMHO there was also an effect that is different on the background LRGB.

AIP:

http://www.astrobin.com/195562/0/

NBRGBC:

http://www.astrobin.com/195562/B/

You can flip back and forth between the images... I am slowly learning.....



« Last Edit: 2015 July 20 07:43:08 by jerryyyyy »
Takahashi 180ED
Astrophysics Mach1
SBIG STT-8300M and Nikon D800
PixInsight Maxim DL 6 CCDComander TheSkyX FocusMax

Offline akulapanam

  • Member
  • *
  • Posts: 60
This is really helpful BUT just mixing Ha into Red is not a very scientifically sound approach. For example, as far as I can tell going through the script for NBRGB all this does is take Max(HA/R)GB.

Ha is also representative of the same atoms emitting Hb only at a third lower intensity in many cases.  So some Ha should be mixed into blue as well.  As Ha/Hb emissions aren't uniform even this will result in more or less blue than in reality in certain areas of the nebula. 

The synthetic luminance should be the best approach for the reasons noted above but I'm finding that it causes pinkish nebula.  I would like to try Vicent pixel math approach that is on the main PixInsight site but as far as I can tell the complete math isn't documented anywhere.


Offline jkmorse

  • PixInsight Padawan
  • ****
  • Posts: 931
  • Two questions, Mitch . .
    • Jim Morse Astronomy
Guys,

Thanks for a great discussion.  Just shot a great set of NB images of NGC7000 this past weekend with my e130D system and can't wait to try out the SHO script which I have dying to get my head around for a while now.  Finally have the right data and, thanks to your thread, have some tools to play with.  I will let you know how it turns out.

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 jerryyyyy

  • PixInsight Old Hand
  • ****
  • Posts: 425
    • Astrobin Images
Guys,

Thanks for a great discussion.  Just shot a great set of NB images of NGC7000 this past weekend with my e130D system and can't wait to try out the SHO script which I have dying to get my head around for a while now.  Finally have the right data and, thanks to your thread, have some tools to play with.  I will let you know how it turns out.

Best,

Jim

Jim,

Please post them where we can compare methods if possible... have to say the Really Red variant got a lot of excess likes on Astrobin.... :)  Guess it is like a red breakfast cereal box on the supermarket shelf. 

If you put it on Astrobin we can flip back and forth....

JY
Takahashi 180ED
Astrophysics Mach1
SBIG STT-8300M and Nikon D800
PixInsight Maxim DL 6 CCDComander TheSkyX FocusMax

Offline akulapanam

  • Member
  • *
  • Posts: 60
One thing I have noticed with the NBRGB and HARVB scripts is that when used on non-linear images the result needs a dose of SCNR or another round of color calibration to eliminate green.  I'm curious if anyone has the same issue?

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4729
IMHO the right way to do SHO these days is to simply do a straight S,H,O -> R,G,B combine and then use RickS's colormask script to fix the colors.

rob

Offline akulapanam

  • Member
  • *
  • Posts: 60
Ok I have made a set of comparison images from my recent NGC6559 data.

NBRGBCombination - Ok stars and nice red color.  Dark nebulas are hard to spot.
http://www.astrobin.com/194889/B/

SHO-AIP method suggested in this thread - Better contrast on dark nebula.  Still nice red color. Crazy looking stars.
http://www.astrobin.com/194889/C/

Other Program - Synthetic Luminescence + LRGB Combine - Awesome stars and dark nebula detail.  Awful pink color and the reflection nebula completely gone.
http://www.astrobin.com/full/194889/0/

Other Program - Synthetic Luminescence + Color Transfer Method - Nice stars and dark nebula detail.  Color not quite as pink.  Reflection nebula still missing.  EDIT.  Looking at it now it maybe that the reflection nebula is just better blended?
http://www.astrobin.com/194889/D/

« Last Edit: 2015 July 20 20:06:18 by akulapanam »

Offline akulapanam

  • Member
  • *
  • Posts: 60
Since I'm at it one more attempt.  This time using NBRGB combo but using the synthetic luminance for ha. 

http://www.astrobin.com/194889/E/

Offline jkmorse

  • PixInsight Padawan
  • ****
  • Posts: 931
  • Two questions, Mitch . .
    • Jim Morse Astronomy
Jerry,

Any issues if I post using Google Drive?  I have my own website that I built with Dreamweaver so I don't use astrobin.

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