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

OSC imaging with dual band filters - Ha/Oiii plus Sii/Oiii

  • Please log in to reply
11 replies to this topic

#1 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 31 October 2024 - 03:24 AM

Hello,

A couple of years ago it was just Ha/Oiii - it was simple enough.

Now we have additional DB filters available - Sii/Oiii, Hb/Oiii and such.

 

I have an Antlia ALP-T Ha/Oiii and Askar D2 Sii/Oiii.

Things got more complicated.

 

I only imaged four targets with both filters so far and still struggling to find a way to combine them pleasantly.

 

I extract Ha and Oiii, Sii and Oiii. Combine the two Oiii.

In the end I get a set of Ha, Sii, Oiii masters to work with.

 

So, I investigated several options available in PI.

 

-- Simple Channel Combination sometimes produces amusing results with SHO or HSO, but generally I prefer a more natural look.

-- NB Colour Mapper (by Mike Cranfield and Adam Block) - has a great potential. I like it.

-- NB Hue Combination (Toolbox, from the makers of GraXpert). Similar to the one above. Like it.

-- Foraxx Palette Utility (Paul Hancock). Very red but I like it.

 

Is there anything else I could try?

 

Over the last week I shot Sh2-188 (the Shrimp nebula):
100x360s with Antlia ALP-T Ha/Oiii

058x360s with Askar D2 Sii/Oiii (planning to collect more to get to 100 frames).

 

Here's different renderings:

 

HSO:

6-hso-crop.JPG

 

SHO:

7-sho-crop.JPG

 

NB Colour Mapper:

8-nbcm-crop.JPG

 

NB Hue Combination:

9-toolbox-crop.JPG

 

Foraxx Palette Utility:

10-Foraxx-crop.JPG


Edited by maxsid, 31 October 2024 - 03:47 AM.

  • psandelle and Josef1968 like this

#2 klangwolke

klangwolke

    Mariner 2

  • *****
  • Posts: 213
  • Joined: 24 Dec 2023

Posted 31 October 2024 - 04:29 AM

I like how you’ve show the different options. Looking forward to reading the comments. I wasn’t aware of “NB Hue Combination” and can’t seem to find it with my Google foo. Do you have a link?  My biggest challenge has been how to “normalize” the different NB channels.  I am constantly going back and forth with linear fitting, color calibration, etc to try and find the right balance.  Perhaps this is more challenging than with traditional NB as the filters are from often from different manufactures, with differing performance.  Any advice?



#3 scanner97

scanner97

    Explorer 1

  • -----
  • Posts: 99
  • Joined: 09 Aug 2024
  • Loc: New Hampshire

Posted 31 October 2024 - 07:27 AM

I wasn’t aware of “NB Hue Combination” and can’t seem to find it with my Google foo. Do you have a link?  My biggest challenge has been how to “normalize” the different NB channels.  I am constantly going back and forth with linear fitting, color calibration, etc to try and find the right balance.  Perhaps this is more challenging than with traditional NB as the filters are from often from different manufactures, with differing performance.  Any advice?

You got me curious and I found this, which looks interesintg although I have not read closely yet.



#4 dac4215

dac4215

    Messenger

  • -----
  • Posts: 429
  • Joined: 24 Jun 2021
  • Loc: Owensboro, Kentucky, Usa

Posted 31 October 2024 - 07:33 AM

I like how you’ve show the different options. Looking forward to reading the comments. I wasn’t aware of “NB Hue Combination” and can’t seem to find it with my Google foo. Do you have a link?  My biggest challenge has been how to “normalize” the different NB channels.  I am constantly going back and forth with linear fitting, color calibration, etc to try and find the right balance.  Perhaps this is more challenging than with traditional NB as the filters are from often from different manufactures, with differing performance.  Any advice?

Here is a video from Adam Block where he uses that process.

https://www.youtube....h?v=td4Z4_fIQOE



#5 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 31 October 2024 - 07:05 PM

Toolbox: https://www.idevicea...ht-toolbox.html

Includes NB Hue Combination and other interesting/useful scripts.

 

 

I like how you’ve show the different options. Looking forward to reading the comments. I wasn’t aware of “NB Hue Combination” and can’t seem to find it with my Google foo. Do you have a link?  My biggest challenge has been how to “normalize” the different NB channels.  I am constantly going back and forth with linear fitting, color calibration, etc to try and find the right balance.  Perhaps this is more challenging than with traditional NB as the filters are from often from different manufactures, with differing performance.  Any advice?



#6 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 01 November 2024 - 02:34 AM

I actually used expressions from that article to produce HOO from Ha/Oiii data.

Created a pixelmath script.

Stopped using it since it didn't make any big difference.

 

Just tried expressions for three channels: Ha/Sii/Oiii.

Worked quite well producing the output similar to Foraxx Utility but not as red.

 

However, I tried to process same Ha/Sii/Oiii data but with only two channels included: Ha/Oiii (HOO) - the result was completely identical to Ha/Sii/Oiii!

Somehow Sii signal has zero contribution to the resulting image.

 

Expressions:

R = (Oiii^~Oiii)*Sii + ~(Oiii^~Oiii)*Ha
G = ((Oiii*Ha)^~(Oiii*Ha))*Ha + ~((Oiii*Ha)^~(Oiii*Ha))*Oiii
B = Oiii

 

and (HOO)

R = Ha
G = ((Oiii*Ha)^~(Oiii*Ha))*Ha + ~((Oiii*Ha)^~(Oiii*Ha))*Oiii
B = Oiii

 

 

You got me curious and I found this, which looks interesintg although I have not read closely yet.


Edited by maxsid, 01 November 2024 - 02:34 AM.

  • scanner97 likes this

#7 Mert

Mert

    Voyager 1

  • *****
  • Posts: 11,453
  • Joined: 31 Aug 2005
  • Loc: Spain

Posted 01 November 2024 - 04:28 AM

Not sure but there is little difference in wavelenght between SII and Halfa. ( SII 672,4 nm and Halfa 656,28 nm )
Filters are 6nm bandwidth so worstcase you could get even less difference between SII and Halfa due to production tolerances.
Not sure if it has to do with the little difference you see between the 2 processes. :shrug:

#8 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 01 November 2024 - 06:47 PM

Ha and Sii are very different in my dataset.

However, in that expression for R, (Oiii^~Oiii) produces very small values close to 0, while ~(Oiii^~Oiii) produces very large numbers close to 1.

In the end, R = (Oiii^~Oiii)*Sii + ~(Oiii^~Oiii)*Ha roughly equals to just Ha.

 

So, with my dataset Sii contribution is very close to 0.

 

 

 

 

Not sure but there is little difference in wavelenght between SII and Halfa. ( SII 672,4 nm and Halfa 656,28 nm )
Filters are 6nm bandwidth so worstcase you could get even less difference between SII and Halfa due to production tolerances.
Not sure if it has to do with the little difference you see between the 2 processes. shrug.gif



#9 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 01 November 2024 - 06:59 PM

OK.. Expressions work good on stretched channel images.

The result is very close to what Foraxx Utility produces (Foraxx is based on those expressions, probably somewhat modified).

 

Ha and Sii are very different in my dataset.

However, in that expression for R, (Oiii^~Oiii) produces very small values close to 0, while ~(Oiii^~Oiii) produces very large numbers close to 1.

In the end, R = (Oiii^~Oiii)*Sii + ~(Oiii^~Oiii)*Ha roughly equals to just Ha.

 

So, with my dataset Sii contribution is very close to 0.


Edited by maxsid, 01 November 2024 - 07:05 PM.


#10 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 01 November 2024 - 07:46 PM

Here's what I've got with expressions and Foraxx applied to stretched channel images.

 

Dynamic SHO (PixelMath)

323-dynamic-sho-crop.JPG

 

Foraxx Palette Utility

322-Foraxx-crop.JPG



#11 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 02 November 2024 - 07:58 PM

Alright... I think, I sorted it all out and now know how I want to process.

 

Below is my semi-final image. 

I still need to collect some more of Sii/Oiii.

 

Right now it's (all Bortle 8):

10hr of Ha/Oiii

07hr of Sii/Oiii

 

Stretched channels combined with "Dynamic SHO" expressions,

then quite heavily processed. Curves, etc.

Stars slightly reduced.

 

333-dynamic-sho_on_stretched-Image50-chrominance-crop.jpg


Edited by maxsid, 02 November 2024 - 08:21 PM.


#12 maxsid

maxsid

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,769
  • Joined: 11 Sep 2018
  • Loc: Sunnyvale, CA

Posted 02 November 2024 - 08:06 PM

Full frame.

RC6/reducer @ 940mm, ASI2600MC.

 

dynamic-sho_on_stretched-Image50-chrominance-psh-sm.jpg


Edited by maxsid, 02 November 2024 - 08:07 PM.



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






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics