Jump to content


Photo

Bi-Color narrow band imaging test

  • Please log in to reply
9 replies to this topic

#1 Hilmi

Hilmi

    Soyuz

  • *****
  • Posts: 3755
  • Joined: 07 Mar 2010
  • Loc: Muscat, Sultanate of Oman

Posted 16 January 2013 - 12:33 AM

OK, first of all some clarification to save you all some some questions:
-That weird white band on the right is my pick-off mirror from my OAG, I didn't realize it needed re-adjustment, must have unlocked it by mistake and sort of moved into the FOV.
-Yes, I don't have enough data, the clouds roll in after only 1 hour of OIII data.


OK, so this is what this is all about, I have been experimenting with roughly approximating natural color using only narrow band filters. So I came up with the idea of using the following for LRGB, I'm sure somebody has tried it before, but to me it's new.

L = Ha
R = Ha
G = OIII
B = OIII

I have 15 10 minute subs for Ha and only 6 for OIII, the image was combined, calibrated stretched and then had some minor smoothing added to reduce image noise, unfortunately, the smoothing also partially kills the crispness of the image.

So, what do you think of my way of doing this? Any suggestions for improving the technique? Is there a better way to approximate natural color purely from narrow band filters?

Posted Image

#2 Mike Wiles

Mike Wiles

    Viking 1

  • -----
  • Posts: 950
  • Joined: 04 Feb 2009
  • Loc: Goodyear, AZ

Posted 16 January 2013 - 02:24 AM

Hilmi,

It's a good image you have there. Typically in a nebula like this, when you find an abundance of Hydrogen Alpha emission you'll also find Hydrogen Beta emission which is in the blue end of the spectrum. Try mixing the blue channel to be 75% OIII data and 25% H-Alpha data.

Mike

#3 Hilmi

Hilmi

    Soyuz

  • *****
  • Posts: 3755
  • Joined: 07 Mar 2010
  • Loc: Muscat, Sultanate of Oman

Posted 16 January 2013 - 02:30 AM

Thanks for the tip, I will try it out soon as I get home.

#4 stephen63

stephen63

    Viking 1

  • *****
  • Posts: 511
  • Joined: 19 May 2010
  • Loc: Central Pa

Posted 16 January 2013 - 01:24 PM

Looks pretty good! Here's a link to another commonly used technique.
http://www.starrywon...chniquenew.html

#5 Hilmi

Hilmi

    Soyuz

  • *****
  • Posts: 3755
  • Joined: 07 Mar 2010
  • Loc: Muscat, Sultanate of Oman

Posted 16 January 2013 - 01:43 PM

Interesting technique, can't see how to try it without owning photoshop. Mike, I tried your technique and I got some odd gradients that were not there before. I wonder what I did wrong.

I did the combine steps as normal in MaximDL but I adjusted the weighting of the different colors so the blue channel is formed as you had suggested.

Must be due to the bad data.

#6 jgraham

jgraham

    Voyager 1

  • *****
  • Posts: 13844
  • Joined: 02 Dec 2004
  • Loc: Miami Valley Astronomical Society

Posted 16 January 2013 - 02:17 PM

Looks pretty good! +1 on the tweak of adding a bit of H-a to O-III for the blue channel (simulated H-beta). My recipe for bi-color is L = Ha + OIII, R = Ha, G = OIII, B = OIII + 30% Ha. For some reason that I've never figured out I get odd gradients in my OIII data that doesn't get removed when using conventional flats. I fix this using synthetic flats (flats made from the image itself) before combining my data.

Have fun!

#7 rigel123

rigel123

    Voyager 1

  • *****
  • Posts: 11158
  • Joined: 29 Jun 2009
  • Loc: SW Ohio

Posted 16 January 2013 - 02:28 PM

The colors came out pretty good with that mix Hilmi! You are getting some nice detail in the horse as well.

#8 David Ault

David Ault

    Messenger

  • -----
  • Posts: 438
  • Joined: 25 Sep 2010
  • Loc: Georgetown, TX

Posted 16 January 2013 - 10:49 PM

Hilmi,

What are you using for processing? Anything with per pixel math operations (i.e. PixInsight) can do this. Depending on what program you are using it may take more steps than Photoshop. You can see the math behind typical blend modes here:
http://en.wikipedia....iki/Blend_modes

They leave out the blend value, which does an alpha blend between the base layer and the new calculated value: [(new*alpha)+{base*(1-alpha)}]
When alpha is 1 (100%) you get only the new calculated value, when it's 0 you get only the base layer. I may have the blend value backwards as I haven't used Photoshop in a long time, but the idea is the same.

Hope this helps.

Regards,
David

#9 Hilmi

Hilmi

    Soyuz

  • *****
  • Posts: 3755
  • Joined: 07 Mar 2010
  • Loc: Muscat, Sultanate of Oman

Posted 17 January 2013 - 01:10 AM

I use MaximDL for processing but I have barely scratched the surface of its features. I still need to learn a lot. Il hqve to go dig up the pixel math stuff to see if it is in there

#10 David Ault

David Ault

    Messenger

  • -----
  • Posts: 438
  • Joined: 25 Sep 2010
  • Loc: Georgetown, TX

Posted 17 January 2013 - 02:07 AM

Hilmi,

That isn't to say you need to process it like that. I think the image you've got now is fantastic! Still, it might be fun playing with the pixel math stuff.

Regards,
David






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics