14 replies to this topic

[imtl]

Hi all,
 
Nights are short and now with a bright moon so not much time to image anything broadband. I decided to do a first SHO image. To learn some new stuff. I usually do LRGB with added Ha and OIII when its meaningful. 
 
Just a few technical notes.
I captured about 18 hours of data (yes I know I need more )
4 hours Ha (5nm Chroma)
5 hours OIII (3nm Chroma)
8 hours SII (3nm Chroma)
1 hours RGB for star color (Chroma)
 
The moon was >50% at all acquisition nights and conditions were kind of hazy and even high clouds sometimes. So not the best data I ever took but meh, good enough for experimenting and learning. The NB filters including the OIII cut through the moon glow as if it was not there. I'm very pleased with that. 
 
I did a blend of
R=.4*Ha + .6*Sii
G=.4*Oiii + .3*Ha + .3*Sii
B=Oiii
 
I replaced the star color with my RGB and I know there are different more advanced methods out there for that but I think I like the simple way I've done it so I'll stick with that for now.
 
It would be nice to get some feedback as to the color tone and also if you think I overdid something here. I sometimes like what I got and sometimes I think it lacks dimensionality. 
I did not do any noise reduction, and just the slightest of sharpening on the L layer in the stretched stage. All the rest is just curves and HT.
 
Thanks in advance for any pointers/feedback


Abin

[StephenW]

Spectacular image Eyal! .

 

As you already stated,18 hours is not nearly enough - you need at least 100 .   But seriously, the only thing I would suggest would be to bring out the dust clouds a it more - maybe it's just the laptop I'm viewing on, but the darker bands don't seem to be popping as much as they could?

 

Overall though, spectacular image!  Thanks for sharing!

[endless-sky]

Very beautiful image and something that us Europeans will rarely if ever have a chance at imaging.

 

Now you only need to add 82 more hours to it, as stated above... 

[imtl]

Spectacular image Eyal! .

 

As you already stated,18 hours is not nearly enough - you need at least 100 .   But seriously, the only thing I would suggest would be to bring out the dust clouds a it more - maybe it's just the laptop I'm viewing on, but the darker bands don't seem to be popping as much as they could?

 

Overall though, spectacular image!  Thanks for sharing!

Thanks Stephen. I wish I could add a lot more hours but nights are short and I only have 10 more days to image before moving

 

I actually darkened this a bit just before saving! I will look at it again. Thanks.

 

I guess what is bothering me is that I get very very different results if I combine the SHO after stretching each filter or if I combine in linear stage and then stretch. I guess it makes sense since I bring out the details different. But it feels that if I do it in the non-linear stage I am losing the structure differences intensity fidelity. i.e. Ha is much stronger. I am ending up with a lot more balance between the 3 structures then what is actually going on in the nebula.

 

Now, I don't mind it so much in the tone map since nothing in the colors here is real, but the structures are.

 

Opinion?

[StephenW]

Can't say I've ever applied individualized stretches before combining, but I do of course apply LinearFit, so that's already mucking with the balance between the signals...

 

In the end these images are just our own artistic interpretations of the data.  I personally like to see as much structure as I can, so whatever brings that out is ok with me

[imtl]

Thanks Stephen. So, I actually used Jon Rista's suggested linear alignment instead of the regular Linear fit and I thought it was much better for me.
 
The other thing I tried now is to create a super lum from integrating all 3 narrow band images. This of course will change the structure ratio between them as well, but it obviously does not change the noise levels, it just distributes them differently.
 
So, I guess its a matter of what I want to pull out the most. 
 
I did a tiny amount of adaptive stretch to pull out the dust more. But now it looks with less contrast. Meh, I just need a lot more integration time.
 
This region is so great and rich. I can actually see details on NGC2074 the seahorse nebula! Barely, but its there!
The Tarantula nebula is so huge and bright that if it was at the same distance as Orion from us it will actually cast a shadow! its almost 80 times the size of Orion! I love it.
 
Thanks for all the feedback. Much appreciated.


Abin

[imtl]

I know I'm squeezing this thread a lot more than it deserves but something in my previous version just didn't feel right. Over processed maybe. I went back and did a "regular" SHO (R=SII,G=Ha,B=OIII) and made a desirable tone map and just played with curves and HT.
Maybe this version shows less details but it feels smoother and actually has more color diversity which does give the feeling of complexity of structure. I think in both versions I'm hitting the limits of my total integration time. oh well.
 
Thanks all.

[endless-sky]

As far as details go, I definitely prefer this last version. Beautiful!

But I also liked the brown/golden dust of the previous shot...

 

Maybe a blend of the two (with different percentages, 0.25/0.75 - 0.5/0.5 - 0.75/0.25. etc.)? Worth a try?

[elmiko]

Excellent image Eyal! Color and details are superb! You mentioned that you used the simple method to add the rgb stars. Would you mind elaborating?

Edited by elmiko, 28 November 2020 - 05:10 PM.

[imtl]

Thanks Mike. I'm still not sure which version I want as my "final" image (yeah right. I'm gonna processes this all over again in a week's time because its driving me crazy). I changed the "rusty" image as well as the last version. Both are on Abin. Thoughts?

 

Now, regarding your question.

Well, maybe my statement sounded better then the stupid thing I actually did but (and it was stupid trust me), I did the following,

After I finish all processing of my SHO image regardless of what star colors come out, I extracted the L layer and used starnet++ to get a star mask. I blurred it just a tiny with with a convolution of I think it was 1.2 pixels (really nothing).

 

With my RGB image, I just PCC it and stretch. I saturate it a LOT. I then extract the a and b (LAB) of my RGB and I use channel combine with these a and b and apply it to the star masked SHO image. 

 

That's it. It was clean and pleasing and I managed to get all of the stars.  

 

I'm sure there is more interesting better way to do it but I don't know, it worked for me and really its not the most important in matters of complexity in an SHO image so I decided its good enough for me. It was really nice to see the star colors change from magenta and green etc to nice blue-white and red-orange stars while still having the tightness of stars of a narrowband filter.

 

Does this make sense? :O

[imtl]

As far as details go, I definitely prefer this last version. Beautiful!
But I also liked the brown/golden dust of the previous shot...
 
Maybe a blend of the two (with different percentages, 0.25/0.75 - 0.5/0.5 - 0.75/0.25. etc.)? Worth a try?

Okay so I've played with blends of the two versions and masked some parts and basically made a mess of everything. I lost the file at some point with all the windows currently opened in PI .

 

Here is something that came out:

[StephenW]

I think I like #3 best so far, with the more natural (blacker) background.

 

...so many ways to process this data...

[imtl]

Thanks a lot Stephen. I think so too. Yes NB images have infinite amount of possibilities for sure. Just imagine that at the bottom right of this image begins the heart of the Large Magellanic cloud. Its a big nebula!!!
 
Might get a chance to grab 10-15 more hours before I wrap this up. I wonder if I should get a lot more SII or go for the Ha OIII.

[endless-sky]

Okay so I've played with blends of the two versions and masked some parts and basically made a mess of everything. I lost the file at some point with all the windows currently opened in PI .

Very nice! I like how you managed to pull out the colors and details, while also retaining the black background and the golden dust. Well done!

[imtl]

Thanks. Much appreciated. . On the abin version I now also reduced the star size a bit which I think helps the nebula to pop out. The little stars are still there only tighter and a bit fainter