Back to forums

How to get a neat Oiii blue background ? (aka where is the path to the blue oceans in the sky)

Pleiades Astrophoto PixInsight Russell Croman Astrophotography NoiseXTerminator Processing Deep Sky IC1805
10 replies•1.1k views
patrice_so avatar
patrice_so avatar
Hi there, 

I am trying since this summer to get the type of results I want when processing my dual narrow band data. I have some good results, but also others that I am less satisfied with because I would like the Oiii to be as even as the neat oceans I see on some pictures. Here are two cases of mine :
https://www.astrobin.com/prgom9/ (IC1805)
https://www.astrobin.com/t9yznf/B/ (IC1396)

I am indeed very attracted to these very beautiful image where the object seems to pop out like an island of a blue ocean of Oiii. Here is a few examples : 
https://www.astrobin.com/328006/C/?q=IC%201805&camera= (IC1805)
https://www.astrobin.com/ahf224/?q=IC%201805&camera= (IC1805)
https://www.astrobin.com/lgvmrf/?q=IC%201805&camera= (IC1805)
https://www.astrobin.com/nmjd33/B/?q=IC%201805&camera= (IC1805)
https://www.astrobin.com/r1xwt7/ (IC405, I did not dare showing mine...)
https://www.astrobin.com/zwrjcm/ (IC 1396)

I would like to improve, working on the IC1805 image currently. My DNB workflow is usually allong the following lines : 
- calibrate, align and stack and extract Ha and Oiii initial layers in SIRIL; then move files to pixinsight
- dynamic crop, DBE and noiseXterminator
- Linear fit to the less bright file
- Pixelmath to tast (usually HOO with G= 0.7*Oiii+0.3*Ha)
- mild GHS --> duplicate 
- Starnet2++ to get a starless 
On the starless
- full stretch with GHS
- color saturation
- curve to adjust colors
- curve/unshapr/HDR/localhistogram to extract detail
On the duplicate for the star (less relevant to the post topic)
- Image solver and photometric calibration 
- starnet2++ to get the stars
- finish the strech with GHS
- color saturation
- extract ACDRNR luminence mask and invert it
- dilate the mask by morphological transformation
- apply the mask to the stars
- erode the stars by morphological transformation
- deconvolution
- adjust stars with curves
Assemble the final image with pixelmath >~((~starless)*(~stars))

Any hint most welcome !
Helpful Engaging
Sergey Kholmanskikh avatar
Sergey Kholmanskikh avatar
Hello, couple of ideas for you:
1. SCNR on starless will shift it to blueish hues
2. Try dynamic narrowband or "Foraxx palette" (you can easily google it by this name)

Cheers!
David Serquera avatar
David Serquera avatar
Hello, I saw your pic of IC1805. You posted 8 h of total integration time under 88% of moonlight. It seems athat you would need that time just for Oiii under this conditions to start getting better SNR..
Benny Colyn avatar
Benny Colyn avatar
Have a look at the NarrowbandNormalization script
Steve avatar
Steve avatar
I suggest using GHS and stretching the R (SII), G (Ha), and B (OIII) channels separately.  Paulyman Astro has a YouTube about this.   This is a recent effort of mine where I'm pleased with the OIII blue.
David Serquera avatar
David Serquera avatar
Have a look to this
https://youtu.be/2QS2Pyhf7as
patrice_so avatar
patrice_so avatar
Dear all, 

So many usefull comments here ! I now have plenty of options to test. 

To go in the order : 
- @Sergey Kholmanskikh  did try the SCNR technic but I left it behind as I learned that photometric color calibration include that step. This is however non-sens in the case of HOO color. I will tray the foraxx palette as soon as I return from my business trip. 

- @David Serquera  I know that 88% moon is not great. However, I expected that to  produce noise, as opposed to introduce some relief on a larger squale in the Oiii layer. Besides, after investigation, and omitting the noise, it seems that my Oiii is indeed fairly neat and that the issue comes from somewhere else. I will post the layers. Thanks also for the link to the Cuiv video This is a great one. 

- ​​@Benny Colyn Thanks. It is installed, but a the time, I prefer playing with pixelmath. I have the same type of issue using the script. 

- @Steve I will try that asap. It is correct that my stretching of the blue strenthen the blue contrasts to too great extent. Thanks. 

MAny thanks for all these comments. I will take some evening after my return to digets them. 

Clear skies !
patrice_so avatar
patrice_so avatar
Hi guys, 

Here is a combination of the advices of @Sergey Kholmanskikh (foraxx) and @Steve  (stretching channel separately). That looks already much better, because the colours are nice separated but without abusing saturation such that the whole thing look completely over cooked. This one might look overcooked, but it is reasonnable. 

https://www.astrobin.com/prgom9/B/

Any further hint most welcome

CS

Patrice
Steve avatar
Steve avatar
I like it!
Norman Tajudin avatar
Norman Tajudin avatar
patrice_so avatar
patrice_so avatar
Norman Tajudin:
Check out these tutorials from PixInsight as well (1 through 6).  I found them super helpful in managing your color channels...

https://youtu.be/WYPVtmhXJYk?si=RhpP8SF8kMg_K34V
https://youtu.be/c-JnE7LXtn8?si=Al8zh82WSjSMjQgU
https://youtu.be/mEdgzaiWmC0?si=9z6WE9Nh60PMdh6G
https://youtu.be/TYAgHnzs67Y?si=oYuuC34rjv_FysLt
https://youtu.be/YzTKHWDXl-0?si=y-_S07yG7Qd4RE1W
https://youtu.be/upeNoiyFR7M?si=7T3LfNx2wzchj8PW

Thanks @Norman Tajudin ! I had a quick look and I am familiar with a part of the concept, but it is great to learn about the other part and to have the whole package explained in a much clearer way than ever before !