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

Nothing to do? Want to practice processing?

  • Please log in to reply
606 replies to this topic

#326 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • Posts: 896
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 28 January 2021 - 05:36 AM

Had a chance to give the glob another go!  And I even managed to get it into a gallery!  Now, we will have to see about linking.  If it ends up a little mini picture like Jonny, I might have to edit and try something different lol.

 

I for sure got deeper into the core this time, even with the yellows.  Less saturation, not as blue.  Not real sure I got the evenness right in the main blob there, but overall it was fun and I think it's pretty good.  Best I've done, anyway.  (Since I've done 2 globs now).  No star reduction in the background field.  One of the modules in ST already does that a wee bit, but I halved the strength on that.  It may be busy, but they are all so pretty I couldn't go about erasing stars.

 

And turns out the line wasn't a line, so no satellite trails to erase or heal (I have plenty of practice on that already).  Some kind of artifact when using Windows 10 Photos and zoom it in all the way.  But I didn't see it anywhere else.

 

So, the link....let's see here...and which size will it give me?

 

gallery_345094_15786_74021.png


  • F.Meiresonne, cybermayberry, imtl and 1 other like this

#327 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 28 January 2021 - 05:43 AM

Alright! So forget about my laziness comment from before. I'm the most industrious person in the world :p :p

 

You got some nice stars and colors. You do have some brightness issue in the core. Looks like you did some sort of HDR overcorrection. Or maybe the glob is about to explode :)

 

I'm glad you're having fun with it.


  • Mike in Rancho likes this

#328 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • Posts: 896
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 28 January 2021 - 02:07 PM

Alright! So forget about my laziness comment from before. I'm the most industrious person in the world tongue2.giftongue2.gif

 

You got some nice stars and colors. You do have some brightness issue in the core. Looks like you did some sort of HDR overcorrection. Or maybe the glob is about to explode smile.gif

 

I'm glad you're having fun with it.

I imagine we all dream of photographing something exploding up there!  Alas, probably won't happen.

 

Yes I used a different HDR scheme that allowed me to tame the core and go deeper with the yellows, and then the deconv - which I had to go pretty light on - resolved fairly well I thought.  The problem is the HDR sort of created a Target/bullseye look to the glob, and I couldn't find the right levels to smooth that out evenly.  Might have to give it a third try then, with different settings or see if that can be manually cleaned up.



#329 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 28 January 2021 - 02:10 PM

Have fun Mike! Let me know how it works for you!



#330 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 07:25 AM

Another approach.

 

I find the color of the galaxie nicer, i put a mask on stars and leave them out for coloring. To me i looked to work somehow better ohterwise you wind up with that strange light cyan colors.

Then i inverted the mask and tried to get some color in the stars.

 

full image

https://www.cloudyni...911_1754067.png

 

Rather nice imo, but my fat stars are still to fat...too much sugar and carbs...grin.gif

 

Dunno what you guys will think...

Attached Thumbnails

  • M101SCO2.jpg

Edited by F.Meiresonne, 29 January 2021 - 07:30 AM.


#331 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 29 January 2021 - 07:32 AM

Freddy,

 

This is my personal approach so take it as such.

 

Your background is very clipped. besides the fact that background is not black like that, it also makes your object edges be very sharp and look unnatural and over processed. I would suggest a much smoother transition between background and object. 


  • F.Meiresonne likes this

#332 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,625
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 29 January 2021 - 09:37 AM

Freddy,

 

the key is the right stretching operation. Histogram transformation is not the way to go. In PI there is arcsinh stretch that allows for a fine balance between background and object. Check out what stretching methods you have in your software. During the initial stretching you want the background clearly visible all over the image. You can dim it down at the end of your processing if you want it darker. If you dim down the background too early you loose the fine transitions at the edges of the object.

Here is an example stretch sequence. Starting point is after background removal, the image still pitch black.

left to right, upper row:

- arcsinh so that the background is there and the object is flat and boring. This is a color preserving or even enhancing stretch.

- histogram to brighten ob the object.

- is curves (luminance) for more contrast, dimming the background a little bit

- is color yellow/blue balance and saturation

lower row:

- local histogram equalization under a luminance mask, operation restricted to bright features

- curves transformation (all channels)

- unsharp mask using a luminance mask, operation restricted to bright features

 

The only operation done before this sequence was background removal, subtraction of a linear mostly red gradient. Well, a crop of course to match the file size limits. Nothing else. No color calibration, no denoise, no nothing. Maybe you can resemble the steps in your software (StarTools I think).

 

stretching.jpg


Edited by the Elf, 29 January 2021 - 09:57 AM.

  • imtl and Ethanwyh like this

#333 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 10:14 AM

Yes i guess i over stretch it somehow. There was kind of a gradient in the picture above M101, guess by darkening the background i got rid of it but maybe winded up with a too black background.

 

Allthough i don't mind it so much , personally i find my stars not that good or at least a bigger issue

 

Also that is why i like to have opinions of people because i don' t see it always good, i was never a artistic guy, i have no feeling for it. 

 

Yes i only use Startools. Stretching is done sort of by default but you have to tune it by eg masking or narrowing zones what you want to stretch and the use of parameters settings. i have difficulty sometimes in evaluating these things.

Probably way different then PI.

In GIMP you use curves, you won't find these in Startools.

The module there for stretching is called Auto Develop

 

Attached Thumbnails

  • Stretch.JPG


#334 Manitu

Manitu

    Explorer 1

  • -----
  • Posts: 56
  • Joined: 04 Mar 2020
  • Loc: South Germany

Posted 29 January 2021 - 11:05 AM

In my approach I used background extraction, photometric color calibration and green removal in Siril and then started Startools processing.

In Startools I used stretching very carefully and did a final Asinh stretch in Siril.

 

When I do stacking in Siril I also use linear gradient removal in every single frame before registration and stacking operation. But here it is already stacked ;-)

 

I am working on a second image with Asinh stretch in Siril, then histogramm transformation in Siril and after that processing in Startools. But then I cannot use gradient removal and noise removal in Startools because these are only available for linear data images.

 

CS Tom



#335 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 29 January 2021 - 11:34 AM

When I do stacking in Siril I also use linear gradient removal in every single frame before registration and stacking operation. But here it is already stacked ;-)

 

 

I'm not sure this is a good move, with single frames your snr is quite low which will cause a gradient removal took to also remove quite a bit of good data as well. I need to think about this for a bit but this is sort of a gut feeling. It will also affect the noise profile of the frame which might cause later artifacts when doing noise reduction. But again, I need to think about this more.



#336 sbharrat

sbharrat

    Viking 1

  • -----
  • Posts: 602
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 29 January 2021 - 12:33 PM

I'm not sure this is a good move, with single frames your snr is quite low which will cause a gradient removal took to also remove quite a bit of good data as well. I need to think about this for a bit but this is sort of a gut feeling. It will also affect the noise profile of the frame which might cause later artifacts when doing noise reduction. But again, I need to think about this more.

I also had a question about this when the Siril team suggested this... but more from the perspective that this gradient removal is completely automatic and I couldn't understand how it would work across the range of object types. Pretty diametrically opposed to what Ivo would say I imagine (and now you) ... 



#337 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 12:33 PM

I installed Siril , but looks strange to me, i think again needs some studying what i can use it for....

 

My brain hurst allreadygrin.gif



#338 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 29 January 2021 - 12:35 PM

I also had a question about this when the Siril team suggested this... but more from the perspective that this gradient removal is completely automatic and I couldn't understand how it would work across the range of object types. Pretty diametrically opposed to what Ivo would say I imagine (and now you) ... 

Shaun,

 

This is starting to be off topic so maybe better to open a new one to discuss this. I find it somewhat strange that they suggested this. But again, I need to think this through first.



#339 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 12:37 PM

I also had a question about this when the Siril team suggested this... but more from the perspective that this gradient removal is completely automatic and I couldn't understand how it would work across the range of object types. Pretty diametrically opposed to what Ivo would say I imagine (and now you) ... 

Also, it is allready difficult enough to deal with one program...takes time to get it a bit the fingers...let alone going along with other stuff like Siril ,APP or whatever...



#340 sbharrat

sbharrat

    Viking 1

  • -----
  • Posts: 602
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 29 January 2021 - 12:55 PM

Shaun,

 

This is starting to be off topic so maybe better to open a new one to discuss this. I find it somewhat strange that they suggested this. But again, I need to think this through first.

I had already started a discussion on this in the past:

cloudynights.com/topic/746430-gradient-removal-in-siril-make-sense

 

so no need to continue... I was just commenting because I was glad to see that you are also skeptical of the approach (from a different and more important perspective).

 

Thanks and cheers



#341 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 01:06 PM

Another go at that amazing Omega Centauri

 

Bit more color, , though, still maybe a bit too much blue stars around the  core. Globs are old so maybe not much blue stars around in such objects dunno. It is a stunning object.

 

https://www.cloudyni...911_1955168.png

Attached Thumbnails

  • OC.jpg

Edited by F.Meiresonne, 29 January 2021 - 01:17 PM.

  • imtl, the Elf and Mike in Rancho like this

#342 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 29 January 2021 - 01:10 PM

This is an awesome image you produced. Well done. I love it.


  • F.Meiresonne likes this

#343 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 01:11 PM

How much time took this to capture, Eyal?



#344 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 01:11 PM

This is an awesome image you produced. Well done. I love it.

Thanks,



#345 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 29 January 2021 - 01:34 PM

How much time took this to capture, Eyal?

Freddy, the details are in the Abin link in the signature. But here is the direct link

this is 5.3 hours integration in total.



#346 jonnybravo0311

jonnybravo0311

    Surveyor 1

  • *****
  • Posts: 1,711
  • Joined: 05 Nov 2020
  • Loc: NJ, US

Posted 29 January 2021 - 02:40 PM

Since there was so much action around the M101 data, I figured I'd give it a shot. I needed a break from imtl's mosaic of the fighting dragons - not because I'm not interested in completing the work on it, but rather, because I have been staring at it for so long, I needed a distraction :).

 

Here's what I ended up with:

 

med_gallery_347158_14910_958475.jpg

 

I blended in some of the Ha data. That actually turned out a bit more of a pain than other HaRGB combinations I've done. I ended up creating a mask from the Ha image and clipping it and using the clone stamp tool so that only a few areas in the galaxy were unprotected. I then used pixel math on the masked RGB image to bring in the Ha. I would have liked the Ha I did bring in to be more red than pink - and maybe I'll go back to it at some point to evaluate ways I could have gotten a better/deeper red there.


  • the Elf likes this

#347 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 29 January 2021 - 03:02 PM

Since there was so much action around the M101 data, I figured I'd give it a shot. I needed a break from imtl's mosaic of the fighting dragons - not because I'm not interested in completing the work on it, but rather, because I have been staring at it for so long, I needed a distraction smile.gif.

 

Quitter! :p I had faith in you.......

 

Great processing jonny, the blending of the Ha and the smoothness of the image is really good. Well done!


  • jonnybravo0311 likes this

#348 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,232
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 29 January 2021 - 03:03 PM

Nice image indeed, but the Ha indeed comes in pink. On the Elfs image also actually...



#349 imtl

imtl

    Gemini

  • *****
  • Posts: 3,491
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 29 January 2021 - 03:16 PM

Jonny and Freddy,

 

If you want to get more "red" than "pink you can try and blend the Ha directly into the red channel and add also a bit into the blue (i.e. Hb). In case you used Ha as Lum, then pink is what you're going to get. White+red=pink.

 

You can also play with Hue and shift the colors. Make an Ha mask and apply first.

 

 

***By the way, getting non red for string hydrogen regions is not a problem, it actually is part of the real signal sometimes. Especially with nebulae with strong Hb. Those cases the Hydrogen will actually be sort of Magenta color (Ha:red + Hb:blue = magenta).



#350 jonnybravo0311

jonnybravo0311

    Surveyor 1

  • *****
  • Posts: 1,711
  • Joined: 05 Nov 2020
  • Loc: NJ, US

Posted 29 January 2021 - 04:23 PM

Jonny and Freddy,

 

If you want to get more "red" than "pink you can try and blend the Ha directly into the red channel and add also a bit into the blue (i.e. Hb). In case you used Ha as Lum, then pink is what you're going to get. White+red=pink.

 

You can also play with Hue and shift the colors. Make an Ha mask and apply first.

 

 

***By the way, getting non red for string hydrogen regions is not a problem, it actually is part of the real signal sometimes. Especially with nebulae with strong Hb. Those cases the Hydrogen will actually be sort of Magenta color (Ha:red + Hb:blue = magenta).

That's what I was thinking about when I said I might get back to it... Pretty sure my pixel math expression was something like...

 

R/K = $T[0] * 0.9 + Ha * 1.4

G = $T[1]

B = $T[2] * 0.95 + Ha * 0.05

 

The multipliers might not be exactly what I used... and I was also goofing around with max ($T[0],Ha * multiplier) and iif (expression, true, false) as well.




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