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

Losing Color on M101

  • Please log in to reply
11 replies to this topic

#1 Trenor

Trenor

    Explorer 1

  • *****
  • topic starter
  • Posts: 86
  • Joined: 01 Mar 2021

Posted 12 May 2021 - 08:05 AM

I was able to get some data on M101 a few days ago when it was clear. I was able to stack the individual LRGB files fine. Due to issues and bad frames, I had 104 Light frames, 34 Red frames, 16 Green Frames, 31 Blue frames. Other than a small water drop that seems to be on the lens all night when I stretch(to view) them in pixinsight they look to have good data in all the channels. Processing seems to go ok until I combine the RGB into one image. The Reds dominate even when using LinearFit but I can still get some blue and a bit of green to show ok with tweaking. When I combine that to my Light file, I lose all color or if it is there it is very faint, and I cannot tell it is there.

 

Is this because I have a lot more lights than the other channels and that is drowning out the color? How is the best way to stack when you have such a big difference in the number of frames? Should I be stacking less of the L, R and B to make up for the fact there are so few G frames?

 

I can post some files when I get home if that helps.

 

Thanks


Edited by Trenor, 12 May 2021 - 08:05 AM.


#2 pyrasanth

pyrasanth

    Skylab

  • *****
  • Posts: 4,198
  • Joined: 08 Jan 2016

Posted 12 May 2021 - 08:29 AM

I would work with a subset of the data to identify the issue. The green currently will be your weakest channel currently if you tried to integrate what you have so try 10 subs of each to build a master light in each filter. Using 10 of each will allow you to assess the channels so you can see which channel you potentially need to pack more data into- this is a good way to see the balance between your filters. 

 

The data channels can be more accurately aligned if you can use photometric calibration on the RGB integration and even better after LRGB integration. You can use curves to experiment with the saturation.

 

I would also stretch the RGB image with arcsinhstretch to extract as much vibrant colour as you can. If this looks too agressive try a very slight histogram stretch then use a masked stretch- both these techniques will do a great job of preserving stars and giving you rich colour.


Edited by pyrasanth, 12 May 2021 - 08:32 AM.

  • Trenor likes this

#3 Trenor

Trenor

    Explorer 1

  • *****
  • topic starter
  • Posts: 86
  • Joined: 01 Mar 2021

Posted 12 May 2021 - 08:36 AM

I would work with a subset of the data to identify the issue. The green currently will be your weakest channel currently if you tried to integrate what you have so try 10 subs of each to build a master light in each filter. Using 10 of each will allow you to assess the channels so you can see which channel you potentially need to pack more data into- this is a good way to see the balance between your filters. 

 

The data channels can be more accurately aligned if you can use photometric calibration on the RGB integration and even better after LRGB integration. You can use curves to experiment with the saturation.

 

I would also stretch the RGB image with arcsinhstretch to extract as much vibrant colour as you can. If this looks too agressive try a very slight histogram stretch then use a masked stretch- both these techniques will do a great job of preserving stars and giving you rich colour.

Awesome, I'll give that a try and see how it goes. Thanks for the tips on those tools. I'm still trying to figure out pixinsight.



#4 Poynting

Poynting

    Mariner 2

  • *****
  • Posts: 252
  • Joined: 01 Apr 2012
  • Loc: San Antonio, TX

Posted 12 May 2021 - 09:31 AM

Try using PhotometricColorCalibration OR ColorCalibration, this should balance the color channels. Use either just HistogramTransformation OR MaskedStretch and a slight HistrogramTransformation to taste to stretch to non-linear state. To recover color, I like using CurvesTransformation with the Saturation option, pull the curve up on the right end a bit, and repeat if needed. Check my Astrobin in the signature for the latest image which is M101. It was lacking in color after stretching but the color comes back after the CurvesTransformation. 

 

I am not an expert, just sharing my relevant datapoint. 


  • Trenor likes this

#5 Trenor

Trenor

    Explorer 1

  • *****
  • topic starter
  • Posts: 86
  • Joined: 01 Mar 2021

Posted 12 May 2021 - 01:15 PM

Try using PhotometricColorCalibration OR ColorCalibration, this should balance the color channels. Use either just HistogramTransformation OR MaskedStretch and a slight HistrogramTransformation to taste to stretch to non-linear state. To recover color, I like using CurvesTransformation with the Saturation option, pull the curve up on the right end a bit, and repeat if needed. Check my Astrobin in the signature for the latest image which is M101. It was lacking in color after stretching but the color comes back after the CurvesTransformation. 

 

I am not an expert, just sharing my relevant datapoint. 

Thanks for the tips. I'll give it a shot when I get home.



#6 Trenor

Trenor

    Explorer 1

  • *****
  • topic starter
  • Posts: 86
  • Joined: 01 Mar 2021

Posted 13 May 2021 - 10:10 AM

I was able to get some of the color back but I was still unhappy with the result. I think that is mostly on me and still learning.

 

Here are my stacks if anyone would like to see what they can do with them. 

 

Meanwhile, I'll keep working and see if I can get better results. I might try processing someone else's known good data and see how that goes.



#7 DubbelDerp

DubbelDerp

    Surveyor 1

  • *****
  • Posts: 1,680
  • Joined: 14 Sep 2018
  • Loc: Upper Peninsula of Michigan

Posted 13 May 2021 - 10:47 AM

I don't really know what I'm doing, nor do I have PI, but I was able to come up with this in APP:

Annotation 2021-05-13 114451.jpg

 

I just composed the image as LRGB, removed the residual gradient, calibrated star colors, and removed the green tint. Boosted saturation a bit and desaturated the background. I think you have some nice color in the galaxy, so your data looks pretty good. I'd say keep at it with this data - there's definitely color in the image.


  • Trenor likes this

#8 pyrasanth

pyrasanth

    Skylab

  • *****
  • Posts: 4,198
  • Joined: 08 Jan 2016

Posted 13 May 2021 - 11:29 AM

Hi,

 

I've processed the data for you. Actually it's very nice data to work with- so it was not too onerous. The saving grace is a very nice L data set that can rescue the most awful RGB data. There was quite a lot of colour data to work with but I guess its knowing how to get that data acceptable to the eye. You can of course play around with whatever colour you like on the galaxy- I left it a tad red as it tended to help remove the green cast which SCNR did not fully remove. I've not calibrated the star colours- I think it looks okay as it is.

 

Let me know what you think.

 

Best wishes

Mark

 

M101_LRGB_CN.jpg


Edited by pyrasanth, 13 May 2021 - 11:37 AM.

  • Trenor likes this

#9 Trenor

Trenor

    Explorer 1

  • *****
  • topic starter
  • Posts: 86
  • Joined: 01 Mar 2021

Posted 13 May 2021 - 11:55 AM

Wow, both of those look much better than what I was able to get. Guess I need a lot more practice.

 

Thanks for taking the time to process it for me. It's good to know where the weak link was even if that turned out to be me. :p



#10 DubbelDerp

DubbelDerp

    Surveyor 1

  • *****
  • Posts: 1,680
  • Joined: 14 Sep 2018
  • Loc: Upper Peninsula of Michigan

Posted 13 May 2021 - 12:08 PM

No worries! I just started trying to learn LRGB with my DSLR, so it's good to practice with as much data as possible.

 

And I've come to realize that I'm almost always the weakest link. We tend to solve problems in this hobby by throwing money at it, but in my experience it's the guy driving the equipment that's usually the problem. grin.gif


  • Trenor likes this

#11 Trenor

Trenor

    Explorer 1

  • *****
  • topic starter
  • Posts: 86
  • Joined: 01 Mar 2021

Posted 13 May 2021 - 06:16 PM

Since you all renewed my faith in the collected data I decided to give it another try. Thanks to the images provided I now had a good idea of what was there. I went back and tried again using a guide I found online and added in some of the tools Mark mentioned in his post (SCNR was a great help in dealing with that green). I think my biggest mistake earlier was combining L with the RGB before trying to color balance and do the color processing. I found it much easer to combine RGB using the LRGB combine. Then clean up the background and color calibrate. Then add the L channel and adjust saturation and balance everything.

 

 

So I ended up with this image after fighting with curves to bring out the color without making it look too processed.

 

 

get.jpg?insecure


Edited by Trenor, 13 May 2021 - 06:21 PM.

  • Poynting and DubbelDerp like this

#12 Trenor

Trenor

    Explorer 1

  • *****
  • topic starter
  • Posts: 86
  • Joined: 01 Mar 2021

Posted 13 May 2021 - 06:24 PM

I think the next clear night I get I'll run some more RGB frames and try to bring the color frame count up to the light frames. For now though I'm happy with the result.




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