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

Question about star color - please look

  • Please log in to reply
15 replies to this topic

#1 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 25 May 2020 - 12:01 PM

Hello,

 

I just got started in AP about two years ago, I am shooting with a Nikon D5600, ha modified; I have an older Stellarvue SV080S refractor and I use the SFF6-2.5 flattener. I am auto guiding with a Lacerta MGEN 2 using a 50mm guidescope. All this in on an AVX mount. I believe I have some minor tracking or alignment issues (tilt?).

 

I was wondering if folks could offer an opinion of the attached snippet? This is a serious zoomed-in view of an issues I see on a lot of my stars. There is some color mis-alignment at the edges of the star.

 

What do yo suppose the root cause of this is? And how might one address it?

 

Thanks in advance.

Ray 

 

 

 

 

Attached Thumbnails

  • Screen Shot 2020-05-25 at 10.52.40 AM.png


#2 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 25 May 2020 - 12:05 PM

Here's another zoomed in view new an image of M13 I took. I believe that "over processing" is a contributor, but I wonder about other possibilities (chroma, alignment, tracking, etc)

 

 

Attached Thumbnails

  • Screen Shot 2020-05-25 at 11.02.41 AM.png


#3 mxpwr

mxpwr

    Vostok 1

  • -----
  • Posts: 198
  • Joined: 08 Dec 2018
  • Loc: Sweden

Posted 25 May 2020 - 01:10 PM

What are you using for processing?
Certain processing steps are prone to enhance chromatic aberrations.

Things you can try to fix this up:
- separate RGB and align them against each other
- use Starnet++ to separate your stars from the rest and patch up the colours (maybe try to separate RGB and align against each other)
- use a pure gamma stretch for the stars on a duplicate of your linear image and merge it back into your main process (arcsin stretch really pops out the chromatic aberrations)
  • Ettu and Juno18 like this

#4 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 25 May 2020 - 02:11 PM

I use PixInsight.

 

Here's the preprocessing flow: create master bias, create master dark, calibrate flats, create master flat, blink lights (find those worthy to stack), calibrate lights, debayer lights, align, stack. Then crop edges, use dbe for background, neutralize the background, use color calibration.

 

Then I switch to another set of saved process icons and I follow the following:

RGB to 1; create lum mask, HT for lum mask, decon (if desired, PSF, big star mask), mlt noise reduction and then stretch (I've used masked stretch, arcsinh stretch, HT / Levels stretch, .....), HDR if desired, SCNR green removal, mlt post stretch noise reduction, acdnr noise (lightness and chrominance), contrast boos, color boost, saturation boost, small detail enhance (mlt), small star mask, star color boost (if desired), star shrink.

 

That's a lot of details, but thanks for looking.

 

Ray



#5 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 25 May 2020 - 02:18 PM

The longer I stay at home, the worse my typing gets. Here's what I meant to type:

 

Then I switch to another set of saved process icons and I follow the following:
RGB to 1; create lum mask, HT for lum mask, decon (if desired, PSF, big star mask), mlt noise reduction and then stretch (I've used masked stretch, arcsinh stretch, HT / Levels stretch, .....), HDR if desired, SCNR green removal, mlt post stretch noise reduction, acdnr noise (lightness and chrominance), contrast boost, color boost, saturation boost, small detail enhance (mlt), small star mask, star color boost (if desired), star shrink.



#6 Hypoxic

Hypoxic

    Messenger

  • -----
  • Posts: 448
  • Joined: 24 Dec 2011
  • Loc: Oise, France

Posted 25 May 2020 - 02:39 PM

 

 

What do yo suppose the root cause of this is?

 

French chromatic aberration.

Attached Thumbnails

  • post-296150-0-74555200-1590426047.png


#7 mxpwr

mxpwr

    Vostok 1

  • -----
  • Posts: 198
  • Joined: 08 Dec 2018
  • Loc: Sweden

Posted 25 May 2020 - 03:29 PM

The longer I stay at home, the worse my typing gets. Here's what I meant to type:

Then I switch to another set of saved process icons and I follow the following:
RGB to 1; create lum mask, HT for lum mask, decon (if desired, PSF, big star mask), mlt noise reduction and then stretch (I've used masked stretch, arcsinh stretch, HT / Levels stretch, .....), HDR if desired, SCNR green removal, mlt post stretch noise reduction, acdnr noise (lightness and chrominance), contrast boost, color boost, saturation boost, small detail enhance (mlt), small star mask, star color boost (if desired), star shrink.


My recommendation is to check the different stretching for how they effect the stars.
Duplicate your linear image and do for each duplicate one of these
- arcsin + gamma
- adaptive
- gamma

I think you will see that the arcsin amplifies your aberrations most.
If that's the case, you can use the stars from a pure gamma stretch and merge it into your arcsin stretch (using Starnet++).

Other then the obvious to use a reflector, trading chromatic aberrations for other defects. And you kind of always have some glass in the optical path anyway...

#8 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 25 May 2020 - 04:10 PM

Thanks, I've not seen the Gamma stretch, I will look for it.

 

Ray



#9 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 25 May 2020 - 06:39 PM

Can anybody help me with the location of Gamma Stretch? I've done many google searches, but  I can't figure out where to find it on the PI website. It would seem it is no longer a process / script that is in the mainstream.

Thanks,

 

Ray



#10 fmeschia

fmeschia

    Apollo

  • *****
  • Posts: 1,364
  • Joined: 20 May 2016
  • Loc: Mountain View, CA

Posted 25 May 2020 - 08:36 PM

Hi Ray,

 

While I also don’t know what gamma stretch refers to, I think the comments you received are spot on: a bit of residual chromatism from either the main objective, or the corrector, tends to be enhanced exponentially by the arcsinh stretch.

 

What I normally do to keep it under control is a combination of the techniques that have already been posted: I register the three R, G, and B channels while correcting distortion, I limit my arcsinh stretch to an initial step with moderate stretch factor, followed by StarNet++ which then allows me to process independently stars and non-stellar objects. On the stars, I often run a pass of MultiscaleLinearTransform to remove the first few small-scale chrominance layers (chrominance only!), which takes care of the most stubborn chromatic residue.



#11 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 25 May 2020 - 09:08 PM

Thanks to all, I have some new tools to try.



#12 whwang

whwang

    Gemini

  • *****
  • Posts: 3,166
  • Joined: 20 Mar 2013

Posted 25 May 2020 - 09:18 PM

Your first thing to do is to check individual debayered subs, to see if the same exists there.  If yes, then the problem occurs prior to stacking.  If not, then the problem occurs during the stacking.

 

I am saying this because I sometimes saw similar things in my images stacked by PI, and it is caused by the rejection during the stacking process.  Once I relax the rejection threshold, the problem is reduced.

 

The problem may also occur during debayering, or during calibration, or some other steps.  The principle is the same, check if the problem occurs right after certain step to determine exactly what step causes it.

 

Cheers,

Wei-Hao


  • redtag likes this

#13 mxpwr

mxpwr

    Vostok 1

  • -----
  • Posts: 198
  • Joined: 08 Dec 2018
  • Loc: Sweden

Posted 25 May 2020 - 11:53 PM

Can anybody help me with the location of Gamma Stretch? I've done many google searches, but I can't figure out where to find it on the PI website. It would seem it is no longer a process / script that is in the mainstream.
Thanks,

Ray

The histogram transformation does a gamma stretch. Almost any image you see on a screen is gamma stretched with gamma=1/2.2 where y=x^gamma.

https://en.m.wikiped...amma_correction

Edited by mxpwr, 25 May 2020 - 11:54 PM.


#14 CorralesRay

CorralesRay

    Explorer 1

  • *****
  • topic starter
  • Posts: 93
  • Joined: 13 Jun 2018

Posted 26 May 2020 - 06:07 PM

thanks again, I will look at the various subs prior to stacking.

 

I wonder if the HT was the tool for gamma stretching....Thanks.

 

Ray



#15 sharkmelley

sharkmelley

    Gemini

  • *****
  • Posts: 3,229
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 27 May 2020 - 01:26 AM

The histogram transformation does a gamma stretch. 

Does it? 

 

I don't see a gamma exponent appearing anywhere on the dialog for histogram transformation.

 

The documentation for histogram tranformation ( https://pixinsight.c...sformation.html ) refers to an MTF (midtones transfer function) which it describes as a rational function.

 

The documentation then goes on to say:

 "A common misunderstanding is the assumption that MTFs are gamma curves in PixInsight. Actually, rational interpolation allows for much more flexible, accurate and controllable MTF curves in PixInsight than gamma curves are in other applications."

 

Mark


Edited by sharkmelley, 27 May 2020 - 01:29 AM.


#16 mxpwr

mxpwr

    Vostok 1

  • -----
  • Posts: 198
  • Joined: 08 Dec 2018
  • Loc: Sweden

Posted 27 May 2020 - 02:57 AM

Does it?

I don't see a gamma exponent appearing anywhere on the dialog for histogram transformation.

The documentation for histogram tranformation ( https://pixinsight.c...sformation.html ) refers to an MTF (midtones transfer function) which it describes as a rational function.

The documentation then goes on to say:

"A common misunderstanding is the assumption that MTFs are gamma curves in PixInsight. Actually, rational interpolation allows for much more flexible, accurate and controllable MTF curves in PixInsight than gamma curves are in other applications."


Mark

I think it's close enough and gets across the point.


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