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

New Pixinsight SpectrophotometricColorCalibration (SPCC) Tool

  • Please log in to reply
26 replies to this topic

#1 Dan_I

Dan_I

    Viking 1

  • -----
  • topic starter
  • Posts: 680
  • Joined: 22 Sep 2010
  • Loc: France

Posted 24 November 2022 - 06:55 AM

   Hi,

 

as probably most PI users noticed, there is a new color calibration tool available in PI, which is based on the GAIA catalog of star spectra, and seems to be very accurate.  It even takes into account the sensitivity curve of the sensor and the transmission curves of the filters. More details there:

 

https://pixinsight.c.../SPCC/SPCC.html

 

 

Here is a comparison between the different color calibration tools on a test image. Nothing has been changed in the workflow except the color calibration (same background reference, and same broadband galaxy white reference for PCC and SPCC).  I saturated the colors more than usual to highlight the differences.

 

 

First SpectrophotometricColorCalibration (SPCC):

 

test_scc_cn.jpg

 

 

Second PhotometricColorCalibration (PCC):

 

test_pcc_cn.jpg

 

Third ColorCalibration (CC):

 

test_cc_cn.jpg

 

 

Which one do you prefer?

 

 

 

 

Dan


Edited by Dan_I, 24 November 2022 - 12:25 PM.

  • R Botero, calypsob and CharLakeAstro like this

#2 james7ca

james7ca

    Voyager 1

  • *****
  • Posts: 11,385
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 24 November 2022 - 07:04 AM

They are all pretty good, but I think I prefer the PhotometricColorCalibration.



#3 Arie

Arie

    Viking 1

  • *****
  • Posts: 764
  • Joined: 01 Sep 2015
  • Loc: Netherlands

Posted 24 November 2022 - 07:04 AM

3rd


  • Juno18 likes this

#4 Daveone

Daveone

    Ranger 4

  • -----
  • Posts: 310
  • Joined: 18 Jun 2021
  • Loc: Czech republic

Posted 24 November 2022 - 07:11 AM

For my eyes...second  PhotometricColorCalibration is the best ...but probably only god knows wink.gif  



#5 view_into_space

view_into_space

    Lift Off

  • -----
  • Posts: 7
  • Joined: 20 Nov 2022
  • Loc: Switzerland

Posted 24 November 2022 - 08:02 AM

If you still have questions how to install or use the new SPCC process, here is a tutorial: https://youtu.be/5xE9V4OIgEw



#6 view_into_space

view_into_space

    Lift Off

  • -----
  • Posts: 7
  • Joined: 20 Nov 2022
  • Loc: Switzerland

Posted 24 November 2022 - 08:10 AM

Just 3 points to the initial post of Dan:

 

a) Did you enter all the setting accurately? Because the tool is complex.....

b) Accuracy and beauty are not always the same - so the question should not be which version people like more, but which version reflects more the reality.

c) At the end SPCC should not produce per se better images than PCC, but it should be more accurate and especially work in situations where PCC failed, caused by the limited data of the APASS catalog.


  • jdupton, JuergenB and KC_Astro_Mutt like this

#7 Dan_I

Dan_I

    Viking 1

  • -----
  • topic starter
  • Posts: 680
  • Joined: 22 Sep 2010
  • Loc: France

Posted 24 November 2022 - 08:45 AM

Just 3 points to the initial post of Dan:

 

a) Did you enter all the setting accurately? Because the tool is complex.....

 

Of course I would not have posted this otherwise... I assume that I've enough background in AP processing to understand how the tool works.

 

 

Capture d’écran du 2022-11-24 14-31-15.jpg

 

 


b) Accuracy and beauty are not always the same - so the question should not be which version people like more, but which version reflects more the reality.

 

This is obvious, but it depends on the goals. Very few of us are imaging for scientific purposes...

 

What I'm looking for is a tool that gives consistent result (to my eyes) from one image to the other, which is not always the case with PCC.


Edited by Dan_I, 24 November 2022 - 11:42 AM.


#8 pyrasanth

pyrasanth

    Fly Me to the Moon

  • *****
  • Posts: 6,277
  • Joined: 08 Jan 2016

Posted 24 November 2022 - 08:54 AM

I tested this as well see thread https://www.cloudyni...bration-module/ however I don't have a before and after test.



#9 JuergenB

JuergenB

    Messenger

  • -----
  • Posts: 465
  • Joined: 15 Nov 2015
  • Loc: Lilienthal, Germany

Posted 24 November 2022 - 12:21 PM

The broadband galaxy white reference approach is very limited because it assumes that the average color of a galaxy is white. That will only apply to a small proportion of galaxies because there are many examples of reddish, yellowish or bluish galaxies in the sky. For example, the integral spectral type of M 33 is A7, which is bluish white.

 

As a better white reference, I would suggest to use the spectral type of the sun, G2V.



#10 idclimber

idclimber

    Gemini

  • *****
  • Posts: 3,473
  • Joined: 08 Apr 2016
  • Loc: McCall Idaho

Posted 24 November 2022 - 12:31 PM

In reading the announcement thread what I found interesting is Juan's hint towards the future where this data can be used to do localized normalization of the images. 



#11 Dan_I

Dan_I

    Viking 1

  • -----
  • topic starter
  • Posts: 680
  • Joined: 22 Sep 2010
  • Loc: France

Posted 24 November 2022 - 01:07 PM


As a better white reference, I would suggest to use the spectral type of the sun, G2V.

 

 

Let's add it into the poll then. Here it is (with SPCC)

 

 

test_scc_g2v_cn.jpg

 

It seems that there's a strong red cast...


Edited by Dan_I, 24 November 2022 - 01:59 PM.

  • calypsob likes this

#12 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 6,080
  • Joined: 14 Jun 2009

Posted 24 November 2022 - 01:09 PM

The broadband galaxy white reference approach is very limited because it assumes that the average color of a galaxy is white. That will only apply to a small proportion of galaxies because there are many examples of reddish, yellowish or bluish galaxies in the sky. For example, the integral spectral type of M 33 is A7, which is bluish white.

 

As a better white reference, I would suggest to use the spectral type of the sun, G2V.

if you read the documentation they explicitly address this. not that G2V vs. other white references hasn't been a longstanding point of argument. at least they give their rationales in section 2: https://pixinsight.c.../SPCC/SPCC.html

 

rob



#13 andysea

andysea

    Fly Me to the Moon

  • *****
  • Posts: 5,804
  • Joined: 03 Sep 2010
  • Loc: Seattle, WA

Posted 24 November 2022 - 07:20 PM

Thank you Dan for posting this. I tried SPCC on an image of Barnard 13 the other day and the result was slightly different than what I was expecting, based on other images that I had seen of that object.

There are subtle, and sometimes not so subtle, differences between the different color balance tools. I will be using SPCC going forward.


  • R Botero likes this

#14 sharkmelley

sharkmelley

    Fly Me to the Moon

  • *****
  • Posts: 5,840
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 25 November 2022 - 01:06 AM

I have tried SPCC with my (unmodified) Canon cameras.  For both the Canon 600D and Canon EOS R, SPCC gives white balance factors within 1% of those produced by PCC.  SPCC is certainly far more powerful and versatile but that's impressive consistency between the two tools!



#15 drmikevt

drmikevt

    Viking 1

  • *****
  • Posts: 966
  • Joined: 22 Jun 2015
  • Loc: Burlington, VT

Posted 25 November 2022 - 10:48 AM

if you read the documentation they explicitly address this. not that G2V vs. other white references hasn't been a longstanding point of argument. at least they give their rationales in section 2: https://pixinsight.c.../SPCC/SPCC.html

 

rob

I came away from that wondering - so, what are we supposed to do?  They start out by saying how the average spiral galaxy is the "new standard white reference on which color calibration will be based in PixInsight".  And then give a bunch of examples which illustrate why we would not want to choose that as the white reference.  As usual with those two, they leave things a bit open to interpretation but what I think they are saying is that its best, in most cases, to not use the standard white reference and instead to use an object-based white reference.... I think(?)

 

I also found it interesting that, after introducing the Narrowband mode they tell us not to use it.  Or at least, not for SHO:  "To achieve a good Hubble's palette representation, we recommend using the ColorCalibration tool instead"


  • pfile likes this

#16 BenKolt

BenKolt

    Vanguard

  • *****
  • Posts: 2,311
  • Joined: 13 Mar 2013

Posted 25 November 2022 - 01:05 PM

Dan:

 

I have also been testing out the new SPCC tool on M33 data.  Choosing white reference G2V also results in my image being a bit redder than expected.  I'll go back and use the spiral galaxy reference as that seems to be what PI is recommending.

 

Ben



#17 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 6,080
  • Joined: 14 Jun 2009

Posted 26 November 2022 - 02:17 PM

I came away from that wondering - so, what are we supposed to do?  They start out by saying how the average spiral galaxy is the "new standard white reference on which color calibration will be based in PixInsight".  And then give a bunch of examples which illustrate why we would not want to choose that as the white reference.  As usual with those two, they leave things a bit open to interpretation but what I think they are saying is that its best, in most cases, to not use the standard white reference and instead to use an object-based white reference.... I think(?)

 

I also found it interesting that, after introducing the Narrowband mode they tell us not to use it.  Or at least, not for SHO:  "To achieve a good Hubble's palette representation, we recommend using the ColorCalibration tool instead"

yes i have the same takeaway, most of the time average spiral is good enough. i guess in the end we're all just making pretty pictures so i try not to get too hung up on color accuracy.

 

the SHO thing is not too surprising because photometrically the colors in SHO are all mixed up, red/red/blue-green. the "color balance over the whole image" thing for SHO has been their recommendation for a long time.

 

rob



#18 calypsob

calypsob

    Cosmos

  • *****
  • Posts: 7,917
  • Joined: 20 Apr 2013
  • Loc: Virginia

Posted 26 November 2022 - 02:48 PM

I like the 3rd warm image the best



#19 JuergenB

JuergenB

    Messenger

  • -----
  • Posts: 465
  • Joined: 15 Nov 2015
  • Loc: Lilienthal, Germany

Posted 27 November 2022 - 12:34 PM

It all goes back to the human eye. The color calibration would be very different for cats, dogs, fish or insects. Our eyes in daylight situations are perfectly matched to the solar spectrum, emitted by a star with G2V spectral class. Night vision is of course different as most people are aware of.

 

If anyone wants to deviate from the G2V = white approach for the reason of personal taste, why not? As it was written already, most people do not do astrophotography for scientific reasons.



#20 andysea

andysea

    Fly Me to the Moon

  • *****
  • Posts: 5,804
  • Joined: 03 Sep 2010
  • Loc: Seattle, WA

Posted 27 November 2022 - 03:33 PM

I find the SPCC colors more pleasing to my eye, that's why I think it's going to be my standard CC process.

When someone criticizes it at least I will have the rigorous approach of SPCC to cite in my defense LOL


  • CharLakeAstro and tjz like this

#21 dudleyjohn

dudleyjohn

    Sputnik

  • -----
  • Posts: 35
  • Joined: 05 Jan 2016
  • Loc: Pennsylvania, USA

Posted 03 December 2022 - 12:32 PM

The Gaia files are really hard to download. So far I've only managed the small set. I'm getting an error the says it's getting too few samples. Has anyone gotten this error and learned how to fix it?



#22 KC_Astro_Mutt

KC_Astro_Mutt

    Vostok 1

  • -----
  • Posts: 185
  • Joined: 14 Dec 2021

Posted 03 December 2022 - 12:39 PM

What I'm trying to figure out is, at what point in the workflow should SPCC be performed?  Before or after DBE? Does it even matter?



#23 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 6,080
  • Joined: 14 Jun 2009

Posted 03 December 2022 - 12:43 PM

it does matter - if you do DBE afterward the color balance can be upset. if you tick Normalize in the Target Image Correction settings area then the color balance will be preserved, but Normalize is unticked by default.

 

there was a lot of debate about this when PCC came out - whether or not the color calibration would be wrong if you didn't do DBE first. i honestly don't remember the answer, but for me my gradients are so bad that i just have to do DBE first in order to really even see what the image looks like. i just do DBE first and then don't have to worry about Normalize.

 

rob


  • jdupton and KC_Astro_Mutt like this

#24 andysea

andysea

    Fly Me to the Moon

  • *****
  • Posts: 5,804
  • Joined: 03 Sep 2010
  • Loc: Seattle, WA

Posted 03 December 2022 - 03:24 PM

I use DBE prior to color calibration.


  • Juno18 likes this

#25 BKMaynard

BKMaynard

    Viking 1

  • *****
  • Posts: 622
  • Joined: 21 Feb 2013
  • Loc: Maryville, Il. (st.louis area)

Posted 03 December 2022 - 03:35 PM

Adam Block has a new YouTube video and, according to him, SPCC is the most accurate color calibration to date. I'm sure this is true, but it does prove that prior cc techniques are quite good and very close to "real" color.




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