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

Need Help - Please Critique and Process my Color Data

  • Please log in to reply
10 replies to this topic

#1 Daniel Dance

Daniel Dance

    Eastern Voltage Research, LLC

  • *****
  • Vendors
  • topic starter
  • Posts: 1651
  • Joined: 08 Jun 2016
  • Loc: Philadelphia, PA

Posted 15 February 2019 - 04:12 PM

Hello everyone,

 

I'm having some difficulty processing some recent color data.  I am trying to determine whether or not its simply my color processing skills or if I don't have enough color data to begin with.

 

This is color data captured with a Takahashi TOA-130 (f5.15) and an ASI183 camera.

M78 Nebula in Orion

Skies - About 21 SQM

 

36" minutes each (18 x 120s) of R, G, B

 

It would be fantastic is some of the more experience imagers could take a quick look at the data and let me know what they think. 

 

Again, i'm trying to figure out if my color data is adequate and I need to improve my processing skills, or if i'm merely lacking enough color data to begin with here.

 

https://drive.google...iew?usp=sharing
 

https://drive.google...iew?usp=sharing

https://drive.google...iew?usp=sharing

 

Thank you all

 

Daniel

 



#2 mistateo

mistateo

    Apollo

  • *****
  • Posts: 1119
  • Joined: 02 Feb 2017
  • Loc: San Diego, CA

Posted 15 February 2019 - 04:17 PM

I'm nowhere near god's gift to photo processing, but I will try and give it a go after work, IF you post your attempt here first.



#3 Daniel Dance

Daniel Dance

    Eastern Voltage Research, LLC

  • *****
  • Vendors
  • topic starter
  • Posts: 1651
  • Joined: 08 Jun 2016
  • Loc: Philadelphia, PA

Posted 15 February 2019 - 04:25 PM

I'm nowhere near god's gift to photo processing, but I will try and give it a go after work, IF you post your attempt here first.


I dont have an attempt. Im not able to process it well enough.

Im trying to figure out if my processing skills or the actual data is the limitation here. Thank you

#4 roofkid

roofkid

    Viking 1

  • -----
  • Posts: 563
  • Joined: 13 Jan 2016

Posted 15 February 2019 - 05:43 PM

Hi Daniel,

 

this is my attempt:

rgb.jpg

 

There isn't much there to be honest. I am really surprised given the aperture, speed, pixel size and SQM you mentioned. I would have expected a lot more to be visible in ~30m but maybe I am mistaken. I have only imaged M78 once and I found out quickly how hard it is. I had about 4 times as much color data but in retrospect I am not satified with my result ~2.5 years ago.

 

My suggestion would be to capture a lot more :| Of course it will depend on the luminance in the end, but right now I would suspect it wouldn't look pleasing even with a really noise free luminance layer - it would just look very grey in the shadow areas.

 

That are my impressions,

Sven


Edited by roofkid, 15 February 2019 - 05:44 PM.


#5 Daniel Dance

Daniel Dance

    Eastern Voltage Research, LLC

  • *****
  • Vendors
  • topic starter
  • Posts: 1651
  • Joined: 08 Jun 2016
  • Loc: Philadelphia, PA

Posted 15 February 2019 - 06:00 PM

Hi Daniel,

 

this is my attempt:

attachicon.gif rgb.jpg

 

There isn't much there to be honest. I am really surprised given the aperture, speed, pixel size and SQM you mentioned. I would have expected a lot more to be visible in ~30m but maybe I am mistaken. I have only imaged M78 once and I found out quickly how hard it is. I had about 4 times as much color data but in retrospect I am not satified with my result ~2.5 years ago.

 

My suggestion would be to capture a lot more :| Of course it will depend on the luminance in the end, but right now I would suspect it wouldn't look pleasing even with a really noise free luminance layer - it would just look very grey in the shadow areas.

 

That are my impressions,

Sven

Wow!  That is really nice.  I can tell right away that my color processing skills are pretty lacking.

 

If you wouldn't mind, could you describe your processing steps - especially on how you brought out the blue color?

 

That was my biggest challenge - getting that blue out - but you did a really nice job of doing it!

 

Thank you so much!



#6 roofkid

roofkid

    Viking 1

  • -----
  • Posts: 563
  • Joined: 13 Jan 2016

Posted 16 February 2019 - 05:35 AM

Hi Daniel,

 

of course, no problem. I am using Pixinsight for processing. I am not sure if you have it or not. Here's the list of processes I used in order of usage:

 

DynamicCrop --> crop some stacking artifacts away at the edges

DynamicBackgroundExtraction -> not really many gradients but cannot hurt, used in substraction mode

ChannelCombination --> pretty obvious, RGB wink.gif

PhotometricColorCalibration --> explaining that below b/c it's the most important step imo

HistogramTransformation --> from auto STF stretch

Curves --> darkening the background a little b/c it was a little too bright / noisy for my taste

Curves --> increase Saturation a tad

 

I think PhotometricColorCalibration is the most important step to get the colors right at the end. You need to set the coordinates and date/time right. You can usually extract those from a single sub. Just open the single sub and have it as the active windows (blue header) and click "acquire from image". Depending on what is logged in your FITS header you may need to set telescope focal length and pixel size individually. The second important thing is the BackgroundNeutralization which you can tick in the settings at the bottom. Create a preview in the image where there are no stars and preferably no nebulosity. Click the "from preview" option in PhotometricColorCalibration and select the preview you have created. Then analyze your preview and check the values. Either by using the Statistics process or just by using your mouse cursor and checking the numbers at the bottom of Pixinsight. Make sure the readout is set to "real" values, meaning between 0 and 1. In your case the RGB values were around R: 0.0008x, G: 0.0011x, B: 0.0011x. So I set the upper limit of BackgroundNeutralization to 0.00125. This makes sure that you only neutralize the background of your image and not important color information like stars or nebula information. If you leave it at 0.1 for a linear image the result will be extremely bland looking.

 

If you have more questions, just ask.

 

I hope this helps,

Sven


Edited by roofkid, 16 February 2019 - 05:35 AM.


#7 Daniel Dance

Daniel Dance

    Eastern Voltage Research, LLC

  • *****
  • Vendors
  • topic starter
  • Posts: 1651
  • Joined: 08 Jun 2016
  • Loc: Philadelphia, PA

Posted 16 February 2019 - 08:17 AM

Hi Daniel,

 

of course, no problem. I am using Pixinsight for processing. I am not sure if you have it or not. Here's the list of processes I used in order of usage:

 

DynamicCrop --> crop some stacking artifacts away at the edges

DynamicBackgroundExtraction -> not really many gradients but cannot hurt, used in substraction mode

ChannelCombination --> pretty obvious, RGB wink.gif

PhotometricColorCalibration --> explaining that below b/c it's the most important step imo

HistogramTransformation --> from auto STF stretch

Curves --> darkening the background a little b/c it was a little too bright / noisy for my taste

Curves --> increase Saturation a tad

 

I think PhotometricColorCalibration is the most important step to get the colors right at the end. You need to set the coordinates and date/time right. You can usually extract those from a single sub. Just open the single sub and have it as the active windows (blue header) and click "acquire from image". Depending on what is logged in your FITS header you may need to set telescope focal length and pixel size individually. The second important thing is the BackgroundNeutralization which you can tick in the settings at the bottom. Create a preview in the image where there are no stars and preferably no nebulosity. Click the "from preview" option in PhotometricColorCalibration and select the preview you have created. Then analyze your preview and check the values. Either by using the Statistics process or just by using your mouse cursor and checking the numbers at the bottom of Pixinsight. Make sure the readout is set to "real" values, meaning between 0 and 1. In your case the RGB values were around R: 0.0008x, G: 0.0011x, B: 0.0011x. So I set the upper limit of BackgroundNeutralization to 0.00125. This makes sure that you only neutralize the background of your image and not important color information like stars or nebula information. If you leave it at 0.1 for a linear image the result will be extremely bland looking.

 

If you have more questions, just ask.

 

I hope this helps,

Sven

Oh, that is awesome!  Yes, this absolutely helps!

I wasn't aware PI had a tool like this.

I will give it a try.

Thanks again Sven



#8 Daniel Dance

Daniel Dance

    Eastern Voltage Research, LLC

  • *****
  • Vendors
  • topic starter
  • Posts: 1651
  • Joined: 08 Jun 2016
  • Loc: Philadelphia, PA

Posted 16 February 2019 - 09:19 AM

Hi Daniel,

 

of course, no problem. I am using Pixinsight for processing. I am not sure if you have it or not. Here's the list of processes I used in order of usage:

 

DynamicCrop --> crop some stacking artifacts away at the edges

DynamicBackgroundExtraction -> not really many gradients but cannot hurt, used in substraction mode

 

 

Sven

Sven,

 

How did you approach DBE in this example given that most of the field is nebula / dust?  Can you describe your sample sizes as well as where you placed your sample markers?

 

Thanks again!  smile.gif

 

Dan



#9 Daniel Dance

Daniel Dance

    Eastern Voltage Research, LLC

  • *****
  • Vendors
  • topic starter
  • Posts: 1651
  • Joined: 08 Jun 2016
  • Loc: Philadelphia, PA

Posted 16 February 2019 - 02:21 PM

Hi Daniel,

 

this is my attempt:

attachicon.gif rgb.jpg

 

There isn't much there to be honest. I am really surprised given the aperture, speed, pixel size and SQM you mentioned. I would have expected a lot more to be visible in ~30m but maybe I am mistaken. I have only imaged M78 once and I found out quickly how hard it is. I had about 4 times as much color data but in retrospect I am not satified with my result ~2.5 years ago.

 

My suggestion would be to capture a lot more :| Of course it will depend on the luminance in the end, but right now I would suspect it wouldn't look pleasing even with a really noise free luminance layer - it would just look very grey in the shadow areas.

 

That are my impressions,

Sven

Sven,

Did you do anything special to get that nice deep blue color for the reflection nebula?

 

I'm getting close, but I end up getting a turquoise color no matter what I do. 

 

Thanks again!



#10 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 22314
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 16 February 2019 - 05:03 PM

Your data is a little shallow, but given the f-ratio, image scale and integrated exposure, I guess that isn't too surprising. I would aim for 2 hours a channel with those skies and that scope/camera.

 

Anyway, here is what I get with a photometric calibration, BG_ref on an area of the dark dust just a little below and to the left of the primary reflection nebula, and whitepoint at G2V.

 

L3OaT42.jpg

 

A bit of TGV, stretch, then a bit of saturation curves to enhance the color a bit.


Edited by Jon Rista, 16 February 2019 - 05:05 PM.


#11 roofkid

roofkid

    Viking 1

  • -----
  • Posts: 563
  • Joined: 13 Jan 2016

Posted 17 February 2019 - 05:44 AM

Hi Daniel,

 

well you can basically skip DBE with your data, there were not major gradient in them. But just in case: I usually just use a boosted STF stretch (shift click on the radioactive icon in STF) to determine my sample points. I leave the default sample size, increase the tolerance until the background is mostly light greyish in the sample preview window and make sure that I do not have a sample on a star by clicking through them individuallyand moving them if needed. On your image I only used samples at the edge of the screen and two where I could make out no nebulosity in the middle.

 

For stretching the color I used the CurvesTransformation in S (saturation) mode (rightmost selection option) and a curve that looks like this https://www.desmos.c...ator/aka7e7xaii ... you have to ignore the axis at the link though.

 

The only thing I could offer you is packaging up a pixinsight project. Then you can look at the individual settings on your own time


Edited by roofkid, 17 February 2019 - 05:44 AM.



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