Author Topic: NGC 7479 - PI vs. Photoshop  (Read 8846 times)

Offline Josh Lake

  • PixInsight Old Hand
  • ****
  • Posts: 424
    • View Profile
NGC 7479 - PI vs. Photoshop
« on: 2012 November 15 17:48:32 »
I'll occasionally engage skilled astrophotographers on Reddit and ask them if they've tried PixInsight. Several people have come over to the PI side once they see what it can do with their data (same as me!).

One user recently posted this nice shot of NGC 7479: http://i.imgur.com/iuJzU.jpg

I made my pitch about working it out in PixInsight, and he graciously allowed me to download his data and try it. I worked with it for a bit but did not come close to the result he got in Photoshop CS2.

He said his process was:
Quote
Acquisition, calibration, 2x resize, stack, and align LRG&B with MaxIm DL.
Open LRG&B in Ps CS2 using fits liberator (32 bit conversion). Convert to 16 bit using Local Adaptation (1 pixel) and adjusting the curve.
Combine the RG&B, Convert to lab color, replace L with real Lum.
Then, begin the magic; color adjustment, noise reduction, unsharp masking, cropping, resizing, repeat several times over several days. Finally get a version that looks ok.

Anyone else want to give it a shot? I may try again when I have a bit more time, but I wanted to see what the real pros in here could do.  :D

Here is the data: https://www.dropbox.com/s/7yp0qt6gne3l2l4/ngc7479.zip  (just two files, the RGB and L masters, from Maxim I assume)

Edit: updated link
« Last Edit: 2012 November 16 12:05:06 by jlake »

Offline Jim_

  • Newcomer
  • Posts: 30
    • View Profile
Re: NGC 7479 - PI vs. Photoshop
« Reply #1 on: 2012 November 16 09:55:05 »
I get a page not found error on the link.
Jim

Offline Josh Lake

  • PixInsight Old Hand
  • ****
  • Posts: 424
    • View Profile
Re: NGC 7479 - PI vs. Photoshop
« Reply #2 on: 2012 November 16 12:05:41 »
I get a page not found error on the link.
Jim

I guess it was pulled off the original server. I have edited the link to point to my Dropbox. It should work now!

Offline Alejandro Tombolini

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1267
    • View Profile
    • Próxima Sur
Re: NGC 7479 - PI vs. Photoshop
« Reply #3 on: 2012 November 17 20:58:38 »
Hi, I made a try

The image is a little complicated regarding saturation, and besides stars are quite big. It demanded work a lot with masks. I would like to have the data aligned and integrated in PixInsigth

Align the RGB and L using DynamicAlign tool



Then generate a crop of both images using DynamicCrop to eliminate dark areas.



Using AberrationSpotter script generate an image of the corners of the RGB image to be use as Reference image in BackgroundNeutralization tool



Idem for ColorCalibration. Generate an image of the corners to be use as Reference image for Background Reference and a preview to be use as Reference image for White Reference.



Now apply DynamicBackgroundExtraction for both, RGB and Luminance.
DBE to RGB



DBE to Luminance



I need a starmask to be use during deconvolution. For build it, clone Luminance and apply HistogramTransformation with the actual ScreenTransferFunction



And generate the starmask over the clone.



DynamicPSF to generate de PSF to be used in Deconvolution tool. (I just noticed that the values of A are a little high, Juan recommend A<0,25 ..., but at the moment I am writing this, the picture is already finished...)



Apply Deconvolution using the external PSF and protecting stars with starmask, to avoid problems in the core of the stars.



Stretch the Luminance using HitogramTransformation tool



Now I am going to use MultiscaleMedianTransformation to increase details in the galaxy. At the same time is needed to protect both stars and background.
Using StarMask generate a mask that cover all the image but the galaxy.



To increase the protection in stars cores added using PixelMath the previous Starsmask built for deconvolution



Protecting stars and background apply MMT



Once again I will use MMT to increase luminosity in the galaxy and make the background a little darker, working in different layers. For this I needed a starmask more adjusted to stars.



Protecting stars apply MMT.



Invert the starmask and use MorphologicalTransformation Tool to reduce stars. (two times)




The stars in RGB image are nicer than Luminance, then I am going to protect them (the bigest at least) during  LRGB combination
Generate another starmask



Apply LRGBCombiantion tool protecting RGB image with the starmask generated over Luminance.



The background has chrominance noise. To reduce it, using ATrousWaveletTransform I need a mask to protect galaxy and stars.
To generate the mask apply ATWT over the clone of the luminance that I had generated before, to smooth the image and then use RangeSelection tool on it.




Protecting galaxy and stars apply ATWT, having chrominance as target



A little movement of Curves to decrease background luminance and adjust color



Using the previous RangeMask to protect background, UnsharpMask to the galaxy.



Noise reduction in the background protecting the galaxy with the clone of the luminance that was made before.



HT to clip shadows (same mask)



And finally resample



Final image



Saludos, Alejandro.

Offline Jim_

  • Newcomer
  • Posts: 30
    • View Profile
Re: NGC 7479 - PI vs. Photoshop
« Reply #4 on: 2012 November 19 13:25:10 »
Alejandro,
What a great tutorial!!!
Thanks, this gave me some great pointers and ideas to try   :P
 
Jim

Offline Herbert_W

  • PixInsight Addict
  • ***
  • Posts: 202
    • View Profile
    • Skypixels - Astrophotography by Herbert Walter
Re: NGC 7479 - PI vs. Photoshop
« Reply #5 on: 2012 November 19 23:31:26 »
Hi Alejandro,
the next great tutorial - thanks for sharing with the forum.

Question:
When and how did you stretch the RGB?
What is your method to equalize the lightness from RGB and LUM - LinearFit or just by eye?

Thank you and best regrads.
Herbert, Austria

« Last Edit: 2012 November 20 00:09:16 by Herbert_W »

Offline Alejandro Tombolini

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1267
    • View Profile
    • Próxima Sur
Re: NGC 7479 - PI vs. Photoshop
« Reply #6 on: 2012 November 20 08:51:06 »
Hi Herbert,

When and how did you stretch the RGB?
I didn't. I do not know if it is correct (remember that I am Canon guy, and I do not have experience with L - RGB combination). As I saw too much noise in the RGB image only tried to see what happened if combining it without stretching, and the result was satisfactory. 8)

What is your method to equalize the lightness from RGB and LUM - LinearFit or just by eye?
just by eye :)
 
Regards,
Alejandro.