My Tutorial Workflow for PixInsight Narrowband LRGB

[kayronjm]

Read here: http://lightvortexastronomy.blogspot.com/2013/02/tutorial-nebulae-pixinsight-workflow.html

This is my latest tutorial for PixInsight, producing my entire workflow for narrowband post-processing (S-II, H-a and O-III) to make an LRGB image. As usual, this tutorial workflow goes from the raw calibrated monochrome images to the final result with everything in-between. S-II and H-a are blended for Red, H-a and O-III are blended for Green and O-III is used purely for Blue. For Luminance, H-a data is used. Post-processing is very similar to my other HaRGB tutorial workflow but with distinct differences in colour-combining and noise reduction. This tutorial workflow is also applicable to anyone interested in pure LRGB (non-narrowband) image post-processing as besides colour-combining, the techniques very much apply. It is particularly useful for post-processing images of nebulae.

[Alejandro Tombolini]

Hi kayronjm,

Thank you for sharing it, very clear and you have achieved a nice result.

Saludos, Alejandro

[kayronjm]

Hi kayronjm,

Thank you for sharing it, very clear and you have achieved a nice result.

Saludos, Alejandro

Muchas gracias. Espero que ayude a alguien buscando algo por el estilo!

[niteman1946]

Hi kayronjm,

Thanks so much for this excellent tutorial/cookbook.  I followed it fairly closely and got very nice results.

A question which really does not have much to do with your tutorial is regarding the image color.  Overall I was pleased with the results but would have preferred a bit more red accent on some of the image's surfaces.  I tried multiple combinations of Ha, O3 and S2 during the PixelMath color combine, but this (mostly) blue-green is all I could get.

Do you think that there is a means of teasing more of the red into the image, without messing up the color balance?

Again, great work and thanks for your effort.

Mark

[niteman1946]

Well, I've answered my own question.  The second image has more of the color I was looking for.

Mark

[kayronjm]

Well, I've answered my own question.  The second image has more of the color I was looking for.

Mark

Hi Mark, many thanks for your comments, I'm glad the tutorial helped you out. My apologies for the late reply - I rarely check all the astronomy forums I'm part of!
How did you achieve your desired colour balance in the end? Did you boost the Red channel with Hydrogen-Alpha data? Your image is wonderful, by the way, congratulations!

[niteman1946]

Hi Kayronjm,

Thanks for the very kind words.  I really appreciate that.
The best I can tell the initial problem was a case of the DA. 
After revisiting my integrated narrow bands, I found the O3 (OIII) looked just a little too much like the Halpha. I reprocessed the O3 and combined this new version with the existing Halpha and S2 (SII).  Voila, there was red!

So, all is well.

A couple of questions if you don't mind,

1.  Do you see a hierarchy in the noise reduction scheme?  First ATWT, then MMT and then ACDNR? 
2.  Once you start saving as 64bit, should each subsequent save be also at 64bit?

Cieli Sereni,

Mark

[kayronjm]

Hi Kayronjm,

Thanks for the very kind words.  I really appreciate that.
The best I can tell the initial problem was a case of the DA. 
After revisiting my integrated narrow bands, I found the O3 (OIII) looked just a little too much like the Halpha. I reprocessed the O3 and combined this new version with the existing Halpha and S2 (SII).  Voila, there was red!

So, all is well.

A couple of questions if you don't mind,

1.  Do you see a hierarchy in the noise reduction scheme?  First ATWT, then MMT and then ACDNR? 
2.  Once you start saving as 64bit, should each subsequent save be also at 64bit?

Cieli Sereni,

Mark

Ahh I see, I'm glad it worked out - looking great!
No problem at all.

1. I personally do indeed use ATWT first as it provides the most effective noise reduction in my honest opinion, massively reducing small scale noise more-so than I've ever achieved with MMT or ACDNR on their own. ACDNR for example tends to blur it out and make noise blobs that look awful. ATWT seems to kill the small scale noise and then MMT can smooth it all out a little more. ACDNR can then be used to some small extent to ensure any remaining background noise is smoothed out. On my other workflow tutorial (HDR LRGB with Orion Nebula data) I did not use ATWT at all but only because I had not adopted it as a noise reduction technique just then. All worked out well in the end there with Orion Nebula but it really does depend on your data. I guess Rosette Nebula being dimmer than Orion Nebula meant more noise with the same number of subframes and exposure times, hence why ATWT was very helpful there.

2. When you first saved your image in 64-bit to start the non-linear stretch and further post-processing, subsequent saves of that image will automatically be in 64-bit format. If you are unsure, you can always make the initial save in 64-bit, close the image and re-open it (the 64-bit one). After the first 64-bit save, I keep it 64-bit all the way up to the very, very end result. I indeed keep the end result in 64-bit as well. I only go back down to 16-bit when I save my FITS end result from PixInsight into TIFF format for opening in Photoshop and adding a nice frame around the image with my name, object name, capture date and data used.

I hope that's helpful!

[mstriebeck]

Hi,

I am completely new to Pixinsight (I'm currently using CCDStack and Photoshop). As I am mostly doing narrowband imaging, I was very interested in this tutorial. But ran into a few challenges:
- when combining the three channels to a color image, is there any faster way to see the effect of changes? Right now, I have to adjust the ratios in the PixelMath tool and then wait until I see how it looks like. In Photoshop I have (almost) a WYSIWYG feature which makes this much easier to play with until it looks good
- when I try to apply the stretch with the HistogramTransformation tool, I don't get any color but just various gray images (with a slight green hue)

Any ideas how to get through this?

Thanks
MarkS

[niteman1946]

Hi Marks,
I'll take a stab at this and then let Karon pick up the pieces.  I've found Karon's process very useful and use it for both narrow band and RGB (maybe not technically correct, but it works for me).

1. The first things you do are calibrate, align and integrate each group into masters Ha, OIII and SII. 
2. Use the same crop for each to remove edge issues.
3. Make a copy of the Ha master and use it for luminance.
4. Conduct Dynamic Background Extraction on the two Ha's along with OIII and SII masters.  I have better luck when I do the DBE before combining the three colors.  Set aside the luminance Ha for later.
5. PixelMath combine the Ha, OIII and SII masters (not the Ha intended for luminance).  This will create a new image in color.  Keep your PixelMath window open and you can create several "what-ifs" if you feel like it, but all of your new images will be initially dark.

Now apply an autostretch (STF) to the new image(s) and they will show color.  At this point the colors are likely fairly garish.  Click on the "chain link" button on the upper left hand of the STF and the colors should take on a more familiar look.

The rest is just processing.  Hopefully I'm not just stating the obvious and this is of some help to you.

Good luck,

Mark

[Alfredo Beltran]

Thank you very much for this detailed tutorial kayron!

This gives me A LOT to think of on how I process my images.

It will be for sure of high usefulness.

Best regards

Alfredo