Cocoon Nebula Processing (from Cloudy Night Discussion)

[gminder]

Hi Ivo,

The data for the Cocoon Nebula image I referenced at CN can be found at https://www.dropbox.com/s/727mhgniargrr ... reProc.fit

I just downloaded the latest ST beta and will take another shot at it. Images like this with lots of dark dust clouds are hard for me.

I really wonder what you think about the LLRGB method. If you haven't seen Scott Rosen's video, it's at http://www.astronomersdoitinthedark.com ... torial.php

[admin]

Thanks for letting me test the latest beta with your data Gary!
It really is deep data and it contains a good bit of the interstellar medium.

Note is used the 1.3.5.249 beta (just uploaded) and here is what I did;
As always, AutoDev to see what we got.
I see lots of stacking artifacts (to be expected), green bias, oversampling, coma or rotation in the corners.
I decide to bin to 50% to make the most of the oversampling.
I next crop the image to get rid of any stacking artifacts and to frame the nebula.
A wipe with default values took care of the green bias.
Next, I did my final global stretch using AutoDev. I actually rarely use the manual Develop anymore! I created a ROI square around the nebula. And specified that AutoDev should only assign 25% importance to the area outside the ROI. I also set 'Ignore Fine Detail <' to 4 pixels. As I said, I'm keen to preserve some detail in the interstellar medium, but you might want to decide otherwise.
I tried my luck with Decon (Auto-generated mask). A radius of 2.1 tightened the stars a little, but not much further detail was recovered (perhaps a tiny bit better definition in the Cocoon).
I Next used the same mask with the improved Sharp module. It now comes with an 'Intelligent Enhance' option (on by default), which uses the Tracking information to only enhance areas that it thinks contain enhanceable detail. It works quite well most of the time. I used default settings but bumped up the 'Amount' to 250%.
As one of the last few steps, I used the Color module. If you haven't used the 1.3.5 betas before you'll be pleasantly surprised with the improvements. Just by launching the module, StarTools manages to often get quite close to a good color balance. So too in this case, although the Saturation was a bit much - I reduced it to 150%. I also set 'Cap green' to 'yellow' in order to remove any gree color information.
We could use the Life module's Isolate preset here to push back the noise, but I want to show you something new. I'm going to try and keep the interstellar medium gas, as I think it's a great feature of your data.
A bleeding edge new feature in the 1.3.249 beta is called Read Noise Compensation and it helps in the darker areas where a noise floor makes noise development non-linear (e.g. areas where noise vs signal is no longer considered 1:1 due to noise influences other than those that come with the recorded signal, for example read noise which is *always* present).
To find a good setting for Read Noise Compensation, I create a preview area that has both background and DSO in it, turn off noise reduction completely (set Brightness Detail Loss to 0%) and bring up Read Noise Compensation until it almost starts affecting the DSO. Then increase Brightness Detail Loss to denoise the rest of the DSO and tweak noise settings to taste as normal.

I used these parameters for your image;
Parameter [Scale 1] set to [90 %]
Parameter [Scale 2] set to [96 %]
Parameter [Scale 3] set to [96 %]
Parameter [Scale 4] set to [96 %]
Parameter [Scale 5] set to [49 %]
Parameter [Brightness Detail Loss] set to [21 %]
Parameter [Redistribution Kernel] set to [8.4 pixels]
Parameter [Read Noise Compensation] set to [19.01 %]

And this is the result;

IC 5146 Cocoon Nebula Neb PreProc.jpg (210.5 KiB) Viewed 5883 times

Hope this helps!

[gminder]

Thanks Ivo!

I tried again with the LLRGB workflow and posted a 2nd attempt at CN. I will go through your workflow above with my data and see what I can get. Your advice on working with the dust lanes should be really helpful.