Long Eagle

[kosborn (Kevin)]

This is the most time I've spent on any one object. This was a total of nine hours - three hours (36 x 5 minute subs) each for Ha, OIII and SII. I'm happy with the data but not so happy with the processing, especially the full field. Although the field of view is wide with the Esprit100 I did a drizzle integration so the cropped version of the core would look better (higher resolution at https://www.astrobin.com/full/352725/B/?nc=user).

Any constructive criticism welcome - especially from any PixInsight gurus.

Kevin

[xelasnave]

Wonderful image.
Alex

[Christopher-26]

Both versions are really good! Great detail into the Pillars of Creation.

[Anth10 (Anthony M)]

Kevin,
A mountain of data has gone into this for a classic shot congrats, deserves applause.
Well done

[kosborn (Kevin)]

Thanks Alex, Christopher and Anthony,


I'm going to keep the data safe and reprocess in maybe 12 months time when I have more PI experience. I really don't like the muddy brown in the wide field image .


Kevin

[raymo]

Lovely Kevin, but I'm wondering why you have almost obliterated the
orange/red, of which there is plenty in M16. I notice the nebula has a
carry handle on the left, I suppose in case somebody wants to move
it somewhere else in the sky.
raymo

[kosborn (Kevin)]

Originally Posted by raymo

Lovely Kevin, but I'm wondering why you have almost obliterated the
orange/red, of which there is plenty in M16. I notice the nebula has a
carry handle on the left, I suppose in case somebody wants to move
it somewhere else in the sky.
raymo

Yeah I know - muddy brown instead of orange/gold. I need to go back and try again. The carry handle is so that I can pick it up and pass it on to someone who knows what they're doing!


Kevin

[Zubenel (Wes)]

Great Job. WOW !!! The question must be asked. When does one stop being a beginner

[raymo]

I have suggested to several members recently that it is time they moved
to the main forum, but most seem reluctant to do so, being I suppose,
intimidated by many of the images there, and maybe thinking [incorrectly]
that the advanced imagers wouldn't want to be bothered with them.
raymo

[kosborn (Kevin)]

Originally Posted by Zubenel

Great Job. WOW !!! The question must be asked. When does one stop being a beginner

Thanks Wes. I don't think I've graduated yet.

Originally Posted by raymo

I have suggested to several members recently that it is time they moved
to the main forum, but most seem reluctant to do so, being I suppose,
intimidated by many of the images there, and maybe thinking [incorrectly]
that the advanced imagers wouldn't want to be bothered with them.
raymo

I've lurked in the background on the forum that the big people post on. I would rather constructive criticism like your orange/red comment than someone saying "it's a tad soft". Maybe if I'm brave I'll dip my toe in the water with the next one...


Kevin

[cometcatcher (Kevin)]

Originally Posted by raymo

Lovely Kevin, but I'm wondering why you have almost obliterated the
orange/red, of which there is plenty in M16.
raymo

It's a narrow band SHO palette Ray. The red hydrogen we are so used to is mapped to green. The red channel is for SII and OIII to blue. I think. As such it looks about right to me. Great shot Kevin!

[Atmos (Colin)]

You have some really nice and sharp data there Kevin.
I assume that you're going for the more "traditional" SHO palette as done by Hubble? The one that doesn't have as much green and shows more of a fiery background?

I've attached a screen shot that shows what I would be doing for colour balance while it is in the linear stage. I did try it on your JPEG but it didn't work in the slightest I did notice that your background does have a blue hue to it, it isn't obvious mind you but it does show up in the histogram.

My only other suggestion would be to either not do as many iterations when running deconvolution OR to change your global dark/white so as to not get the "worms" on the edges of high contrast areas. If you look around the pillars it has a white rib, that is caused by the deconvolution.

Other than that, you've done an excellent job

[kosborn (Kevin)]

Originally Posted by cometcatcher

As such it looks about right to me. Great shot Kevin!

Thanks Kevin!

Originally Posted by Atmos

You have some really nice and sharp data there Kevin.
I assume that you're going for the more "traditional" SHO palette as done by Hubble? The one that doesn't have as much green and shows more of a fiery background?

I've attached a screen shot that shows what I would be doing for colour balance while it is in the linear stage. I did try it on your JPEG but it didn't work in the slightest I did notice that your background does have a blue hue to it, it isn't obvious mind you but it does show up in the histogram.

My only other suggestion would be to either not do as many iterations when running deconvolution OR to change your global dark/white so as to not get the "worms" on the edges of high contrast areas. If you look around the pillars it has a white rib, that is caused by the deconvolution.

Other than that, you've done an excellent job

Thanks for the comments Colin. I generated a synthetic luminance from the SHO data as suggested by RickS. I did 100 iterations of deconvolution (is that too many?) but then further sharpened luminance with a multiscale linear transform. Maybe that was where the worms came from...?


The (non drizzle) data is available at https://drive.google.com/file/d/1Ue0...ew?usp=sharing. If you or anyone else has the inclination I would really like to see what could be done with it.


Kevin

[Atmos (Colin)]

This is a 1:1 crop of the centre of your data. There is some JPEG compression from the TIFF I saved it as but it's pretty close.

My workflow was:
- Rotate images 180º
- 5 iterations of deconvolution on the three channels individually. You have to create a new PSF for each channel as star sharpness goes S > H > O with OIII having the largest PSF.
- LRGBCombination to create the SHO image
- Set Black and White areas as shown in my previous post
- Crop a 1200x900 image for further processing
- DGVDenoise
- Stretched via HistogramTransformation
- HDRMultiscaleTransformation
- CurvesTransformation for saturation boost
- ColorSaturation to reduce the purple
- SCNR to reduce some of the latent green cast

[kosborn (Kevin)]

That was quick! Nicely done. The pillars in your version are much clearer than mine. Did you generate a synthetic luminance or just use the SHO? I used a single deconvolution (with 100 iterations) on the luminance alone. I hadn't thought to deconvolve all the channels.


Kevin

[Atmos (Colin)]

I just used the SHO but I have only very recently started doing deconvolution on the individual channels as opposed to a synthetic lum. It is more work but I have found that it gives a better result.

I did a single deconvolution with 5 iterations. I have found that the only way that you can do more than 25 iterations is with heavy and very precise star masking and not using any deringing protection.
To give an example of this, check the link below
100 Iterations with kinda okay star masking


With this I opened your Ha file:
- Cropped a rough region I wanted to work on
- Upsized it 2x2
- Created a PSF
- Spent 5-10 minutes creating several star masks and using pixel math to add them together
- Masked the stars with the master star mask
- Ran 100 iterations on the Ha
- Stretched Ha and used it as a luminance on the previous SHO image from the last post. That's why the colours have been messed up a bit

You can see that there is a dramatic increase in resolution, a bit more could probably be done with drizzled data though.
- Spent a few

[xelasnave]

Wow Colin.
Great work.
Alex