NGC6946 fireworks

[Datalord]

I have about 16 hours of LRGB on this one. Before I call it the end on this one, I would like to hear some feedback. Yay or Nay? Things to improve or redo?

[wimvb]

Very nice. Is this with the remote rig?

As for yay or nay; Yay with a but: a bit too colourfull for my taste. But that's just me.

[Datalord]

2 minutes ago, wimvb said:

Is this with the remote rig?

yep. I can't get this close with the RASA. 🙂

3 minutes ago, wimvb said:

bit too colourfull for my taste. But that's just me.

Fair point. I did give it an extra notch in the end and might have pushed it too far. This is before the last curves transformation (and thus without the last sharpening and final touches...)

[ollypenrice]

It looks a bit colour-warm to my eye, suggesting the reds are a little high. I think a cooler background would bring the galaxy forward.

The big challenge with this galaxy is the control of the stars. It's in a dense and bright starfield. My own approach in Ps was to blend two stretches, one with smaller stars. It's quite time consuming, though, because those over the galaxy need to be removed carefully first. I reckon this is the second hardest galaxy I've imaged, second only to IC342 and for the same reason - the starfield.

Olly

[Datalord]

3 hours ago, ollypenrice said:

I think a cooler background would bring the galaxy forward.

I'm always having a hard time figuring out the balance on the background. I'm not a fan of the black-black-black blackness, so maybe I was on the opposite end of the spectrum now.

3 hours ago, ollypenrice said:

My own approach in Ps was to blend two stretches, one with smaller stars

I've dodged this for a while, as I have a hard time figuring out how to do it, and as you say, it is very time consuming. But I probably have to sink my teeth into this process...

[ollypenrice]

32 minutes ago, Datalord said:

I'm always having a hard time figuring out the balance on the background. I'm not a fan of the black-black-black blackness, so maybe I was on the opposite end of the spectrum now.

I've dodged this for a while, as I have a hard time figuring out how to do it, and as you say, it is very time consuming. But I probably have to sink my teeth into this process...

Ps and other programs will allow you to measure the background brightness per colour and PI routines aim to set the background values to parity, so the same brightness per colour. I generally go for parity myself but not necessarily all that dark. Some people like red to be a tad higher, some blue.

Background and stars away from the galaxy are quite quick and easy: you just do a curve from scratch which lifts strongly at the bottom and then turns to a straight line fairly early. A few similar stretches will get your background to the values of the original image (the background must be identical in RGB). The galaxy and any faint fuzzies will be feeble but that's fine, and the stars will be small. I paste the 'star stretch' over the original and then erase it where it is masking the main galaxy and any faint fuzzies.

If you have removed the stars from the main image galaxy you can use the 'star stretch' again as a top layer in blend mode lighten and it will put the small stars onto the starless galaxy painlessly.  (Or painlessly-ish! )

Olly

[Datalord]

So, making a starfield, easy.

I did this with a starmask and an iterative stretch and erosion of stars, probably done 5 times. For the sake of learning the process, I did it quickly.

Now, my problem comes when stretching the rest. First of all, how do I stretch the galaxy without blowing up the halos of the stars in the main picture? And if I delete the stars and the halos before stretching, the smaller stars here won't cover those holes?

[wimvb]

http://pixinsight.com/tutorials/NGC7023-HDR/index.html#High_Contrast_Small_Scale_Structures

This is the PixInsight version of star replacement.

You create two images, one using MS which has the smaller stars, the other to reveal the galaxy. Create a starmask. Dial down the whitepoint, so what was white (1.0) is now at or below 0.25. Apply mask to galaxy image. Then use pixelmath to replace original with ms stretched version. Do this in several iterations.

[Datalord]

10 hours ago, wimvb said:

This is the PixInsight version of star replacement.

Spent too much of the day grappling with it, but it finally clicked in my head. Below is a comparison, which is not apples to apples, because I didn't pay the same care to every step on the left one, but I hope the difference is visible. It's going to require quite a lot of tries to figure out how exactly to land the two first stretches correctly in order to achieve the best result:

Anyways, there is no way around that I have to figure it out. At this focal length the stars will always become too large.

[Gina]

The one on the left looks more "natural".  The right hand one seems "over cooked" to me, saturated colours and a purple background.  It might be possible to "lift" the left one slightly but care would be needed not to overdo it.  Nicely captured anyway.

[Datalord]

Once again, salute to the forum. This time especially @wimvb who pointed me to the PI specific solution. I'm pretty happy with this edition.