Star artifacts - optical or processing

[ajh499]

OK, that's a bit better. Following Mike's suggested steps results in this:

mikeCrop.jpg (189.55 KiB) Viewed 2222 times

At least I know that it's an optical problem and that I can process my way around it if I have to. I expect it was a little cloudy, I'll check the optics when I can.

As a full image I get this

mike.jpg (548.98 KiB) Viewed 2222 times

Not too bad for only 50 mins of exposure in RGB.


On that note though, most of the exposure time was in the Ha, which isn't used here. I wanted to use it as Lum in "L, RGB" mode, but I end up with grey stars as shown in my first post. Used in "L + synthetic L from RGB, RGB" mode, I'm sure it will blend fine and processed just like above I'm sure it'll be OK, but why the grey stars?

If the Ha is providing lum, and the G and B is smeared out, how can I end up with no colour in some stars (and artifacts in others, but that's dealt with)?


Alex

[Mike in Rancho]

Hey Alex,

Not sure on the gray halos. I've noticed that too in some of the various composites I tried. Likely some kind of pixel thing, and that stuff can get weird when you mix unmatched color and luminance. Pretty sure Ivo has cautioned on that. And not too long ago Freddy did a mono Ha blended with color for a Soul nebula. Interesting. If the L isn't there to support it, the color will just vanish. Not quite sure what happens the other way around. But yes I would imagine the synth L could ameliorate that somewhat?

You can also see that kind of thing in action maybe too. If you've loaded all your data files into L, RGB, just change the compositing from L+Synth L, RGB to L, RGB, and you'll see a bunch of stuff immediately disappear.

On a few tries with your data, a non-shrinking Shrink with high deringing and no color tame can blend out many of those gray halos, but not necessarily on the largest stars.

Here's maybe my best version. Or at least best looking maybe? Though with all the compositing I'm not sure what it shows. I think still something, so hopefully not too much "art."

It's pretty similar to Ha as L, but has more red. What I did was first blend Ha and R to make newR. Then I composited with Ha as L, then newR,G,B, as L, RGB. So without the synth L that is cutting off some of the B and G if the L (Ha) isn't there to support it. This does lead to some blown out areas, but I made adjustments in Contrast and HDR to bring that back down as much as I could. Some of the nebula edges are still way white, but not super terrible I think.

In any event, this does allow for reveal of the more far flung and nifty nebula clouds. When I was done, I still wanted a little more normal hue and pop to the stars than mostly white, so with a modified star mask I did a color-of-fg Layer blend using a prior, resolution matching, version that I had done up in legacy RGB, RGB with only the original RGB files.

Obviously more time could be spent perfecting it, but...proof of concept right?

alex M16 Ha as L Ha+R GB ST8 1A and RGB stars2.jpg (507.07 KiB) Viewed 2212 times

[ajh499]

Mike, this is my attempt at a straight "L, RGB" image from yesterday before you had posted your much more detailed attempt

EagleNebula-Forum.jpg (283.84 KiB) Viewed 2206 times

There was no attempt here to deal with the grey stars, but I did try to blend out the odd coloured artifacts the way you did on the RGB image.


There's obviously something I'm missing about LRGB compositing. The Ha (as Lum) is sharper than each of the R, G and B images, surely then there should be colour "available" for each pixel of a star. Maybe the colour would "bleed" into the background away from the point of the star, but the star should always have colour.

I think I have some reading and experimentation to do...


Alex