StarTools

And the segment of Barnard's Loop in pure red makes a nice Christmas garland.


This is a work in progress I'd say, at least for processing. Acquisition for 2022 is done with. This holiday weekend has bee nice, clear, and over 80° to boot, but I've spend the time overeating and napping, and a week straight of rain through and past New Years is on tap.

I considered posting this in image troubleshooting as well, for questions on NB Accent.

Specifically - is there any "correct" level of narrowband accenting, or is it just a matter of imager's choice when blending NB into BB? Or rephrased, there's no correct way to do it but lot of ways to botch it up?

I had a prior version of this on CN in the monthly target progress thread, and was critiqued for "too much red." In review, I think there was something to that.

In NB Accent, it seems a matter of eyeballing both screens. But here, where I am accenting with a field filling nebula file (rather than small galactic knots), it can be hard to tell what is being altered. Though I was able to see it later after opening in Gimp and putting an RGB levels pointer to pixels. That would be a pain to save out iterations for double checking and then undo to try again if needed.

It also takes some doing (though I am getting better) in the NBA stretch screen to get a good accenting to pass through, avoiding black areas being "lit up" so that only structure gets added rather than, I'm not sure -- background levels or noise floor?

I suppose some extra integration with the Ha here could help separate signal.

In any event, in the image at hand there actually seems to be significant Ha strewn about the area, as seen in various reference images. Including some nice striated details below M78. Those are what I really wanted to show, but it was hard to do without, again, apparently bringing an Ha floor along with it. I think maybe a bit harder in that the NBA file rides along unprocessed, and doesn't get the benefit of Contrast, HDR, etc., and maybe particularly in this case Sharpen.

I was also wondering if Wipe sets any kind of floor for the NB Accent file, the way I think it does for RGB?

Anyway I had some massive LP + Christmas lights gradients here, but too much aggressiveness in Wipe was erasing much of the Ha file, which was mostly gradient free. So I cropped, binned, and separately wiped each file ahead of time for recomposition.

The LRGB was minimally stretched, in hopes of not overwhelming the dimmer Ha striations. I think I found a decent compromise here using NBA screen 1 tight ROI and a lot of IFD, plus lower shadow linearity. I tried to avoid much threshold usage as that can get choppy.

I used pure red for better visibility (and Christmas cheer) plus bumped the strength. I believe I avoided improperly raising the floor in the red channel where there shouldn't be any, and at least some of the Ha striations can be seen if the image is against a dark border.

I did need to reverse the order of SS and NB Accent from recommended, but didn't need a layer undo for star colors.

I did try separately processing LRGB and Ha, which did make for better nebula detail visibility, and then tried a lighten layer to bring them together. Meh. It worked for that aspect of it, but overall I think the NB Accent version came out much better.

Sill too red? Or not enough? I did check a few pixels in most black and gray spots, and the RGB levels are even, so I don't think I passed through too much of a global red floor this time.

As a more direct comparison, this was my earlier rendition. Composed from the original stacks, not separately wiped first. So, since the LRGB required a strong wipe, it seemed to me that wipe also ate away at a lot of the Ha stack.

But, whatever I did in NBA screen 1 passed along a fair bit of the Ha structure over on the right side, the only problem being that as discussed it seems I passed along too high of an Ha or red floor.

I like the first one best, a nice red Barnard loop.

Just whished i could get my IC434 like that....

Thanks Freddy. No red in your Horsey?

Well, going to try this again since the first time I hit logout during my preview without pressing submit.

I had another go at my M78, and I think I got a better balance of the Ha, the dark shadowy nebula, and whatever minimal dust I managed to pick up. Some details I thought could be artifacting, reflections, bad flats, but was able to match up to references so I think it's good. Just low SNR in some regions.

This was also my first try at stacking in PI, the trial of which I got a couple days ago. Was able to throw both night's sessions of all filters plus flats and my archive darks and flats into WBPP. With a little helping hand and of course, being PI, trying to fathom the unfathomable, it actually does a nice job of taking full directory loads and figuring out the proper calibration workflow to get me my 5 registered masterlights. Took a long time, even though I did a little PI tuning of my swap files.

I'll have to start another thread perhaps of my PI experience, and to garner some do's and don'ts of settings for PI pre-processing with ST post in mind.

I had already planned on buying it, which I will at the end of trial. As predicted it managed to star align those difficult JSWT files nicely. It's blink for reviewing subs is probably the best blink type module I've used of any program so far. And because it can stack everything to a single best reference sub, unlike ASTAP which forces you to register finished filter stacks afterwards, I didn't have to use any highlight repair for multi-colored stars.

Here was my restacked and reprocessed wide view of M78.


Then last night I did a deep crop of the composition in ST to see how that worked out. Better details in M78 itself, though I am pushing the limits of my resolution and the limited integration here I think.