StarTools

The July BSDI Challenge on CN was the Cygnus Loop. I participated, though had limited time on target, and really it was more about experimenting with new equipment. Like my 7x2" EFW and some filters. It was also my first real try at a Mosiac, both in sequencing NINA and stitching it all together.

All that did not necessarily go as planned.

In any event, in order to fit both sides of the Veil as best I could, I only picked a 10% panel overlap in NINA. That may have led to trouble down the road. As did the fact that I perhaps haven't properly dialed in the new backspacing to take the EFW into account. Mathematically I should have, anyway, as I set b/s to 55.7mm, being longer by about 1/3 of the 2mm filter thickness. But, the periphery stars do not look fully corrected, and of course that's right where panels need to be aligned and stitched.

Integration total was 3.75 hours, including both panels and all SHO channels.

6x300s P1 (West) and 5x300s P2 (East) of Ha.
9x300s P1 and 9x300s P2 of "OIII" (Created by putting the L-eNhance on the end of the CC and moving the EFW to Green).
8x300s P1 and 8x300s P2 of SII.

Various iterations in different orders of stacking, cropping, registering, and stitching in ASTAP, Siril, and DSS until I finally got something halfway workable. Alas, there was a seam issue in the SII, which mostly vanished due to binning, downsampling, and the chosen stretch, and the star registration/alignment was flawed in a thin strip up the middle. Luckily, only two small stars were greatly affected, and I used heal on the misaligned bits so as to not create a new double star. The offset tiny stars, however, were too numerous to do much about. I did an auto-star mask on red and just desaturated them to reduce their presence.

For color mapping I kept to a straight-up SHO = full primary RGB, no blend from the matrix. Not a normal thing, but I think this unique target is amenable to seeing some bright red, green, and blue. The channels were first cropped, Wiped, and resaved separately before composing, as the SII and OIII were of course much noisier, and the OIII picked up a lot of NP gradient.

Turning this into a mosaic and squeezing it down into postable form does lose some fine detail. I think these might look okay as independent West and East Veils processed separately. At least with more integration and perhaps fixing my backspacing.


This link goes to the CN Gallery which should pop up a bigger (~1550x1200) jpg at 100% quality compression level.
https://www.cloudynights.com/uploads/ga ... 696553.jpg

Not great, but maybe passable considering all the firsts. I have much to work on.

Mike,
I’m not familiar with this Northern Hemisphere target but nevertheless it’s a brilliant image
Well done !!
Gauging from your stacking and processing descriptions it looked like a difficult time consuming job to produce this image.
Colours and detail are superb
Cheers
Martin

Thanks Martin!

It's a fun target and for sure I want to do it right, one of these days. But all things considered I think this turned out better than I should have expected. F/4 probably helped, as did the QE of the 2600.

I only picked up the brightest parts in such short time. There's a lot of extra stuff in between the West and East Veils, invisible or just a hint in my image. Oh and it extends down more also, but a 2-panel try at mosaic was more than enough for me.

Typically this target is shown in HOO, but the SII filter picks up a decent amount also, allowing for some fun SHO possibilities.

Hi Mike,

that looks awesome! I did a mosaic only once and for sure it's not so easy. Given mine was only OSC your attempt is very impressive since you had 2 x 3 filters

Was every stitching run in ASTAP successfull or failed it from time to time? If so, what were the parameters of your successfull runs? I had some trouble with this...

Regards
Stefan

Thanks Stefan!

Hopefully this is just a taste of a better Veil mosaic in the future. But it was a lot of fun, a lot of learning, and a whole lot of frustration.

I just went to the ASTAP website and followed Han's mosaic steps one by one, except for the cropping/saving part. I took the (giant!) output FITS file and used Siril to crop it. Once I learned how to that in Siril, that is. Anyway, it seemed to me that Siril's crop allowed much finer adjustments to the cropping rectangle. Important because I wanted all the channels to be the same size before then registering them (in case I used DSS, which I eventually did).

I believe I had it on astrometric alignment, use FITS solution (easy since the stacking was also in ASTAP), equalize background, no oversize and no crop, no merge overlap.

Ultimately though I ended up trying nearly all the different options...

Anywho, things seemed good at first with my Ha and "OIII" stitching, as well as an HOO process of it in ST. But I admit I may not have looked carefully.

It was when I added the SII that things started going sideways. For some reason, it stitched the two panels together at a slightly different angle than it had for Ha and OIII. Of course I didn't realize that at first. Registration was a mess. So I tried registering in DSS, forced bilinear alignment, and while that was better, I ended up with some serious warpage. That's when I found the mismatched stitching. So I tried all the various different options in ASTAP, to no avail, including even star alignment instead of astrometric, and astrometric ignoring the FITS header and solving anew.

What finally worked (mostly) was this: After stacking all 6 of my panels in ASTAP, I used DSS to register the panels separately (ASTAP would have worked for this too I think), giving me all Panel 1's aligned to each other, and all Panel 2's aligned to each other. Then I went through the ASTAP mosaic stitching for each SHO channel, and yay! they all stitched at the same angle. Then again I used Siril to crop the giant mosiac files to matching xy resolution. Finally, I used DSS to register the three mosiac channels (i.e. standard mode, run registration to get lots of stars for aligning - about 150 on my lowest channel, and select save calibrated/aligned intermediate files. I deleted the autosave and used the saved intermediate aligned files.

Well, that was close, but not perfect, as I described. One seam in the SII, and a thin strip of misaligned stars, fortunately not the entire overlap.

I tried several times before finally surrendering. In the future, hopefully a better coma corrector backspacing and/or larger panel overlap will fix those little issues.

Hi Mike,

thanks for outlining your workflow in such a detailed way!

Mike in Rancho wrote: ↑Mon Aug 01, 2022 6:57 am I believe I had it on astrometric alignment, use FITS solution (easy since the stacking was also in ASTAP), equalize background, no oversize and no crop, no merge overlap.

Interesting...I've only been successfull with no background equalization but with background merging Probably makes sense to test different options /combinations for individual data sets.

Mike in Rancho wrote: ↑Mon Aug 01, 2022 6:57 am I took the (giant!) output FITS file and used Siril to crop it. Once I learned how to that in Siril, that is. Anyway, it seemed to me that Siril's crop allowed much finer adjustments to the cropping rectangle.

Did you also have these huge totally black areas around your mosaic result? I'd have cropped most of the black in order to reduce the file size and then crop carefully in a different software.

Mike in Rancho wrote: ↑Mon Aug 01, 2022 6:57 am It was when I added the SII that things started going sideways. For some reason, it stitched the two panels together at a slightly different angle than it had for Ha and OIII. [...]

[...] Then again I used Siril to crop the giant mosiac files to matching xy resolution. Finally, I used DSS to register the three mosiac channels (i.e. standard mode, run registration to get lots of stars for aligning - about 150 on my lowest channel, and select save calibrated/aligned intermediate files. I deleted the autosave and used the saved intermediate aligned files.

Oh boy, that sounds like torture, especially with thus many files etc. I'll stick to OSC at the moment

Mike in Rancho wrote: ↑Mon Aug 01, 2022 6:57 am In the future, hopefully a better coma corrector backspacing and/or larger panel overlap will fix those little issues.

A bigger overlap will help for sure. In daylight photography at least 20% are recommended I think.

One little thing on your color result: Have you tried to invert the colors and to reduce the magenta/pink stars by capping green to a certain degree? I know that these stars are typical for SHO images and some people like them but I think they are more appealing if the pink isn't too dominant.

Regards
Stefan

Hi Stefan!

Yes, totally huge black space around the stitched panels, and a 1.2 GB file size resulting therefrom. I guess I could have pre-cropped in ASTAP roughly and saved as FITS, then used Siril for more precise cropping down to the pixel level. Unless ASTAP can also precision adjust crops and I just didn't figure it out? But Siril seemed to easily allow dragging each side of the cropping selection rectangle independently, as opposed to just dragging a rectangle and hoping for the best. Though I did have to use the zoom-in buttons in order to get finer pixel gradations.

I'm actually good with SHO colored stars and that their colors reflect the wavelengths passed by the filters. That said, my little tricks to minimize the appearance of the one strip of overlap errors helped reduce some overall red, and I kind of did that generally too. My channel bias balance in order to give a bit more strength to the SII parts of the Veil also had affected the stars and background, of course. That just neutralized the stars by a small amount, which I guess could also be done in Shrink settings or with Highlight Repair in Color, though I think I was already near maxed out there. However, I didn't include big stars like Cyg 52, which looked totally cool retaining the pink.