M13 (maybe work in progress?)

User images created with StarTools.
dx_ron
Posts: 246
Joined: Fri Apr 16, 2021 3:55 pm

M13 (maybe work in progress?)

Post by dx_ron »

[edited with an updated version]
Only the clouds, moon and wildfire smoke know for sure if I will add more time to this.

I imaged M13 last year with the RC6, and wanted to do it again this year with the 130mm f/7 refractor. This is 363x60s at low-conversion gain with a Risingcan IMX571C camera (equivalent to gain 0 for a ZWO ASI2600mc, I believe) and an Astronomik L3 filter. The original version (bottom of this post) was affected by some blue bloat (different uv-ir filter) and then I fought through what proved to be some wonky calibrations + channel alignment issues.

This version was processed at full resolution using Crop, Wipe, Optidev with a fairly small RoI (maybe ~ 1/3 of the cluster), a touch of HDR (gamma tame 1.05), SVD, Color (see Mike's suggestions on page 3 of the thread), Bin to 71%, Noise Reduction, finally a bit of gamma boost in FilmDev. Then resized to 1600 pixels wide and saved to jpg in IrfanView64.
M13_L3_363x60s_1600_500k.jpg
M13_L3_363x60s_1600_500k.jpg (480.32 KiB) Viewed 1481 times


Original 1st post:
Rather minimal processing: Crop, Wipe, Optidev, SVD, Color, NR. I kept the image at native resolution, to keep the smallest stars as round as possible.

In SVD I ended up using the synthetic PSF, mostly to save time, as I kept re-doing the processing (some of that due to ST crashing - only happened with the full-resolution image so I'd guess some memory issue/limitation in my computer). But I think with the refractor the synthetic psf is quite adequate. In acquisition there were about 65 maxed-out pixels per sub, and only two stars were clearly saturated - blue HIP 81673 and red HIP 81848. They gave me fits due to SVD ringing. I ended up with Linearity and Fuzz both down to 5%.

I will wait for Mike to tell me how I've gotten the color balance wrong ;) (another reason to call this a work in progress...). While Mike is pixel peeping, he will probably notice just a touch of pepsi-logo-ing on some stars. If I do collect more data I will try culling based on altitude to see that is reduced.
M13_436x60s_minimal-process_500k.jpg
M13_436x60s_minimal-process_500k.jpg (488.07 KiB) Viewed 1985 times
https://astrob.in/a7v4og/0/ astrobin link, so Mike can pixel-peep
Last edited by dx_ron on Tue Jun 20, 2023 3:08 am, edited 1 time in total.
decay
Posts: 433
Joined: Sat Apr 10, 2021 12:28 pm
Location: Germany, NRW

Re: M13 (maybe work in progress?)

Post by decay »

Oh Ron … I dimmed ambient illumination, disabled the ECO mode of my display and dialed brightness to maximum. And taking a very close look, I think the background might be a tiny bit uneven. So we all know, what Mike will tell you ... ;-) hehe

On topic: Nice shot! It’s so easy to burn out the bright stars in the core of M13, well done! And the seeing must have been pretty good? The processing with full resolution seems to have worked well in your case. I recently tried this as well (with my M57 that I posted in the other thread), but somehow NR did not work. Maybe too much of correlated noise messing up the processing ...

Best regards, Dietmar.
dx_ron
Posts: 246
Joined: Fri Apr 16, 2021 3:55 pm

Re: M13 (maybe work in progress?)

Post by dx_ron »

Thanks, Dietmar! To be honest, noise reduction didn't do all that terribly much - maybe because I did so few operations that it didn't have a good base from which to operate.

One big question is to HDR or not HDR? All M13s on CN are judged by "the Propeller", which is made more obvious by using the HDR module. But I tend to like having the core of the cluster teeter right on the brink of being totally blown out.
Mike in Rancho
Posts: 1141
Joined: Sun Jun 20, 2021 10:05 pm
Location: Alta Loma, CA

Re: M13 (maybe work in progress?)

Post by Mike in Rancho »

Golly. If you're good with it of course that's all that matters, I don't need to suggest anything.

Elsewise, plenty of propeller visible to me. Also that blue HIP star has a tiny close companion (maybe just line-of-sight) that is looking well-separated. Neat. Almost like Jupiter and a little moon.
dx_ron
Posts: 246
Joined: Fri Apr 16, 2021 3:55 pm

Re: M13 (maybe work in progress?)

Post by dx_ron »

I look forward to your comments because I appreciate your attention to detail in matters of color and sharpness.

Looks like that dim star is just LoS - about 19th magnitude? It's not in Stellarium, and I've not had a lot of success lately with plate-solving and annotation in ASTAP. Maybe I just forgot how I had previously succeeded. It will solve the unstretched image and then do a rudimentary stretch, which is how I got the 19th mag figure. But that's not useful for labeling the dim distance galaxies on the processed version.
Mike in Rancho
Posts: 1141
Joined: Sun Jun 20, 2021 10:05 pm
Location: Alta Loma, CA

Re: M13 (maybe work in progress?)

Post by Mike in Rancho »

dx_ron wrote: Fri Jun 02, 2023 7:10 pm I look forward to your comments because I appreciate your attention to detail in matters of color and sharpness.

Looks like that dim star is just LoS - about 19th magnitude? It's not in Stellarium, and I've not had a lot of success lately with plate-solving and annotation in ASTAP. Maybe I just forgot how I had previously succeeded. It will solve the unstretched image and then do a rudimentary stretch, which is how I got the 19th mag figure. But that's not useful for labeling the dim distance galaxies on the processed version.
Ok thanks. If you're sure! I saw the emoticons and all, but still wondered...maybe they just want me to shut up already. :? :lol:

I actually didn't find the background too bad at all. And against a bright beige-yellow background, like CN's, I'd say it would look quite dark and clean. I also can't find much in the way of Pepsi stars, so if they are there they have to be pretty subtle.

The only thing I would have mentioned is the blue stars being purple, something I believe you have raised yourself in prior postings. Now, since you undoubtedly used star or field sampling as a starting point for the channel balance, I'm a bit stumped as to any underlying cause. It also seems like it might be rather difficult to eradicate, without knowing more. I did download the jpg, converted to tiff, and opened non-linear no-tracking in ST. Usual purple fixes like altering red and blue decrease, or just adding some capped green, did not seem to work well. Or if I got the purple field stars more like the shade of blue I typically see, the blue glob stars ending up looking funny. Quite a conundrum. :think:

Maybe with more info...oh good grief nevermind, I didn't see the a-bin link down there even though it called me out. :oops:

BRB :lol:

Hmm, not as much help as I thought. If I hover over your IMX571 under equipment it says mono, but I see no filters listed, and the integration specs look more like OSC. Which if it is, maybe means those various UV-IR cuts you were experimenting with?

:confusion-shrug:

Only longshot I can think of is maybe check on Risingcam's B and R scaling in the driver. You've probably seen this referenced before for ZWO cams, where the desired setting is 50 and 50, which can be thought of as no scaling (and thus no potential quantization problems from multiplying the levels, though there might be a stronger green cast in the finished stack. But hey, that's what Wipe and Color in ST are for!).

In the meanwhile, the quickie band-aid in ST would be a post-tracking purple cap (save first if you want to be safe), accomplished by going into Layer and selecting Invert Foreground and Keep, go into Color and cap green to 100%, and then go into Layer again for another Invert Foreground. Also described in the User Notes under Special Techniques.

Oh and the tiny star pretending to orbit HIP 81673 pops up for me in Stellarium as mag 15.80, and looks even better in the a-bin version. :D
Color index for it is 0.0 though it looks a little pink-red here, but that probably clears up too with the purple cap?
decay
Posts: 433
Joined: Sat Apr 10, 2021 12:28 pm
Location: Germany, NRW

Re: M13 (maybe work in progress?)

Post by decay »

Mike in Rancho wrote: Sat Jun 03, 2023 4:27 am Ok thanks. If you're sure! I saw the emoticons and all, but still wondered...maybe they just want me to shut up already.
In my reply the comment regarding the background was only meant as a joke, Mike! It has obviously gone wrong. I also appreciate all your input here! It is very valuable to keep the discussions going and to keep on learning.

Best regards, Dietmar.
dx_ron
Posts: 246
Joined: Fri Apr 16, 2021 3:55 pm

Re: M13 (maybe work in progress?)

Post by dx_ron »

Thanks, Mike :thumbsup:

If you're extra bored, here's a link to the M13 stack. I'll leave it up for a few days: https://www.dropbox.com/s/termzzjwjx7bm ... s.fit?dl=0

I don't *think* The Color Purple is something in the camera driver (yes, the OSC version - I'll have to check abin's nicely complex equipment system to see how they've differentiated that). Unless some change in the driver itself coincided with when I switched from the AT65EDQ to the AT130EDT. Here's a couple of shots with the quad:
M31_KY_2022_OSC_365x60s.jpg
M31_KY_2022_OSC_365x60s.jpg (499.48 KiB) Viewed 1798 times
Barnards_E_10-28-22_30x120s_500k.jpg
Barnards_E_10-28-22_30x120s_500k.jpg (468.38 KiB) Viewed 1798 times
So my current working hypothesis is that it is something to do with the color correction of lens itself. I see the same colors whether I use the 1.0x flattener (like this M13) or the 0.8x reducer.

I should do some more experimenting with the L3 uv/ir filter now that I have the light leak squashed. What I have done with it leads me to what feels like such a weird starting point for color balancing:
color.jpg
color.jpg (453.53 KiB) Viewed 1798 times
But it doesn't seem to impact star sizes much, if any.


[edit} Hmmm, I haven't updated Stellarium in forever. I shall do so, and figure out how to add the coolest bestest star catalog
Mike in Rancho
Posts: 1141
Joined: Sun Jun 20, 2021 10:05 pm
Location: Alta Loma, CA

Re: M13 (maybe work in progress?)

Post by Mike in Rancho »

decay wrote: Sat Jun 03, 2023 6:59 am In my reply the comment regarding the background was only meant as a joke, Mike! It has obviously gone wrong. I also appreciate all your input here! It is very valuable to keep the discussions going and to keep on learning.

Best regards, Dietmar.
No you're all good Dietmar! I figured it was in jest, but still wanted to double check overall as "pixel peeping" can be derogative, and maybe Ron didn't need me pointing out things he already knows, or dismisses as deminimus.
Mike in Rancho
Posts: 1141
Joined: Sun Jun 20, 2021 10:05 pm
Location: Alta Loma, CA

Re: M13 (maybe work in progress?)

Post by Mike in Rancho »

Gosh Ron, the struggle is real. :?

I did a quick check on indi-kstars (your program, I think) and there should be some settings it calls WB 1 and WB 2. But I have no idea what they mean or what the neutral configuration is. Plus as you correctly note, the two images from the other scope have the shade of blue that we are all used to.

I ran through it several times when I had time today, at 50% bin (way too square for SVD), 71%, and full size. I tried both an aggressive normal Wipe with several clicks of DAF, as well as UnCal1 to really try to knock out color error. I sampled full screen and auto star mask (glob deleted), yet still ended up with more or less the same starting point as your M106 above -- blue zeroed out and red + green decrease. Now, with a few adjustments and green cap, I could get to near-regular blue for most of the stars, but a few larger ones were still afflicted with some purple. Highlight repair helps, but removes too much color from the globular core stars so is a hard compromise. Fiddling with Style and LRGB Emulation can help too, but really shouldn't be necessary for a normal broadband image I think.

I also did quick background neutralizations and autostretches in Siril and PI (I don't know how to do much else in them) and noticed the same purple casting to the stars at issue.

Finally I just split the channels in PI and ran a new star alignment and then composed as RGB in ST, but that didn't seem to help much either. :(

So...could something be going amiss in debayering and stacking? The other thing that sends me down that path are those displaced semi-circular blue artifacts around several stars. They look like patches that don't belong, not even a matter of CA or dispersion, I don't think.

By loading each split channel into ST as mono, the blue and green channels looked fine (though I would posit the blue a bit fuzzy and bloated). The blue artifacts, oddly enough, are in the red channel. Like a dead zone, through which the blue must be dominating, causing the final color.

You could see that quickly also by just loading in Siril, doing a background neutralization and autostretch, and clicking between the RGB channels.

That doesn't make sense to me as an optical artifact. Unless...when did the light leak get fixed in relation to this stack?

Kind of a shame, because other than the undue purple proclivity, and those blue arches, this glob data is really quite nice.
Post Reply