Next project: bat and squid thoughts/planning

[STEVE333]

Hi Steve, I tried out MureDenoise and it worked really well! The SNRWeight value just shot up - it was the first thing I did.

This is done without any masks right? It seems to be smoothing out the nebula really nicely but I think slightly adding some noise in the background. Is it not meant to be used on the background also?

--Ram

Hi Ram - As far as I know, no masks are used. However, I'm no expert on this process, so, I wouldn't take my word for it.

Glad to hear it worked well for you. I was amazed at how well it smoothed out the low level nebulosity for my data. I have already added this process into my regular workflow for NB images. Because I always have limited data I always have low-level nebulosity issues. I'm glad to have found this new "tool".

Cheers

[ram]

I'm wondering what kind of Squid people like, these are four paths I can take. This a crop of the Squid and still have to do a lot of things, but this leads to different outcomes depending on how I go:

A. Starless processing, stars put back, overall a light treatment of the Squid.
B. Normal processing, without doing any starless, also a light treatment of the Squid.
C. More intense process, a pretty bright result, no starless.
D. Result of the Auto STF from PI after initial processing. It produces the brightest Squid but also the brightest background.

My daughter likes (C). I'm torn between (C) and (A). What Squid do you like to see?

Thanks in advance for letting me know your preferences...

--Ram

PS: I forgot to attach the images in my original post.

[STEVE333]

It's your data, so, your choice! Everybody has a different opinion, and, I doubt you could get any consensus. I look forward to seeing whatever choice you make. Just my opinion.

Steve

[ram]

It's your data, so, your choice! Everybody has a different opinion, and, I doubt you could get any consensus. I look forward to seeing whatever choice you make. Just my opinion.

Steve

Thanks Steve! Ultimately it is my choice but I'm fascinated by the sociology of it all, why do people prefer the things they do, etc. It goes back to the issue of everything that social media is based on, and I think it teaches us about people when they do respond in the way they do. Take your own noncomittal response, which I think is informative in its own right (this time---in the past you have sometimes indicated a preference also). Same with me: though I may indicate a preference, most of the time I'm like "up to you" when confronted with other people's choices. Nonetheless, I think making my choice after hearing all of the opinions (whether there is consensus or not) and then making my choice before is different - I may consider something I wouldn't have otherwise, etc.

Whenever there's an article in a newspaper like the NYT, etc. the first thing I do is read all the comments. I read a lot of the posts here and the comments herein and again, the responses sometimes are more interesting than the original post. Like I said, fascinated by the sociology of it all. Especially AP or anything involving artistic opinions to evaluate.

--Ram

[sdbodin]

I'm with you C and A.

Waiting, should be epic,
Steve

[STEVE333]

OK - a blend of B and C.

Steve

[Graeme1858]

C

Good bright squid without looking over processed. Natural looking stars. Natural looking dark background.

Regards

Graeme

[ram]

Thank you all for those who responded. So I'm approaching the end of the mono image with the FC100DF at least that's only the Squid and I tried to take your viewpoints into account (I agreed with (C) and I tried to find a bit more subtle version and also this a lot cleaner, with NR, etc.). Here're my top two choices. The two are very similar except the first one has LocalHistogramEqualization done to it which brings out some internal contrast but also makes it look a bit weird and then I also reduced the star sizes which causes some stars to disappear to I always worry about changes like this. Nonetheless, the second one is just after NR and stretching, and the first is with everything else after. In the old days I'd stop at the second one but now I don't think there's a process in PI that goes by untouched eventually and that's reflected in the first image.

In terms of NR, I did MureDenoise, and Jon Rista's tutorial. I think that part came out pretty good. The SNRWeight of the second image is like 7000! But then it drops the moment I stretch it - hopefully that's a normal thing.

Also I've been comparing to the EZ Processing Suite available in PI - I think it's highly convenient and very useful. I also got an answer back from Juan of PI who explained to me why the SNR of the latest version 1.8.8-6 after ImageIntegration might drop a bit slightly. So that was reassuring. I'm still using 1.8.8-5 integration in the above example though.

Next step is taking an RGB sky background when the moon diminishes, integration of the Ha and S2 channels that's narrowfield, and then taking more widefield Ha and S2 and perhaps additional O3 as well (matching the background noise in all these frames will be fun). There's actually some very very faint nebulosity that's all around the Squid IMO. It's hard to say given how much stretching has to occur here. Most people just focus on the Squid region and ignore everything that comes around it but there's some nebulosity here and there. Nonetheless I think this is a case where that takes away from focussing on the Squid unless one gets a better scope or puts in a lot more time (like 4x as much).

--Ram

[STEVE333]

That's looking beautiful Ram -

Oftentimes after performing the first mild stretch on a nebula I will try creating a starless version. I subtract it from the original to create a Stars image. Then I do the LocalHistogramEqualization (LHE) to the starless version [with a Luminance mask applied] while keeping the Amount fairly low (0.3 - 0.5). I usually do LHE twice, once with Kernel Radius = 26 and then a second time with Kernel Radius = 110. The first application brings out the smaller details while the second pass enhances larger structures. Once LHE is complete on the starless image I recombine with the Stars.

Just a thought.

Steve

[ram]

Thanks Steve, will try it out and report back!

--Ram

[ram]

Hi Steve, here's the version that does your LHE trick but processed in the same way otherwise. I think it brings out the highlights a bit more and has a bit more contrast but the SNR* for this one is a teensy bit slower (less than one db). I also didn't do the HDR on this one to tame down the central star which I thought I had done a good job of taming anyway. Let me know which version you prefer, this one:

Or this one:

download/file.php?id=13758&mode=view

*I've been using the SNR or SNRWeight a lot to guide my choices esp. when I can't tell the differences visually. For this target at least I find it served me well at least in terms of making this decisions. In this case however, the SNR is very similar but there's some visual differences and I think I lean towards the newest one but both are good IMO.

--Ram

[STEVE333]

Thanks for trying the comparison Ram. I agree that the two results are very close from a visual inspection point of view. I slightly prefer your original version because of the better control of the bright star as well as a slightly "more natural" appearance.

Both look very nice for sure.

Really enjoying watching this project progress.

Steve

[ram]

Thanks Steve - without all your suggestions I wouldn't have come this far - this year has seen an exponential growth in my processing skills.

Anyways, as I mentioned, the HDRMT is the difference in how the central star is tamed and less brighter. Here's the same image with your LHE settings that has HDRMT applied to it (the previous one you liked had HDRMT too). What I like is that even without the HDRMT the star core in the centre is discernable if you look closely (i.e., it's not a big blob) but anyways, with the HDRMT the central star is tamed even more:

Now to get a dark night to get some RGB background shots and I'll be done with the first version of the c-orgb composition. I spent a week soaking my t- adapters in the penetrant spray that can loosen stuck bolts, etc. and I finally got it unscrewed today. Washing to remove to oil. Clean the camera sensor.

--Ram

[STEVE333]

Looking really nice Ram. You have captured a lot of details on this one.

Steve

[STEVE333]

Thanks Steve - without all your suggestions I wouldn't have come this far - this year has seen an exponential growth in my processing skills.
--Ram

Hi Ram - Glad that I have been able to provide some assistance in advancing your processing skills. I've learned so much from tutorials and helpful suggestions of forums. It's nice to be able to pass some help forward.

Best of success with your continued growth in processing.

Steve

[ram]

Thanks! Since I posted this elsewhere, here're the original XISF files (as well as FITS in case you don't want to use PI) if you want to take a shot at the raw data. These two are flats calibrated, slightly lower SNR so I didn't use it:

http://ram.org/images/space/downloads/ou4_O3.v2.xisf
http://ram.org/images/space/downloads/ou4_O3.v2.fit

This is not flats calibrated. This is what I've used to produce the images you see, like this one:

http://ram.org/images/space/downloads/ou4_O3.v0.fit
http://ram.org/images/space/downloads/ou4_O3.v0.26.jpg

--Ram

[STEVE333]

Beautiful.

Looking forward to seeing the combination.

Steve

[ram]

Hi all, I've now created decent final images of the Ha data and previously I had processed this framing of the O3 data entirely so I decided to create a HOO image using the method described by Jim Lindelien (https://telescopius.com/pictures/view/6 ... -lindelien). Just playing around but I'm really wondering about how to position the Bat with the Squid: pointing down, pointing up, or sideways (which is how I process it). I've reduced the images a lot so each file is around 1 MB:

So most people prefer to take the head of the bat to be part on the right side of the image when the image is pointing out (see Jim's image) but to me it's more natural to have the head be where the dot of O3 is especially in this framing (but I'd like to know what people think). But f you look at the full widefield (this is a draft copy with 25% of the data), the wings go back a long away so this looks better: http://ram.org/images/space/downloads/Image40.jpg - fascinating target given the multitude of perspectives possible.

I still have to do the S2 and then redo the O3 since I'm going to add the final 4-5 hours of data to it from the widefield collection and reprocess if I have the patience/werewithal, either for this narrowfield version or the widefield version which comes AFTER this which will include the full field.

--Ram