Data set of IC443 in SHO for processing

Questions and answers about processing in StarTools and how to accomplish certain tasks.
riccdavis
Posts: 7
Joined: Tue Jan 25, 2022 11:29 pm

Re: Data set of IC443 in SHO for processing

Post by riccdavis »

Thanks to everyone for the contributions so far. Interestingly I initially produced a result very similar to that of Ivo (OK my sky was a little too blue and I had intensified the colours a little more and I hadn’t quite got the same amount of faint detail) but still remarkably similar. Being from a data handling background in the dim and distant past, The thing I really liked about Startools was the data integrity and the ease of use. The whole ethos of it struck a chord with me somewhere. I have produced so far with it images of M31 and the Pleiades, both in LRGB and I was very happy with them, as I felt they looked like a lot of good images I saw on the web of those objects (I only had the demo version at the time, so just have screen grabs of the final images (I’ll do a full res reprocess soon 🙂)

But then when I started doing NB I thought….hang on these images look different to the vast majority of NB images I see out there. Am I doing something wrong or is my data not good enough or is the post processing software not up to the job? I therefore sent the data to my friend to process in PI to clear it up. Far from clearing it up, it’s actually opened a can of worms it seems 🤣
3BA2580F-3CCE-48F1-9373-702876F29B73.jpeg
3BA2580F-3CCE-48F1-9373-702876F29B73.jpeg (442.66 KiB) Viewed 2136 times
riccdavis
Posts: 7
Joined: Tue Jan 25, 2022 11:29 pm

Re: Data set of IC443 in SHO for processing

Post by riccdavis »

709B7E85-0177-48AB-BB6E-49795F49B44B.jpeg
709B7E85-0177-48AB-BB6E-49795F49B44B.jpeg (725.76 KiB) Viewed 2136 times
User avatar
admin
Site Admin
Posts: 3367
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: Data set of IC443 in SHO for processing

Post by admin »

Mike in Rancho wrote: Thu Jan 27, 2022 4:27 am Was your Wipe essentially default other than the DAF? Being narrowband, I had hit the NB preset, but also noticed the major noise, and actually ended up at aggressiveness 95 with DAF 3. Too much?
I naturally tried the NB preset first as well, but there was a gradient visibile (the NB preset assume no gradients as stray light is much rarer in NB). I therefore just used the default preset.
I believe any of the tricolor matrices, as well as pulling back on a filter (here the Ha), are still legit representations of the concentration? Even though I chose a pretty unusual mostly blue version. :D I've read the docs and website, but am still trying to wrap my head around how they stay relative to each other when the mapping matrices are used. Is it that it is still a tricolor, and the listed combos (for example 70Ha+30OIII) just set the particular color/hue for that element? I know the R G and B remain tied to their filters, so maybe I'm finally coming around to getting it.
You are! The different blends just lead to different hues. By far the most dramatic is the straight up SHO->RGB mapping ("unity matrix"), but it does tend to show a fair bit of green. The alternative mapping used by the SHO(HST) preset picks hues that steer away from the green.

There are good reasons (historically) to not prefer too much green, as it quite literally overpowers other hues (humans are exceedingly sensitive/drawn to green, while rendering media are also predisposed to emit higher brightness in green). Pure blue, in particular, is not as well visible (try some big squares of pure green, pure red and pure blue next to each other on your screen and you'll see what I mean).
I did not think of the two SS runs (or Isolate instead of dim), but will check it out on this data again. I also chose 20% for the Airy parameter, so my PSF grokking still needs some work it seems. Close! But not a precision 13%.
It's just a ballpark/uidance value - what matters most, is that you considered it, and - rightfully - decided you needed something smaller for this image, as it is a widefield-ish image. I may add some visual guide in the next version of StarTools... :think:
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
Posts: 3367
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: Data set of IC443 in SHO for processing

Post by admin »

riccdavis wrote: Thu Jan 27, 2022 8:15 am Thanks to everyone for the contributions so far. Interestingly I initially produced a result very similar to that of Ivo (OK my sky was a little too blue and I had intensified the colours a little more and I hadn’t quite got the same amount of faint detail) but still remarkably similar. Being from a data handling background in the dim and distant past, The thing I really liked about Startools was the data integrity and the ease of use. The whole ethos of it struck a chord with me somewhere. I have produced so far with it images of M31 and the Pleiades, both in LRGB and I was very happy with them, as I felt they looked like a lot of good images I saw on the web of those objects (I only had the demo version at the time, so just have screen grabs of the final images (I’ll do a full res reprocess soon 🙂)

But then when I started doing NB I thought….hang on these images look different to the vast majority of NB images I see out there. Am I doing something wrong or is my data not good enough or is the post processing software not up to the job? I therefore sent the data to my friend to process in PI to clear it up. Far from clearing it up, it’s actually opened a can of worms it seems 🤣

3BA2580F-3CCE-48F1-9373-702876F29B73.jpeg
Those renditions are really rather excellent at the scale presented! That's quite impressive for someone who just got introduced to the software.

It's very good to hear you achieved a similar image, following a similar standard workflow. I try to keep things as vanilla as possible when discussing workflows, so that the basic science and signal processing is out of the way and personal interpretation/vision can take over. :)
Ivo Jager
StarTools creator and astronomy enthusiast
riccdavis
Posts: 7
Joined: Tue Jan 25, 2022 11:29 pm

Re: Data set of IC443 in SHO for processing

Post by riccdavis »

Thanks Evo!

My early success is clearly due to your excellent software!

I also have some images of the Heart and Rosette in NB (which I thought were ‘wrong’ in terms of the colours.) When I reprocess those, I will now have to reassess that opinion. It is interesting that I felt my RGB images were relatively good compared to many I saw (certainly colour wise) but my NB images didn’t have the ‘wow’ factor. I guess we all know instinctively what RGB looks like but not NB?

Keep the crusade going 🙂👍🏻
riccdavis
Posts: 7
Joined: Tue Jan 25, 2022 11:29 pm

Re: Data set of IC443 in SHO for processing

Post by riccdavis »

I naturally tried the NB preset first as well, but there was a gradient visibile (the NB preset assume no gradients as stray light is much rarer in NB). I therefore just used the default preset.
I think this is moonlight leaking into the Oiii channel as the moon was very close and full for some of the time. I also wonder if that is why the Oiii channel seems relatively weak? I’ll add some more moon free Oiii data if the skies ever clear 🤓
Mike in Rancho
Posts: 1141
Joined: Sun Jun 20, 2021 10:05 pm
Location: Alta Loma, CA

Re: Data set of IC443 in SHO for processing

Post by Mike in Rancho »

Nice job on the full spectrum images, Richard. :thumbsup:

I thought something may have been wrong as Ivo (who knows just from glancing at it) and Freddy did, but didn't want to say anything. Other than the mismatching colorization, it holds up at a distance. But up close zoomed in, things starting looking very wrong, though mostly I was noticing the strange grid pattern over the large (and overstretched) stars. I'm not quite good enough yet to notice Topazizing, but I'm starting to get better at seeing things like range mask megablurring and color painting. :D

Anyway I want to thank Richard and Ivo for the conversation, as it's helping me pick up on a few things, and good reminders as to others. :thumbsup:

I too used a pretty similar workflow, despite shifting to blue-heavy with the mapping. But ST is so, hyperdynamic might be a good word for it, that differences in Wipe, AutoDev, and so on can lead to the ST engines seeing things somewhat differently. I became too worried about the noise early on and tried to squash it, rather than waiting it out for noise tracking and things like SS-Isolate (I've been using Dimsmall too much) and Shrink-unglow too.

Just a few small changes, and sticking with SHO, brought the image much closer to Ivo's on a subsequent try. Yay!

The NB color mapping is the biggest thing, I'm getting closer but I don't think I'm there yet with the descriptions. Let's take one example, that I used on my second go-through. Matrix is [SHO 40SII+60Ha,70Ha+30OIII,100OIII].

They all add up to 100% for each of the R, G, and B, and that's where I'm getting lost. Since we are actually thinking of a tricolor image based on the filter wavelengths, each element/molecule emission with its own distinct separate hue (based on the color wheel for our improved perception if I have that right), shouldn't each filter add up to 100, instead of each channel? For example OIII = 77% blue plus 23% green. Is that the blend for the color that now represents OIII, a sort of lightly green-tinted blue? However the way the matrix description reads just gets me crossed up, as it seems that the red channel is a blend of the SII and Ha. And even if we think of red as SII, why would we blend that element with another element, here Ha, and ruin it? Does any of that make sense, or is it just my strange brain? :lol:
Post Reply