StarTools

Hi Ivo
First let let me say, your software is awesome . Now that I've buttered you up ..............I've been able to get to a point with the software that I am able to get better results than I was getting with "other"programs and in less time. With that said, I still feel that I don't really have as good a grasp on its use as I should. Here is a link to my latest effort http://astrob.in/56158/ NGC 281(the Pac Man nebula) totally* processed in StarTools(after stacking in DeepSkyStacker), 50 subs of 10 min each with no background calibration. I like my result but would really appreciate having an expert in the software (you of course) process the image to see what you come up with. I cant help but think that as I fumble around, I may not be getting all I can from my data. Here is the unaltered stacked data https://www.dropbox.com/s/cup9nxi2t2cj6 ... PacMan.FTS When you get the time and if you are so inclined, could you take a stab at it and post your results and the process you used. Feel free to use/share my data anyway you would like. * Let me fess up and say that that image wasn't totaly done in StarTools. As a last step I opened it in PS and dodged a few of the bright bloated stars as I couldn't seem to fix them to my satisfaction in ST. thanks in advance for any help
Che

I forgot to add that it was stacked using flats and bias frames as well
Che

Hi Che,

Thanks for the kind words and buttering up. I'm known to respond well to that sort of thing...
You got some nice data there.

Here's what I did. Please note that I used a development version of ST for this, but most things should be replicable with the 1.3.5 beta.

As always, I use AutoDev to see what we got.
I can see a yellow bias (probably light pollution), and some tiny stacking artifacts.
The data appears to be oversampled at native resolution (e.g. 1 unit of real detail is 'smeared out' over multiple pixels), so really it's probably a good idea to Bin the data a little. I didn't do that in this case, just to see if we could use deconvolution to bring back the detail to 1 unit-to-1 pixel (which turned out to be wishful thinking, but you gotta try these things... ).
I also cropped out the stacking artifacts before doing further processing.
I applied Wipe with default values. The data is very good in the sense that it is evenly lit, with no dark anomalies (ex. dust bunnies or dead pixels). You did a great job with your calibration! Wipe loves this sort of data.
For my final global stretch, I launch the AutoDev module again and (obviously) choose to redo my stretch. Due to the presence of quite a bit of noise in the lower scales, is set Ignore fine detail to < 3.0 pixels. This ensures that AutoDev won't optimize the histogram transformation for bringing out that fine noise, allocating dynamic range to 'interesting' larger detail instead.
I attempted some deconvolution. The oversampling required that I intervened a little in the mask creation. I used the decon preset in the 'Auto' feature of the mask editor, clicked grow about 8 times until the stars and halos were well covered, and then inverted the mask (decon requires 'gaps' where big fat white stars are).
I chose a radius of 1.9 pixels. A slight amount of detail was recovered and some stars appear a bit tighter. With slightly binned data you may have better luck though.
With the same mask still active, I used the Sharp module. Use it to taste oviously, but this is what I used;
Parameter [Scale 1] set to [8 %] to avoid exacerbating the finest noise.
Parameter [Scale 2] set to [50 %] also to avoid exacerbating fine noise.
Parameter [Scale 3] set to [100 %]
Parameter [Scale 4] set to [100 %]
Parameter [Scale 5] set to [100 %]
Parameter [Amount] set to [230 %]
Parameter [Small Detail Bias] set to [97 %] to give priority to small detail if two scales happen to be 'fighting' to enhance the same pixel.
As one of the last steps (recommened), I used the Color module (not that the settings below may not look correct in the current 1.3.5 beta - I'm still doing a lot of development on the Color module currently). Here too, the adage is 'season to taste', but I used the following settings;
Parameter [Cap Green] set to [To Yellow] to remove any spurious green pixels.
Parameter [Dark Saturation] set to [3.90]
Parameter [Bright Saturation] set to [Full]
Parameter [Saturation Amount] set to [100 %]
Parameter [Blue Bias Reduce] set to [1.00]
Parameter [Green Bias Reduce] set to [1.52]
Parameter [Red Bias Reduce] set to [2.09]
Before switching off Tracking and to push back the noise in the background, I ran the Life module with its Isolate preset at 75% strength.
Finally I applied noise reduction (switching Tracking off) with the following parameters;
Parameter [Color Detail Loss] set to [12 %]
Parameter [Brightness Detail Loss] set to [15 %]
Parameter [Redistribution Kernel] set to [6.1 pixels]
Parameter [Read Noise Compensation] set to [0.20]

This is what I ended up with;
I should've probably binned the data and the stars could stand some rounding, but all-in-all it's a rather nice Pac Man!

For the bloated stars, did the Magic module's 'Shrink' or 'Tighten' algorithms help at all?

Hope this worflow helps!

Hi Ivo
Thanks so much for doing this. I think I learned a thing or two I find it very interesting to see how the same data can turn out different ways. As for the bloated stars and the Magic Module, I'm sure it was operator error. At what point in the processing should this be done?
Thanks
Che

Cheman wrote:Hi Ivo
Thanks so much for doing this. I think I learned a thing or two I find it very interesting to see how the same data can turn out different ways. As for the bloated stars and the Magic Module, I'm sure it was operator error. At what point in the processing should this be done?
Thanks
Che

Good to hear you find this helpful Che. Processing is all about giving your own personal interpretation to a dataset (and this is mine - actually not sure i'm totally happy with it). StarTools is all about giving you the tools to attain the interpretation that you envision.

The Magic module is best used after you swithced off Tracking (it is not available during Tracking). It is purely a cosmetic function that can be applied whenever you want to touch up your final result.

OK thanks for that, good to know when to use it.
Thanks
Che