NGC 1566 The Spanish Dancer

[g__day]

So I am going deeper with this target from my Bortle 8 skies - up to 34 hours of data on my 132mm WO refractor.

I am selecting targets to shakedown all my gear and then get rich data sets as I start planning to do some online courses on PixInsight.

I figure it will be interesting to see what I can do before and after training when I have 50 hours or so of data on my chosen targets!

All LRGB 300sec subs - stacked and processed in PI.

PI Workflow
1. WBPP - high quality / no drizzle
2. Gradient correct all LRGB auto-cropped masters
3. Blur Exterminator (Correct only) the outputs of step 2
4. Blur Exterminator
5. Noise Exterminator (latest AI version 3 - defaults apart from iterations set to 3)
6. Channel combine RGB masters
7. SPCC on combined RGB master
8. Statistical stretch RGB master (defaults)
9. Star Exterminator on Lum master
10. Star Stretch (defaults) on Lum_Stars
11. Statisical Stretch on Lum_Nebula
12. Unsharpmask on Lum_Nebula (gentle < 2)
13. PixelMath combine ~(~Lum_Nebula * ~Lum_Stars)
14. ImageBlend combine Stretched Lum (from step 13) with Stretched RGB (from step 10)
15. GHS Linear - slight rise of black point
16. Curves Transformation - slight rise of Chrominance and Saturation
17. Histogram Transformation - slight lowering of mid point
18. Noise Exterminator (defaults)

So I will look back on this in 6 - 12 months and see how my processing workflow and results may have changed!

[PRejto]

Okay, good start!

I will refrain from commenting on the process (though I am tempted!). But no.

I would only suggest that you choose a larger target initially. It will be much easier to evaluate what is going on at a small scale.

Good luck. It's a pretty deep dive but you will emerge from the other side with a lot of appreciation for how the tools can work. Hopefully you have the right instruction; it's not created equally!

Peter

[gregbradley]

That's a good go at the Dancer. Main thing that stands out to me is the background is black clipped and this is not showing the fainter spiral arms as a result.

Greg.

[alpal]

Yes the background is reading zero of 255.
It should be in the range 20 to 30.

[g__day]

Hi guys,


Totally agree - will improve it as I fell into the beginner mistake of clipping the background too far to give a neat black space look - that's my next fix to address in my imaging - and luckily a pretty simple one to correct!

One thing I do wish to explore further and soon is once one has a large data set - how best does one determine what shots go into stacking and which get pre-rejected? When one is considering S/N or other factors like sky quality or star count or HFR how does one draw the line to say either take everything above some weighted set of criteria versus take only the best 10% or best 50% or reject only the bottom Y%. With a large data set does the image improve if you take only your very, very best images or do you let WBPP weighing sort this out and say throw nearly everything into it and let it decide what to keep and reject and how to optimally weigh what it is seeing?

Thanks,
Matt

PS

I purchased the AdamBlock series and am about to go through them - from the basics up to get all I can from it.

[gregbradley]

It’s a bit of an art I think.
Given the obstacles we face: lack of imaging time,
Cloud, unexpected cloud, fog, issues with gear and software
It all seems to conspire against getting an image.

So keeping that in mind you could simply experiment
On a largish data set what improvement you get by getting
Tighter with FWHM and see.

The biggest gains I see are deleting partial cloud affected subs.
They can give weird gradients and colours.
Tighter stars are going to give a gain but at what point is there
Diminishing returns and leaving out exposures on skimpy
Data is a recipe for excessive noise in the image.
Mono camera filtered imaging is also very vulnerable.
You get a nice set but no or little red for example.

Not sure there is a pat formula.
Greg

[PRejto]

Adam Block's tutorials are great. You won't go wrong with those!

P

[gregbradley]

I agree, I think he is the best processing tutor.

Greg.