IC410 - Pointers please

[volcanotop]

Hi Folks,

Attached what i consider my first real astro photograph (> 2hr integration).   IC410 from earlier in the month.   I've been playing and learning and buying the equipment for a year or so and just getting to the stage where I think I'm getting okay at the capture side of things.    ie polar align, guiding, dithering, plate solve etc etc.    The data for this image i think is ok other than I think the focus could be a bit tighter.  (lesson learned - i didn't focus each filter,  just the Ha at the start of the 2 hour session)

Processing is a different story.   I'm a month into the Pixinsight learning curve.  I've been reading and trying to understand.    But still a long way to go.....

Linearfit sounds like a good idea to align the three stacks before combining them....but it seemed to make my pictures worse.  So i excluded that.

So the stacked images were combined, background removal.

For Background neutralisation it feels like i ought to use the black 'holes' in the center of the nebula as the 'sky' but that didn't seem produce a good result for me.  in the end I made 4 or 5 previews in the holes and also to the top of the image where there is less 'dust' and did a backgroundneutralsation on the aggregated previews.  then colorcalibration.

histogram transformation,   range selection,   hubble colors and a bit of curve transformation.

Anyway, pointers gladly received.  I wish i could really nail the hubble colours and also get a 3D looking final picture.   I seem to make flat looking 2D pictures.   No 'billow' to the nebulosity.....

Given that we seem to have non stop clouds for the next week or so.... DSS stacked data is shared here is anyone wants to process and post?  15x3 min subs for each H, S and O.   45 mins per filter.   2 hours 15 in total.

https://1drv.ms/u/s!Am0Xaik4MKPAp3nLTfhx_tG82AEa?e=8GsaV5

Thanks!

Peter / volcanotop

https://astrob.in/wrzh6j/0/

 

 

 

Edited January 26, 2020 by volcanotop

[Davey-T]

Never used PI but the colours don't look lined up.

Dave

[Davey-T]

Software complains that Tiff files do not have embedded greyscale profile whatever that means.

Dave

[Davey-T]

Just put them in P'Shop and aligned channels, stars look better, ignore the funny colours, doesn't look like PI aligned them properly, need a PI expert.

Data looks alright so should be able to extract a decent image.

Dave

You can see from the top edge that they weren't aligned.

Edited January 26, 2020 by Davey-T

[don4l]

Here is what the Gimp did.  Just auto stretched each image, RGB combined and flattened the OIII background a bit.    I haven't done anything with the alignment.

The data is looks good, but the focus could be better.

I'm afraid that I don't know anything about PI.

[Davey-T]

36 minutes ago, don4l said:

I haven't done anything with the alignment.

You could've aligned the colours Don 😂

Dave

[don4l]

28 minutes ago, Davey-T said:

You could've aligned the colours Don 😂

Dave

I'm a bit of an idiot at times.  CCDStack didn't like the tif files.  Anyway, I've now learned a bunch of settings in the Gimp that fix that problem.

"tif" should be a standard format but every software author has their own idea of what is best for us. 

 

[Davey-T]

3 minutes ago, don4l said:

CCDStack didn't like the tif files. 

Got one of those unhelpful messages from Maxim, didn't try CCDStack, I think it maybe because they're DSS 32bit I changed them to 16bit for P'Shop.

Dave

[volcanotop]

The tif were straight from DSS.  Nice processing Davey-t

[Davey-T]

2 minutes ago, volcanotop said:

The tif were straight from DSS.  Nice processing Davey-t

Must be something missing in the DSS file header, didn't do any processing as such just stretched and aligned in P'Shop, didn't even put them in the right colours 😂

May give it another go if I get a moment, got a load of images of the Tadpoles last week that I'm only just getting around to.

Dave

[wornish]

Thanks for sharing the data.

Stacked in APP using SHO pallets

SHO Processing Final Steps

Reduced magenta in stars by using PS Reduce Color Noise slider to 100%
then In PS use Image/Adjustments/Selective Color

Colors : Greens
Cyan : -100
Magenta : -25
Yellow : 0
Black : 0

Colors : Yellows
Cyan : -100
Magenta : +25
Yellow : 0
Black : 0

Colors : Cyans
Cyan : -1
Magenta : 0
Yellow : -100
Black : 0

Colors : Cyans
Cyan : 0
Magenta : -25
Yellow : -100
Black : 0

[volcanotop]

Thanks all for taking the time to process and post.   its interesting to see what others can make of the data with different tools and styles etc.

Aside from the focus being a wee bit off the data seems to be not to bad.

I'm continuing my PI learning curve during the free trial period.     I've played with MLT and Deconvolutuon and they seem to be making my image better.   Still a lot more to learn... 

 

 

Edited January 28, 2020 by volcanotop

[carastro]

For a first real astro photograph I think you have done well.  Yes lessons to learn, (focus as you've already realised), and registration.  I don't use Pixinsight so can;t advise on that, but I can tell you it took me a several months of trial and error before I nailed it.

I think you 2nd effort is probably a little too processed, but I took that one and re-registered the channels so that the stars lined up.  I also adjusted the magenta so the stars look a lot better now.

Carole 

 

However I just noticed you have posted up the actual individual filters, so I will try processing from scratch as well.

Carole 

Edited January 28, 2020 by carastro

[carastro]

I have processed it from scratch in Photoshop.  Registered in Registar and tried to keep as close as possible to the colours you chose though not 100% successfully. 

Unfortunately since you are using Pixinsight I can't really offer anything useful except to say your need to register your filters so that the stars line up.  But it shows that data is OK.  

Carole 

 

Edited January 28, 2020 by carastro

[volcanotop]

thanks again all.    one last attempt from me.  better aligned this time and SHO rather than trying to do the Hubble.

[dannybgoode]

I’d be happy with that as a first effort. Much better than mine was. 
 

I too am learning PixInsight and if you haven’t already got it then order a copy Inside PixInsight. 
 

Watching YouTube etc is ok but the book really nailed down exactly how to use it, the best steps to take and in which order etc as well as digs into the more advanced stuff. 
 

Seeing as I am new to it as well I will try to download your data and have a go

[carastro]

Your latest revision is much better, well done.  Only comment is that it is predominantly green. 

Carole 

Edited January 28, 2020 by carastro

[volcanotop]

Thanks again all.   One last version with the green zapped and as close to Hubble palette as i can get with my skill.

For those on the PI learning curve too here's my process.   (im sure s i learn more these steps will be improved too)

1. individual filters stacked in DSS (with a reference frame - thats what i missed with the misalignment problem)
2. LinearFit H and S using O as the master
3. Chanelcombination as S H and O
4. DynamicCrop to get rid of the edges.
5. ABE - subtract to get rid of the background.
6. Background neutralization.  Using the hole in the middle as 'black sky' preview
7. ColorCalibration using the whole image as while and the hole in the sky preview as black  (there is very little black in the image)
8. Make a DynamicPSF and Start Mask as prep for Deconvolution
9. Deconvolution with the PSF and Stark Mask (messy trial and error of the global dark value)
10. MLT de-noise on 5 layers (trial and error on the values)
11. Histogramtransformation stretch.
12. SCNR to Zap the green and get an approximation of Hubble colours
13. RangeMask to highlight the nebula.
14. Curves to enhance the nebula
15. Invert the mask
16. Curves to darken the background.

 

now if only the clouds would leave I'd get some more data to learn with....

 

 

 

Edited January 29, 2020 by volcanotop

[Filroden]

17 minutes ago, volcanotop said:

individual filters stacked in DSS

I would recommend giving PI a go for the image calibration, registration and integration as I think it gives much better results than DSS.  PI has a batch process that simplifies this process at the expense at removing many of the options. It works in a similar way to DSS in that you load your bias, darks, flats and lights into different areas then select some calibration/integration parameters then leave it to work its magic. Alternatively, there is a really good step-by-step guide here for doing it the longer way. Again, I think this way gives better results.

https://www.lightvortexastronomy.com/tutorial-pre-processing-calibrating-and-stacking-images-in-pixinsight.html

  

17 minutes ago, volcanotop said:

• Background neutralization.  Using the hole in the middle as 'black sky' preview
• ColorCalibration using the whole image as while and the hole in the sky preview as black

PI has a new process called PhotometricColorCalibration that effectively replaces the above two steps into a single step. It uses plate solving to identify the stars within the image and uses that information to properly balance colours based on their spectral types. You have to get the initial scope/camera parameters right to enable the plate solving but once these are set, this is module can be used with default settings almost all the time.

I also like TGVDenoise once you've stretched the image to help further reduce noise. There's a good noise reduction tutorial here.

https://www.lightvortexastronomy.com/tutorial-noise-reduction.html

Finally, I'd recommend Warren Keller's book, Inside PixInsight which received a second edition in 2018 to cover some of the new features in PixInsight. It remains my goto book to help me through my workflow.

Edited January 29, 2020 by Filroden
Added colour calibration comment

[wimvb]

The easiest way to get rid of those magenta stars:

Invert the image (process - intensity transformations - invert) 

Apply scnr (process - noise reduction - scnr) 

Invert