Jump to content


IKO - IC59 & IC63 Ghost of Cassiopeia Data Release - Processing Competition


Recommended Posts

Here's my efforts. I imaged this at home last winter but ran out of clear nights. Once it was clear it was too low to the north and behind the trees.

Two versions processed in PI, one in Hubble palatte which I think suits it as it's ghostly. The other HaS2O3 which after background neutralization ended with warmer colours. Also this version shows more detail in the nebulosity.

Thank you for the lovely data.


Ghost nebula _S2HaO3_final.png

Ghost nebula_HaS2O3_final.png

  • Like 13
Link to comment
Share on other sites

The data has that ethereal feeling and it reminded me of willow the wisp, looking at in colour the duo band blend emphasised this ghostly aberration.

StarTools 1.7.456 compose module | crop |bin | wipe | auto dev with ROI | wavelet sharpern | colour [H(H+O)O Duoband 100R,50R+25G+25B,50G+50B] | superstructure dimsmall | wavelet sharpen | filmdev slyglow lift | denoise (using a variety of masks throughout the process). saved tiff Irfanview to png for upload.

Edit: tightened the stars



Edited by happy-kat
  • Like 5
Link to comment
Share on other sites

My first attempt with this challenge. Processed In PixInsight with some finishing touches in PS.

Done as HSO as I'm not really comfortable with the SHO pallet yet. I wanted to emphasise the ghost and his tail and minimise y-Cas. I did this with much masking and stretching in PI

Lovely data like this makes processing so much easier and I really did not have to do that much to it

In Pixinsight;

Decon and a gentle denoise (MLT) on the Ha

Stretched all three gently using HT - further stretches with Star masks and range masks to focus on the nebula and keep y Cas in check. LHE (very gentle) on Ha

Combined in SHO-AIP script - several different blends until I got what I was looking for but essentially HSO with some of the Sii in red and some of the Ha in blue (I think!) with Ha as a luminance.

Some tweaking with HT and CT

Saved as 16bit Tiff and into Photoshop

Some very minor masked selective colour adjustment to the orange and blue

A selective high-pass on the "head and arms" of the ghost and a few other areas to bring out some structure.

All done!

EDIT -Rev 2 placed after Carbonis actions to bring some colour back to the stars.





Edited by mackiedlm
  • Like 6
Link to comment
Share on other sites

Processed almost entirely in Pixinsight.

Stretched individual channels to non-linear

Linear fit Ha and OIII to SII channel

Chanel combination using SHO (Hubble pallet)

Dynamic Background extraction

Starnet++ to remove stars - from here on the processes are mostly to the nebula sans stars

Boosted Reds and Yellows in Curves transformation

Hit it hard with SCNR

Inverted the image and hit it with SCNR to remove Magenta

TGV Denoise very gently - there wasn't much noise.

Adjusted curves to correct colours

Processed stars with Multiscale median transform to reduce the size slightly

Adjusted star colour in Curves transformation

Adjusted star brightness in Histogram transformation.

In Photoshop pasted the star layer onto the nebula layer at 95% 'lightness' then back to Pixinsight

Masked the big bright star and reduced it as far as possible whilst still leaving it big and bright - which it is!

Final tweaks in curves and Histogram Transformation

Saved as TIFF and Jpeg.

Full res TIFF below.








Edited by Skipper Billy
  • Like 11
Link to comment
Share on other sites

I've only ever used data from a DSLR or OSC so this was new to me. Processed in Startools 7 and tweaked in GIMP.

That star is a right nuisance!

Processed using the following modules in StarTools - AutoDev, Bin to 45%, Wipe (Basic, Dark Anomaly Filter to 5 and precision 128x128), Film Develop (Gamma to 0.90, Dark Anomaly Filter to 16, Home In then adjust manually), HDR, Colour, Shrink, Entropy, Tracking off and denoise. 



Edited by Mikey0368
  • Like 1
Link to comment
Share on other sites

Here's my attempt. Gradient reduction done in APP, everything else done in PS. 

Combined as SHO. Stars removed using Starnet, and then blended back in a lightly stretched Ha layer in blend mode Screen to add back the stars. 

Really amazing data, was fun to process, so once again thanks for sharing and taking pity on us poor folk who haven't seen a clear sky in weeks, if not months!

Edit - made a slight revision, with toned down highlights and slightly different colour palette. 


Edited by Xiga
  • Like 8
Link to comment
Share on other sites

I did a full process in PI and then another process in PS and came up with with very similar results!!!  In the end I blended a PI and a PS process together.  Used Starnet from within PI during the PS process but could have managed without PI by using the command line interface.  I find the hardest bit, by a long way, is making the choices - how much saturation, colour options especially when it comes to leaving in a bit of green, how much contrast, bright and vibrant or soft and wispy.  Shove the stars out of the way  for drama, leave the in or attenuate by a certain amount.  In the end I ask my wife what she thinks!  So this is a combined effort, thank you MrsB.



  • Like 4
Link to comment
Share on other sites

I wanted to do something a little bit different as it's been a while since I processed any images. So I made life easier for myself and processed all 3 channels in mono and then combined them. I removed the stars in starnet to make it more ghost like, in fact , quite like a dementor !


  • Like 4
Link to comment
Share on other sites

Narrowband channels all log stretched in Astroart, then Starnetted.  Layer masks in PSP used to minimise "Burn out" in bright areas.  Gamma Cassiopeiae cloned out of all channels, then RGB combined in PSP, R=Ha, G=SII, B=OIII. H alpha overlaid as luminance to retain detail. For stars, the unstretched narrowband data was RGB combined in PSP, then saturation lifted to bring up star colours.  Star layer pasted over nebula layer in screen mode. For this iteration, the log stretch for the narrow band channels was milder resulting in a less overpowering green from the SII channel and a more subtle rendition of the nebulosity.  I think the nebulosity around Gamma Cas is genuine rather than a gradient so I have left it in (again).


There is a lot of faint detail in the background and I have tried to include it but that does bring the background noise up, which I don't have software tools to handle. 

Edited by Hallingskies
  • Like 1
Link to comment
Share on other sites

Effort 1 from me- 

Masked Stretch on each channel.

Starnet, then switch to GIMP for a bit of layered blurring around Navi to soften some of the star removal artefacts.

Pixelmath to combined in HSO, with O boosted x2 in it's own channel and subtracted from the signal in the other 2 to make it stand out more.

Bit of Local Histogram Equalisation to bring out some of the texture in the nebula. 

Then, for the stars, back to the original channels and another masked stretch, but with background set to 0.05 and combined HSO.

Then combine the stars back in using a formula of iif(Stars>0.5,HSO*.4+Stars*.8,HSO). Finally, back into GIMP to soften Navi a bit.


Link to comment
Share on other sites

Effort 2 from me. Mostly the same process, but I always find with this really strong data that the temptation is to make the Sii and Oiii more dominant than they are (and why not- half the time I barely even get an Oiii or Sii  signal at home!), so I tried a stronger mix of the Ha into the Red channel to make it a bit more stronger.



Link to comment
Share on other sites

Thanks for the data! Here's my attempt in SHO, processed in Pixinsight.

Approximate workflow:

  • Combine to SHO
  • DBE
  • Background neutralisation
  • Linear noise reduction
  • Unsharp mask
  • Histogram stretch
  • Starnet++ to remove stars
  • Histogram adjustments and curves to increase contrast
  • SCNR and colour manipulation
  • Added Ha as luminance 
  • HDMRT & local histogram equalisation 
  • Stars added back in
  • Star reduction (morphological transformation)
  • MMT sharpening
  • Dark Structure Enhancement





Link to comment
Share on other sites

Thanks FLO for the clean data set, really enjoy doing this and amazed at the difference people produce

this is my SHO version, if i get another chance i might submit another try, all processed in PS using  carbonies tools, starnet++  etc image.thumb.png.408ca75b2a9b21148a0ec98854007a9c.png




FLO COMP final 1st process.tif

Edited by newbie alert
upload probs,still not sure if its done
  • Like 1
Link to comment
Share on other sites

Thanks once again to FLO & IKO for this opportunity to work with quality data as the awful weather continues to dominate in my part of the world!

Processing information below image.



Each channel
Balance channels with HistogramTransformation

I like to try something different each time I process an IKO image and this time I used PixelMath "Dynamic narrowband combinations" to combine as SHO (based on the post at https://thecoldestnights.com)

Multiple small stretches with HistogramTransformation then a MaskedStretch 

Remove magenta stars with pixelmath

Stars layer
For the stars, I used several Small stretches via HistogramTransformation to keep them tight with good colour. I opted to keep Gamma Cass big and bright as it's a key part of the composition for IC63

PhotometricColorCalibration for star colour and background neutralisation (narrowband mode)

Export to Photoshop
Combine nebula & stars
Adjust contrast, colour and saturation  of each layer to taste
Flatten and export

Edited by Rob63
  • Like 2
Link to comment
Share on other sites


Here's my entry. I've never processed a NB image before so this was an exciting thing to learn on a rainy Saturday afternoon! It's taken about an hour to do while I figured out what on earth I was doing with the stars, still couldn't get it right but here's the process I followed using Photoshop:

  • Change each individual image to an RGB 16bit image
  • Adjust curves and levels, trying to ensure levels are similar across each image
  • Load each image into its respective RGB channel in a new file
  • Made the stars smaller by selecting highlights, then using the minimum filter (2 pixels)
  • Colour corrected the stars as best I could using the selective colour tool
  • Used camera RAW filter to remove purple halos from stars
  • Used camera RAW filter to reduce noise and colour noise (which really helped with the sky)
  • Used camera RAW filter to sharpen the image to reduce the softening that the noise reduction did
  • A couple more small curves and levels adjustments
  • Saved as PNG
  • Enjoy a well earned cuppa tea while typing this up in the hope it helps others


  • Like 2
Link to comment
Share on other sites

Here's my best effort! I too have never processed a narrowband image before and I'm still learning pixinsight. It continues to amaze me how powerful this software is. Coming from photoshop though I do miss a few features! I thought I'd do something interesting with the narrowband data to see if I could produce a unique colour of the target. I found the most interesting to be OSH because you have the blue nebulosity contrasted against the red glow from the star Navi. Wish I could've extracted more of the fainter data in the HA filter but I'm happy considering I'm still learning!


  • Combine as OIII-SII-HA:R-G-B (pixelmath, no math operators, raw images used)
  • Autocolour script to balance the available data with a black background
  • Histogram Transformation - convert to non-linear data and align RGB channels
  • Range select - white for deep space background
  • Multiscale Liniar Transform noise reduction 
  • Masked stretch to bring out that nebulosity 
  • Curves transformation
  • ACDNR - more noise reduction
  • Local Histogram equalisation (a few runs at a minor strength)
  • SCNR (green) - remove minor green cast over image and stars
  • TGV Denoise (strength 2) - smooth out background noise that was creeping in
  • Colour Saturation - Push the blues to balance with the red from the Navi
  • EZ Star Reduction
  • Local Histogram equalisation
  • Final curves


FLO entry_less_stars.png

Link to comment
Share on other sites

Had to get a bit sweaty over this one, lots of star control required. Also, in OIII there is a noticeable halo on the bright star once heavily stretched (nothing can be done about that) but I modelled the backround for that area and subtracted (in Ps) for an overall flatter OIII layer. Tried to retain some decent star colour while I was there.

I know theres a fair bit of Ha lurking, but in an effort to keep the overall noise down I'll be slightly more restrained. The image should tolerate some pixel peeping too :)



  • Like 7
Link to comment
Share on other sites

Here's my attempt trying to recreate what it looks like in RGB which I did with the following blend..  Ha to Red, 50%Oiii, 50%Sii to Green and Oiii to Blue. Channels stretched individually in Pixinsight using the Screen Stretch and Histogram functions then combined using PixelMath. Exported to Photoshop for colour adjustments, a bit of noise reduction and repair work around Gamm Cass, then back to Pixinisight for LHE, star reduction and MMT sharpening. 




  • Like 4
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.