IKO - SH2-155 Cave Nebula SHO - Processing Competition

[Grant]

We've just released another set of data from the Ikarus Observatory project here.

Following the wonderful entries from the first processing competition we are running another to see who can get the most from this data and create the 'best' image! The winner will receive £150 of First Light Optics gift vouchers and two runners up will receive £50 of gift vouchers.

Details below:

• Closing date: 22nd October 2021 @ 22:00
• Please post your entries into this thread
• Please only use the data we've released - don't blend it with other data or add to it with other sources
• Use what ever processing software and techniques you like, potential bonus points for sharing your workflow and techniques with others - that way, we all get to learn as well 
• Multiple entries are allowed but, please refrain form posting multiple attempts that are very similar - better to update an earlier attempt if it's just an incremental improvement but, it would be interesting to see totally different attempts using different narrowband blends etc...
• If you win or are a runner up, we would really like to use your creation in future marketing materials, on our website etc.. so by entering the competition, you are giving us permission to use your image in this way.
• Judging will be done by a small team of judges from the FLO team - It's completely subjective  but will be based on what image(s) we think do the best job of extracting the most from the data released and, look the 'prettiest' 
• We will announce the winners in early November via SGL

Thanks all and look forward to seeing your entries!

[Budgie1]

I'll kick this one off then.

Beautiful data again, didn't even bother with noise reduction on this one!

So, this is an SHO rendition using PixInsight only. A slight crop, then ABE, followed by the stretch and Starnet.

Worked on the stars to give a more natural look, used the CorrectMagentaStars script and then a little CloneStamp to remove some yellow centres to the larger stars. A slight star reduction and that was about it.

Moved on to the nebula: Started with easing the saturation with several attempts, then enhance the blue, used CIE b* to enhance the blue & yellow then the lightness to bring out the dark dust clouds. Finally played around with UnsharpMask to bring out more detail. Then PixelMath to recombine the background & the stars and resize. 

[JSeaman]

This is an SHO version using Pixinsight and Photoshop

I stretched each layer and combined into an RGB then, using Selective Colour, I adjusted a separate layer to boost vibrance which was merged to pull out detail of the nebula

An agressive removal of red followed before some levels and curves then finished with a morpholigical transform which was selectively masked in

I did crop a tiny bit as there was some noise on 2 edges
 

[Anne S]

Here's my rendition.

DBE on separate channels, then combined in the Hubble palatte. Convert to non-linear.  Remove violet star by inverting image and using SNCR. invert again to obtain blue/gold colour scheme.

Starnet to create a starmask. Extract luminance and perform HDR on that. Saturate colour version then recombine with luminance.

Morphology on Starmask image, then saturate stars a little. Use Pixelmath to add star to starless image.

Slight MMT to reduce colour noise. Use star image as a mask to enable me to adjust the colours slight to strengthen the blue/gold effect. Remove mask. Crop to tidy up edges.

Lovely data, I wish I could do as well.

 

Edited September 29, 2021 by Anne S

[Phillyo]

Beautiful data to work with!! This is my attempt at an SHO version.

[AndyWeller]

Thanks for the chance to play with some really nice data

My interpretation. Mixed and processed in pixinsight and photoshop

Cleaned the backgrounds up using the automatic background extractor. some denoising with the multiscale linear transform tool. these images are pretty clean so didn't do much

I stretched each image and then used pixel maths to do a Combination of the SHO and ForaxX Pallets. i made 2 versions SHO and ForaxX Pallet and then used pixelmath to make a 50/50 blend.

SCNR on the inverted image to reduce the magenta around the stars.

used the EZ scripts for some deconvolution but again this had minimal affect as the data is so good .

Used the Ha layer in the luminance channel to make it pop and bring out the detail.

used starnet++ to allow me to work on the main image and stars separately, i wanted to try and keep as much of the star colour as i could.

and then played with the curves, saturation and hue's to get a look and feel that i was happy with.

Photoshop for a bit of tweaking and logo .

2nd and 3rd image are SHO Hubble and ForaxX pallet respectively processed in the same way.

Good luck all

 

Edited October 8, 2021 by AndyWeller
cleaned the post up. Page tools took some understanding for me haha

[bonzodog]

StarTools ver. 1.8 alpha...

[Budgie1]

I thought I'd have a go at something a little different (and then do it again, so this is now V2 of the image ).

I started by doing a dynamic crop of the three channels and combining them as HSO before removing the stars with StarNet. Then I did a 60/40 combination of the Ha & Sii channels to produce a luminance channel then removed & discarded the stars from this image so they didn't effect the original star colour.   

This Luminance layer was then added to the HSO and I set about using Curves, HDRMulti and Unsharpmask to bring out the detail, lastly I used the DarkStructureEnhance script and added the original stars back in.

This produced a better image and Version 1 with enhanced background nebulosity sharper dark areas, so V1 was replaced.  

Edited October 7, 2021 by Budgie1
Updated image with different workflow.

[alacant]

Hi everyone. Thanks to flo and to all who posted. Lovely detail.

StarTools 1.8.512 with S in NBAccents. 

Thanks for looking.

[A320Flyer]

Hi,

Here is my attempt. Processed in PixInsight with the following workflow:

Combined Sii, Ha and Oiii to give SHO colour image

Stretched using HT

Colours remapped to Hubble Palette with Curves, using appropriate ColourMasks

Saturation increased

Ha used as Lum

Slight Deconvolution

NR using TVGD and MMT

Stretched using HT

Contrast increased on high SNR areas using iterations of LHE

Detail increased in background using HDRMT

Lum merged with SHO

Dark Structure Enhance using the DSE Script

Small stars decreased in size/intensity using MT and a star mask

Transferred to Photoshop for final colour balance and slight sharpening of the smaller details.

I rotated the canvas 180deg so that the highlights appear to be lit from above, as I think this looks more natural.

Sh2-155.tif

[teoria_del_big_bang]

@A320Flyer Great image and so nice to share your workflow 🙂 

Steve

Edited October 4, 2021 by teoria_del_big_bang

[Spikey]

Star Tools 1.8 Beta, NB accents as SII, set orientation and played around with colour and saturation.

[Spongey]

Amazing data as usual! Great fun to play with; I settled on a standard SHO combination in the end. I created a synthetic luminance channel from the three NB channels, weighted towards Ha.

Processed in Pixinsight and Adobe Lightroom.

My lum processing workflow from Pixinsight history explorer can be seen below. Happy to answer any questions on my workflow, please ask!

Cheers

Edited October 5, 2021 by Spongey

[teoria_del_big_bang]

My 1st attempt at this and reasonably happy with it but would like to have another go (but for another day).
I am putting a lot of time into learning PI this year especially as I have a lot of time on my hands with really poor weather conditions North of UK, so this IKO data has been a real Godsend to practice with.
This is a long way from my usual workflow as my usual methods didn't seem to do what I wanted, I could get the nebulosity and some colour but my stars seemed poor with big magenta halos I just couldn't seem to eliminate, I think I need to become better with using masks maybe and will try again.
All done in PI:-

1. Small Dynamic crop as I could seem something odd in the very edges of some of the images.
2. DBE on individual SII, Ha and OIII images.
3. Slight deconvolution.
4. SHO integration
5. ColourCalibration.
6. Initial stretch with Histogram Transformation. but only enough to get good size stars so nebulosity was very slight with little colour.
7. Using the three images SII, Ha and OII obtained after step 3 do  Histogram Transformation on each to permanently stretch them, but not too much.
8. Make each greyframe starless.
9. SHO combination of the starless frames.
10. Further Histogram Transformation and slight curves transformation to enhance colours.
11. Reunite the stars using the Blend Script.
12. I did like the image at this point but the now dark nebulosity on the LH side was quite magenta  and still some small magenta halos on the stars so image inverted, SCNR to remove green and invert back again.
13. Using Ha as Luminance combine with HSO image using LRGB Combination, this seems to bring about more detail and enhances the colours somewhat.

 

 

Steve

 

Edited October 8, 2021 by teoria_del_big_bang
Tweaked the original image

[gorann]

Here is my first attempt. Great almost noise free data. In cotrast to the other entries here I used a HSO palette and tweaked it towards a HaRGB-like palette to give it a more natural look. Processed in PI (a little) and PS (a lot).

Cheers, Göran

[gorann]

I just found this thread on an new RC-Astro star removal plug-in for PS:

 

So I tried it (first image below) and then blended in the stars at a much reduced level compared to my post above. I really like the result. Could not find many annoying artefacts.

Edited October 9, 2021 by gorann

[Anonymous Astronomer]

Hello everyone! 

Image processed in PI:

- Decon and Denoise on each channel
- DBE
- Histogram stretch
- Channel combination with PixelMath
- Applying Ha as Luminance with LRGB Combination 
- ACDNR
- Curves transformation
- Saturation boost
- Reducing Star Sizes with Morphological Transformation
-Dark Structure Enhance

Clear skies,
Vlad

Edited October 6, 2021 by Anonymous Astronomer

[Whistlin Bob]

Here's my effort. Actually it took a couple of goes. Starnet++ always leaves a few artefacts, but this one seemed particularly bad, so I kept the stars in and used masks to protect them during the process. I did a little bit of reduction, but decided I really liked the big clouds of stars as part of the image, so less than I usually would. Approximate process was-

- Combine Channels in SHO

- Rotate 180 degrees

- Neutralise Background and Colour Calibrate

- Mask the brighter bits and very gentle NR with MLT

- Stretch with Histogram transformation

- Mask the stars of and use Curves/Hue and SCNR to get get the colour to my liking.

- Reverse the mask and saturate/reduce the stars

- Take a Luminance image and use it to mask of the darker sections; use USM and a tiny bit of LHE to bring out detail.

- Use Dark Structure enhance to bring out the Dark Nebulae.

 

[Xiga]

Unreal data, as always. A joy to process! This was my workflow:

1. Used AstroPixelProcessor to remove what little gradient there was in each channel. I then combined them as SHO to create the RGB image (also in APP). The advantage to this is, APP will scale the channels accordingly to balance them. Before, i used to stretch each one individually and then combine, but this doesn't produce channels that are anywhere near balanced, so you then need to balance the histograms yourself. Once the RGB image was created, i gave it a light DDP stretch and saved it as a 16bit Tiff. I then made a 2nd RGB image, combined as HSO, to use just for star colour later in processing. 

2. Open the Tiff in PS, crop off the edges and then extract that S-H-O channels from the R-G-B. Save each one for individual processing. 

3. Run Starnet++ on each channel. Clean up any remaining artefacts using the Spot Healing Brush in PS. 

4. Denoise each starless image. I used Topaz Denoise Ai at 1% level. Even this was too much, so i recovered some of the original detail using the slider. No sharpening. 

5. Create the Tonemap by copying the denoised SHO channels into a new RGB image. Then do Selective Colour adjustments to taste - i like Bob Franke's  approach http://bf-astro.com/hubblep.htm

6. Create the Lum layer. For this, i used Topaz Denoise Ai on the original Ha image (not the starless one). Similar settings to above, but this time with some mild sharpening. The small & medium sized stars are now a mess (due to sharpening) so i bring it into PS and layer the original Ha image on top and add a layer mask to only allow through the original Ha stars. This layer mask is created from the Original Ha image and the Starless one using Blend Mode Difference. The mask needs the white point brought waaay in, in order to protect the teeny tiny stars. This is now the base Lum layer.

7. Start a new image in PS, with the Tonemap on the bottom as a Color Layer, with the Lum on top as a Luminosity Layer. Some minor Curve adjustments to both. 

8. Bring the HSO image from pint 1 above into PS as a new image and work on it for star colour. Duplicate the layer, set to Color, run HLVG (to remove green) then invert the image, run HLVG again (to remove Magenta) and then re-invert back again. Boost the saturation, being careful not to overdo the Reds and Yellows. This is now the star colour layer. Add it into the main image, set Blend Mode to Color, and add a layer mask (the same one we made earlier containing only the stars). This adds the nice star colours to the stars only, leaving the colours we created earlier in the tonemap alone. 

9. Added some contrast using the Clarity slider in PS. Making sure to protect the stars, using the same star mask as before (only now inverted). Then add a small amount of Vibrance. 

10. Added a very small % of NR using Carboni's Space Noise Reduction. Only needed in the top portion of the image. 

11. Star reduction pass 1. Minimum Filter on just the very largest stars. I only selected the biggest 10 or 15 for this. 

12. Star reduction pass 2. Used ImagesPlus to do one pass of star reduction at default settings. Then brought this into PS as a new layer and dropped the Opacity slightly, to re-introduce a slightly soft edge to the small stars. 

13. Rotate for better framing (to my eye at least) and done!

Edited October 7, 2021 by Xiga

[mareman48]

On 29/09/2021 at 00:47, Budgie1 said:

I'll kick this one off then.

Beautiful data again, didn't even bother with noise reduction on this one!

So, this is an SHO rendition using PixInsight only. A slight crop, then ABE, followed by the stretch and Starnet.

Worked on the stars to give a more natural look, used the CorrectMagentaStars script and then a little CloneStamp to remove some yellow centres to the larger stars. A slight star reduction and that was about it.

Moved on to the nebula: Started with easing the saturation with several attempts, then enhance the blue, used CIE b* to enhance the blue & yellow then the lightness to bring out the dark dust clouds. Finally played around with UnsharpMask to bring out more detail. Then PixelMath to recombine the background & the stars and resize. 

Very nice, thanks for sharing this 🤩

[mareman48]

On 30/09/2021 at 10:09, AndyWeller said:

Thanks for the chance to play with some really nice data

My interpretation. Mixed and processed in pixinsight and photoshop

Cleaned the backgrounds up using the automatic background extractor. some denoising with the multiscale linear transform tool. these images are pretty clean so didn't do much

I stretched each image and then used pixel maths to do a Combination of the SHO and ForaxX Pallets. i made 2 versions SHO and ForaxX Pallet and then used pixelmath to make a 50/50 blend.

SCNR on the inverted image to reduce the magenta around the stars.

used the EZ scripts for some deconvolution but again this had minimal affect as the data is so good .

Used the Ha layer in the luminance channel to make it pop and bring out the detail.

used starnet++ to allow me to work on the main image and stars separately, i wanted to try and keep as much of the star colour as i could.

and then played with the curves, saturation and hue's to get a look and feel that i was happy with.

Photoshop for a bit of tweaking and logo .

2nd and 3rd image are SHO Hubble and ForaxX pallet respectively processed in the same way.

Good luck all

 

Nice work, wish I could see that with my own eyes 👁️👁️

[bottletopburly]

Processed in startools 1.8.516

 

Edited October 31, 2021 by bottletopburly

[Rob63]

Thanks once again for the great data, this one was more challenging to process particularly in the OIII channel.

Processing details listed below image.

 

Crop & DBE to each channel, 

To get a stars layer, use PixelMath to combine as SHO and starnet++ then delete SHO combine as we will create this differently.

Each channel
Small stretch with HistogramTransformation
Starnet++ to remove stars
this allowed me to fix some of the issues with the OIII halos/artifacts around bright stars

Balance channels with HistogramTransformation
Create SuperLum (90%Ha)
PixelMath combine as SHO 

Multiple small stretches with HistogramTransformation then a MaskedStretch 
LRGBCombination using SuperLum
Remove magenta stars with pixelmath
HDRMultiscaleTransform (masked)
LocalHistogramEqualization 
ExponentialTransformation
SCNR

Stars layer
For the stars, I used several Small stretches via HistogramTransformation to keep them tight with good colour.
PhotometricColorCalibration for star colour and background neutralisation

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

 

[urbko]

My try hope you like it

[AndyWeller]

On 09/10/2021 at 08:41, mareman48 said:

Nice work, wish I could see that with my own eyes 👁️👁️

Thanks Mareman48,

Wouldn't that be something? can you imagine how our night sky would look if all of the deep sky objects were bright enough to be seen with the naked eye. Would be trippy!

Edited October 15, 2021 by AndyWeller