Astrophotography Processing Challenge April 2024

[AstroBee]

Happy April first everyone. This is no fooling, here's your challenge for April!

IC1795, The Fish Head Nebula is a star forming region in the constellation Cassiopeia approximately 6,000 light-years away. It is located near the tip of the Heart Nebula, IC1805
I started obtaining the Ha data from my Bortle 8-9 backyard and really loved the composition so I decided to spend some time at one of my Bortle 3 dark sky sites to obtain the rest of the RGB data as well as a bit of OIII and SII data for this image.
This data was captured through my Explore Scientific ED152 CF APO with the ZWO2600mm camera and Chroma RGB and 3nm Ha,OIII, & SII filters.

30x60sec Ea. RGB = 90min
154x600sec Ha = 25hrs40min
20x600sec OIII = 3hrs20min
20x600sec SII = 3hrs20min

Total Integration Time = 33hrs 50mins

I did not obtain nearly as much OIII and SII data as I wish I had but I still feel as though there is enough to work with here.
Options are a plenty. You can just process the RGB data as there is a good bit of nebulosity that shows up there.
But considering how rich the Ha data is, that would be ashame.
HARGB is another way to go or HOO or HOO with RGB stars or just go for it and give it the SHO treatment and add the RGB stars to it. This is one of those rare targets that actually has a lot of SII data.
These files are straight from PixInsight's WBPP where they were stacked, aligned, auto cropped. Only Cosmetic Correction and Local Normalization has been done to them during the WBPP process.

I'm looking forward to seeing your version of one of my favorite images.
Here's the link where you can find the .xsif, .fits, .tiff, or .jpg files. You only need to download the format you want to edit as each zip files contains the same data, just a different format.
https://drive.google.com/drive/folders/ ... wlJq1-qt6y

[yobbo89]

waiting for 75 iterations of deconvolution to be done ZZzzzZZZzz... fyi i think tif sii file is corrupted, fits folder works

edit: 15mins 22 sec later deconvolution is done ...

[yobbo89]

other users might want to re sample to 60% for upload

[Graeme1858]

... fyi i think tif sii file is corrupted,

Could you try it again please Rob. I've uploaded new copies of the .tif files.

Cheers.

[AstroBee]

Nice go at it Rob. I would love to see folks give more details on the steps they took. Was this every file or did you use narrowband stars?
I'm going to wait till the end of the month to share my version because I processed it quite a while back and was not keeping track of the steps I took.

Greg ~

[yobbo89]

Nice go at it Rob. I would love to see folks give more details on the steps they took. Was this every file or did you use narrowband stars?
I'm going to wait till the end of the month to share my version because I processed it quite a while back and was not keeping track of the steps I took.

Greg ~

used only sii-ha-oiii frames,

histogram stretch each frame
lrgb combine
scnr green removal
automatic background extraction
starmask
psf
clone image for masking
apply cloned mask to image
deconvolution with psf file,starmask for deringing,set to 75 iterations
remove cloned mask
slight adjustment with curvestransformation,adjust lrgb,saturation and hue for the stars to apear white
tvgdenoise 9 strength - smoothness 3.00
resample 60%
save file.

i've seen your version it's very nice , you make the colour pop.

[Graeme1858]

used only sii-ha-oiii frames,

Can I assume the newly uploaded Sii.tif was ok?

[yobbo89]

used only sii-ha-oiii frames,

Can I assume the newly uploaded Sii.tif was ok?

I havnt tried yet, I'll check it out later again for the tif folder, I used the fits folder for my image

[yobbo89]

used only sii-ha-oiii frames,

Can I assume the newly uploaded Sii.tif was ok?

yes the new upload for the sii.tif works

[Graeme1858]

Here's my go at LRGB using the Ha for the Luminance using the .xisf files.

4 x ABE Function degree 2
4 x EZ Denoise - TGV Denoise + Multiscale Median Transform (20 minutes each, I did some Arduino stuff while I was waiting)
4 x EZ Soft Stretch
LRGB Combination (using Ha for L)
Starnet2
Starless image - Curves - S Curve, Saturation
Star mask - Curves - Star reduction curve, Saturation
Pixel math
SCNR
Final Curves tweak
Save as .xisf
Save as .jpg
Opened in IrfanView - Resampled to 1920 x 1080

Cheers Greg, Good data, I'll have to have another go with the NB files.

Graeme

[starfield]

What a great data set, Greg. This is going to be fun to play with. First up I thought I'd try something different and go with the FORAX palete. Went for a high contrast look with this. Ended up with this:

SHO Processing:

On each SHO master:
BlurX Correct Only
DBE
BlurX Sharpen Stars and Nebulosity.
Light NoiseX
StarXterminator
Histogram Midtone Stretch on each channel
Combine in Forax Script Utility
Unsharp Mask
Lite NoiseXTerminator (not much noise here)
Curve tweak for contrast.

RGB
BlurX Correct Stars
DBE
Midtone stretch to get good stars.
Remove stars using Screen Stars Script

Combine Stars and Starless using Bill Blanshen NB combination script

Contrast tweak in curves. Export to Photoshop

In Photoshop:
Contrast adjustments in Camera Raw
Adjust color hues and individual saturation in Camera Raw

[Jockinireland]

Here's my go at LRGB using the Ha for the Luminance using the .xisf files.

4 x ABE Function degree 2
4 x EZ Denoise - TGV Denoise + Multiscale Median Transform (20 minutes each, I did some Arduino stuff while I was waiting)
4 x EZ Soft Stretch
LRGB Combination (using Ha for L)
Starnet2
Starless image - Curves - S Curve, Saturation
Star mask - Curves - Star reduction curve, Saturation
Pixel math
SCNR
Final Curves tweak
Save as .xisf
Save as .jpg
Opened in IrfanView - Resampled to 1920 x 1080



IC1795, The Fishhead Nebula.jpg



Cheers Greg, Good data, I'll have to have another go with the NB files.

Graeme

Thats really good Graeme. the stars have come out really well.

[Jockinireland]

What a great data set, Greg. This is going to be fun to play with. First up I thought I'd try something different and go with the FORAX palete. Went for a high contrast look with this. Ended up with this:

Forax@0.33x.jpg

SHO Processing:

On each SHO master:
BlurX Correct Only
DBE
BlurX Sharpen Stars and Nebulosity.
Light NoiseX
StarXterminator
Histogram Midtone Stretch on each channel
Combine in Forax Script Utility
Unsharp Mask
Lite NoiseXTerminator (not much noise here)
Curve tweak for contrast.

RGB
BlurX Correct Stars
DBE
Midtone stretch to get good stars.
Remove stars using Screen Stars Script

Combine Stars and Starless using Bill Blanshen NB combination script

Contrast tweak in curves. Export to Photoshop

In Photoshop:
Contrast adjustments in Camera Raw
Adjust color hues and individual saturation in Camera Raw

Thats excellent. The colours look great. I've never really been able to get on with Foraxx - I need to try again.

[Graeme1858]

Here's my SHO attempt.

3 x ABE - Function degree 1
Channel combination SHO to RGB
Starnet 2
Stars - Invert, SCNR, Invert - Curves - Reduction, Saturation
Starless - Curves - S Curve, Saturation, Hue
Pixelmath recombination
SCNR - 50% green
Save as .jpg file
IrfanView - Reduce to 1920 x 1080

I didn't do any noise reduction, none required! Perhaps a bit over saturated but it has brought out the orange hues.

Again, excellent data Greg.

[Jockinireland]

This was, yet again, fantastic data - thanks for sharing Greg.

I decided on SHO with RGB stars and started with the XISF files. I know that green in SHO is a bit out of fashion but I like it as it is more remiscent of the original Hubble images - think pillars of creation. Although I have toned the green down quite a bit.

I'm linking to the image in my non-public AB (click the thumbnail) because it really deserves to be seen at decent resolution -even on AB its downsized quite abit.



Starting with the SHO images
GraXpert on each
Channels Combined as SHO
BluXterminator (correct only)
SPCC
BlurXterminator
StarXterminator

On starless
Blanshans HTStretch unlinked
GHS to brighten and contrast (Mode RGB)
MLT Sharpen
3 iterations of LHE (amount 0.150, Kernal rad's of 26, 126 and 200)
Noisexterminator
Took to Photoshop to balance backgroung, tweak colours in CameraRaw and did a selective high pass sharpen to make some of the structures pop
Back to PI

On RGB
Graxpert on each
Channel combination
BlurXterminator
SPCC
Gentle HT stretch (just enough to show stars
removed stars using Blanshans ScreenStars

On stars
GHS stretch (mode colour)

Combined RGBStars with starless SHO using Blanshans ScreenStars

back to PS to resize

I have left the stars more prominent than I usually do because I thought the colours were good and the starfield at top left adds interest to an otherwise sparse area.

This was great fun and I am pretty happy with the final image

[AstroBee]

Very nice images so far. It's really cool to see other people's takes on your own data. So many variations... And it's only April 3rd!

[starfield]

What a great data set, Greg. This is going to be fun to play with. First up I thought I'd try something different and go with the FORAX palete. Went for a high contrast look with this. Ended up with this:

Forax@0.33x.jpg

SHO Processing:

On each SHO master:
BlurX Correct Only
DBE
BlurX Sharpen Stars and Nebulosity.
Light NoiseX
StarXterminator
Histogram Midtone Stretch on each channel
Combine in Forax Script Utility
Unsharp Mask
Lite NoiseXTerminator (not much noise here)
Curve tweak for contrast.

RGB
BlurX Correct Stars
DBE
Midtone stretch to get good stars.
Remove stars using Screen Stars Script

Combine Stars and Starless using Bill Blanshen NB combination script

Contrast tweak in curves. Export to Photoshop

In Photoshop:
Contrast adjustments in Camera Raw
Adjust color hues and individual saturation in Camera Raw

Thats excellent. The colours look great. I've never really been able to get on with Foraxx - I need to try again.

Forax is fairly new to me. Sometimes I find things turn out a bit too orange for my taste, but here it turned more of a dark rust color. For next go round I'm tempted to try using the narrowband color mapper script for a sort of H+S, O+H, O blend. I saw Adam Block demonstrate the technique on the Helix Nebula and thought it might be fun to try.

[Graeme1858]

If anybody would like to contribute to a future processing challenge please post here or PM me.

See how others interpret your data!

Graeme


@Gordon @KathyNS @AstroBee @KingClinton @JockinIreland @Pepper @Yobbo89 @scribbly @Pavebo @gcisko @dcrowson @bernhardhubl @Petrol @bobharmony @Gerald_db @sdbodin @andrewsscope @jthommes @BigKahuna @lostone @MariusD69 @avdhoeven @XCalRocketMan @OzEclipse @Juno16 @ewomack @carastro @ScottFwashere @jnelson @starfield @Rasalgethi

[yobbo89]

This was, yet again, fantastic data - thanks for sharing Greg.

I decided on SHO with RGB stars and started with the XISF files. I know that green in SHO is a bit out of fashion but I like it as it is more remiscent of the original Hubble images - think pillars of creation. Although I have toned the green down quite a bit.

I'm linking to the image in my non-public AB (click the thumbnail) because it really deserves to be seen at decent resolution -even on AB its downsized quite abit.



Starting with the SHO images
GraXpert on each
Channels Combined as SHO
BluXterminator (correct only)
SPCC
BlurXterminator
StarXterminator

On starless
Blanshans HTStretch unlinked
GHS to brighten and contrast (Mode RGB)
MLT Sharpen
3 iterations of LHE (amount 0.150, Kernal rad's of 26, 126 and 200)
Noisexterminator
Took to Photoshop to balance backgroung, tweak colours in CameraRaw and did a selective high pass sharpen to make some of the structures pop
Back to PI

On RGB
Graxpert on each
Channel combination
BlurXterminator
SPCC
Gentle HT stretch (just enough to show stars
removed stars using Blanshans ScreenStars

On stars
GHS stretch (mode colour)

Combined RGBStars with starless SHO using Blanshans ScreenStars

back to PS to resize

I have left the stars more prominent than I usually do because I thought the colours were good and the starfield at top left adds interest to an otherwise sparse area.

This was great fun and I am pretty happy with the final image

you don't use deconvolution ? your data seems pretty reasonable any how ie star fwhm , guessing you don't have large jet streams overhead and deconvolution isn't much needed

[Jockinireland]

This was, yet again, fantastic data - thanks for sharing Greg.

I decided on SHO with RGB stars and started with the XISF files. I know that green in SHO is a bit out of fashion but I like it as it is more remiscent of the original Hubble images - think pillars of creation. Although I have toned the green down quite a bit.

I'm linking to the image in my non-public AB (click the thumbnail) because it really deserves to be seen at decent resolution -even on AB its downsized quite abit.



Starting with the SHO images
GraXpert on each
Channels Combined as SHO
BluXterminator (correct only)
SPCC
BlurXterminator
StarXterminator

On starless
Blanshans HTStretch unlinked
GHS to brighten and contrast (Mode RGB)
MLT Sharpen
3 iterations of LHE (amount 0.150, Kernal rad's of 26, 126 and 200)
Noisexterminator
Took to Photoshop to balance backgroung, tweak colours in CameraRaw and did a selective high pass sharpen to make some of the structures pop
Back to PI

On RGB
Graxpert on each
Channel combination
BlurXterminator
SPCC
Gentle HT stretch (just enough to show stars
removed stars using Blanshans ScreenStars

On stars
GHS stretch (mode colour)

Combined RGBStars with starless SHO using Blanshans ScreenStars

back to PS to resize

I have left the stars more prominent than I usually do because I thought the colours were good and the starfield at top left adds interest to an otherwise sparse area.

This was great fun and I am pretty happy with the final image

you don't use deconvolution ? your data seems pretty reasonable any how ie star fwhm , guessing you don't have large jet streams overhead and deconvolution isn't much needed

Sadly it's not my data, it's Greg's and he's out in Nevada so pretty good skys I think. Here in ireland my sy is reasonably dark B4 but awful in every other way.

But no I don't use deconvolution anymore even on my own data. I use blurXterminator pretty much exclusively because i find it does everything deconvolution does but with much less of the work. I hear all of the anti-AI sentiment but honestly i dont listen to all that esoteric stuff.