Linux Version of Cosmic Clarity Sharpen and Denoise compiled!!! The stand alone is up and running. I still have to fully vet the script interface with PixInsight but that shouldn't stop anyone from using the standalone! Site: www.setiastro.com/cosmic-clarity Drive: drive.google.com/drive/folders/1hB2qeMF7YQH1-Vr81udvNfd_NB6RKBT0?usp=drive_link
Amazing, it does a marvelous job on stars, rounded without black halos, and good sharpening of the DSOs. You are improving a lot our processing capabilities, thanks.
Exceptional piece of software, it's very kind of you to provide this to the community :D Been testing it on a number of objects taken with my 200mm f/4 newt and it really does a fantastic job on the stars, what I noticed is that it puts some little "stars" on the airy discs or spikes, also some small galaxies get a stellar nucleus instead of the original fuzzy cores, appart from that it's fantastic!!!
Very impressive. Just found your channel and can’t tell you how impressed I am with your content. You should be proud of this tool. I have no doubt you can refine this to become the gold standard.
This is great news for those of us that try to keep their astrojourney in a budget -and thus don't own PixIn or BXT. Thank you so so much for this. (At the moment i post this comment I'm running CosmicClarity on some of my photos)
Thank you for your hard work! It would be great if the standalone version of the program works with linear data, and for PixInsight could download the script in the archive.
Many thanks for your excellent scripts. With respect to the PI CosmicClarity scripts I propose to let the user choose the location of the input and output folders. Also this would allowe me to install the program to the systems "Program Files" Folder (no write rights), provided no other rquirements prevent this.
Thanks! Yes that would complicate some things especially with the ability to interface via scripting. I am not saying it is impossible but a lot of effort for minimal return at this point. I need to more fully develop a gui and some other items to make it more user friendly first.
Thanks for the amazing work! I suspect I am still using this tool incorrectly. I have tried noth stretched and prestretched I seem to not get good results. Would you recommend using this tool before or after processing the image. (After seems to result in a overexposed look vs the original image)
If using as a stand alone you should run this immediately after stretching. It should not give you an overexposed look. Can you share the image you are having issues with? frank@setiastro.com
Very impressive, it sharpens hot young blue star clusters in galaxies the way it should, with BlurX even if they are resolved /partly resolved, it merges them in to structures that look like worms. This is evident on a lot of images on Astrobin. Even John Hayes mentioned this and was struggling with his latest image.
Over sharpening leading to wormy appwarances happens way too much. Heck even my latest image of ic342 has it becuase of blurx (i processed it before cosmic clarity was ready). I need to reprocess it now with cosmic clarity.
Overall this is an amazing contribution to the image processing arsenal. The non stellar sharpening is truly impressive. Curious if the star sharpening was trained on reflector images as well as refractor ones: the diffraction spikes on bright stars are turning into strings of pearls. Also seeing a little dimming in the centers of bright stars.
It was partially trained on reflector images. I only have a few masters from my newt. If youhave master lights from a reflector i would love to add them to the modelling. frank@setiastro.com
Great work…stars on the edges which are my problem are rounded with your script…terrific. For some reason I can only use version 2 of BXT and the rounding it does is not as good as version 4…so perhaps by following up BXT ver2 with your new sharpening script the stars on the edges will be even better taken care of…thank you.
I have actually found cases where BXT was hallucinating more artifacts in non-stellar regions than CosmicClarity. If you own BXT it is worth running it vs CosmicClarity and just using whichever gives you the best results. Can't hurt. BXT will probably win 95% of the time, but my free tool is there and only going to get better :)
Really great work man! The only difference I could see between yours and bxt was that bxts stars were a little less saturated in my opinion (other than little rounder) But really an amazing tool! 👍
Thanks! I am still getting more data to train, and honestly i like the stars having a little more color. Bxt really can leave a bunch oh white points I am not a fan of. I really need to double my data again and maybe even use another neural net layer, but for a first production version i think it us pretty darn good, and the price is right :)
Hi Frank. Looks great. I am having trouble getting the program to download on my M1 MacBook Pro. Seems there are some incompatible files?? The PI link works, but not program or models.
Hi Frank, I tried to use it on my macbook as well, everything works, exept that there occur some errors when running the program: “Error reading image”, “ValueError: not enough values to unpack (expected 4, got 3)” and “Failed to execute script due to unhandled exception”. Any ideas on how to solve this? I’ve put a tif file in the input folder before running the program.
Because some people like hitting their images with a bit more sharpening sometime after removing the noise, especially at the end of their processing. The ai was also trained on both noisey and noiseless concolved models
Right. So instead of just erroring out it asks if you want to wait longer. That way on slower systems you can wait for the output file so PI can load it back in.
Hi, I downloaded a few times because I haven´t managed to open the program. I double click in the file "setiastrocosmeticlarity" and a little back window appears and then desappears. The same for the file "setiastrocosmeticlarity_denoise". So as I said, I uninstalled the program and install it again. Same result. What could be wrong ?
@@setiv2 Good news, it is working. I may had done a wrong manipulation. I went to check also the Cuiv The Lazy video on your program. Seems ok. Many thanks, sorry if I scared you.
Same for me. I can't tell if I am unzipping the files correctly or running into security issues on my Mac. I know it's bone-headed, but an idiot's guide to installation on a Mac would be quite helpful. I have BlurX and Graxpert, but really wanted to give this script a try. Thanks.
@@setiv2 Still no luck. I running MacOS 16.4.1. I think I will wait for the next OS update and try again. In PI, the script starts, I get a security warning which I allow and then the script just spins. Totally befuddled at this point.
@@michaeljarvis4612 the script calls the folder where cosmicclarity is. So if cosmicclarity isnt working the script wont. Make sure your cosmicclarity is no in a write orotected folder and that you went to privacy and security and enabked it to run
I did everything according to the instructions, but after starting the Cosmic Clarity console appeared and disappeared in a few seconds, later this error appeared "Timeout waiting for file: C:/Users/arrow/Desktop/Latest Models - Updated 9.27.2024\output\b_clone_sharpened .tif Failed to find the output file after waiting" but the output file is there in the folder
Can you make sure the input and output folders are empty prior to running it. Give it another run and see that it says. Specifically if the console opens and startd sharpening
Mmmm, its not producing output for me, got the very latest files, tried reinstalling several time, it goes thru the process but outputs nothing....standalone version on win11
I like cosmic clearity on brighter star because it kept the star like glowing or sparkling as how we see under atmosphere and that how I like to see it on my photos too
Some detractors would say "this isn't really Deconvolution, it's just sharpening..." (Not me mind you) But what would you say to those who do? Nice tool.
Then they would have to say the same for blurx. Anything with a neural net is trained how to reverse something, whether it is aberrations, convolution, contrast enhancement, whatever.
Or graxpert is really just large scale contrast normalization and not gradient removal too in the strictest sense, since we dont know what the true gradient is
@@setiv2 oh for sure. I looked at the root of the definition of the word deconvolution: "the improvement of resolution of images or other data by a mathematical algorithm..." 🤷 Good enough for me!
Beautiful You are not compared to blur exterminator, but to me it is as good as that for my astrophotography level I do definitely will save me a lot of headache and money. Good job thank you so much.
Trying to get it to work but I get this error message in PI. *** PixInsight API Error: MetaFileFormat: TIFF: Invalid or nonexistent directory: C:/Users/***/Documents/PixInsight Database/Latest Models - Updated 9.27.2024/input/masterLight_BIN_1_6248x4176_EXPOSURE_600_00s_FILTER_Ha_mono_drizzle_2x_autocrop_integration_autocrop.tiff *** Error [000]: C:/Program Files/PixInsight/src/scripts/SetiAstroCosmicClarity.js, line 372: Error: Failed to save image as C:/Users/***/Documents/PixInsight Database/Latest Models - Updated 9.27.2024\input\masterLight_BIN_1_6248x4176_EXPOSURE_600_00s_FILTER_Ha_mono_drizzle_2x_autocrop_integration_autocrop.tiff
You need to download the cosmic clarity program first, then use the wrench icon in the script to tell the script where the program is. Be sure to watch my first video where I talk about that. I have since updated the instructions on my website too to remind people to do that.
Linux Version of Cosmic Clarity Sharpen and Denoise compiled!!! The stand alone is up and running. I still have to fully vet the script interface with PixInsight but that shouldn't stop anyone from using the standalone!
Site: www.setiastro.com/cosmic-clarity
Drive: drive.google.com/drive/folders/1hB2qeMF7YQH1-Vr81udvNfd_NB6RKBT0?usp=drive_link
Thanks. Excellent job on the Linux. tested and it's working fine under kubuntu 24.04. Keep up the great work.
This is bloody brilliant! I can't believe this is free, too! You, ami, are an amazingly positive force in the astrophotography community. Well done!
Thanks a bunch!
You have every right to be proud! This is an awesome tool already!
Thank you for donating your talent to all of the astrophography community. Your tools are excellent.
Amazing, it does a marvelous job on stars, rounded without black halos, and good sharpening of the DSOs. You are improving a lot our processing capabilities, thanks.
Congratulations. Very good work, you should be very proud.
Amazing progress! You SHOULD be proud!
Exceptional piece of software, it's very kind of you to provide this to the community :D
Been testing it on a number of objects taken with my 200mm f/4 newt and it really does a fantastic job on the stars, what I noticed is that it puts some little "stars" on the airy discs or spikes, also some small galaxies get a stellar nucleus instead of the original fuzzy cores, appart from that it's fantastic!!!
Amazing work! I would love to see a video at some point on how you train the network. This is fascinating.
Awesome job Frank.
Very impressive. Just found your channel and can’t tell you how impressed I am with your content. You should be proud of this tool. I have no doubt you can refine this to become the gold standard.
Glad you found it! Welcome to the group!
Awesome job! I will download v2 after work and try it out! Many thanks!!🙏
Really great work. Thanks for providing so many useful tools.
Thank you very much for this great tool !
This is great news for those of us that try to keep their astrojourney in a budget -and thus don't own PixIn or BXT. Thank you so so much for this.
(At the moment i post this comment I'm running CosmicClarity on some of my photos)
Awesome work.
Thank you for your hard work! It would be great if the standalone version of the program works with linear data, and for PixInsight could download the script in the archive.
What do you mean by download the script in the archive?
@@setiv2 ZIP archive for offline download of the script.
@@setiv2 ZIP archive for offline script download.
Impressive job on this tool.
Many thanks for your excellent scripts. With respect to the PI CosmicClarity scripts I propose to let the user choose the location of the input and output folders. Also this would allowe me to install the program to the systems "Program Files" Folder (no write rights), provided no other rquirements prevent this.
Thanks! Yes that would complicate some things especially with the ability to interface via scripting. I am not saying it is impossible but a lot of effort for minimal return at this point. I need to more fully develop a gui and some other items to make it more user friendly first.
Thanks for the amazing work!
I suspect I am still using this tool incorrectly.
I have tried noth stretched and prestretched I seem to not get good results.
Would you recommend using this tool before or after processing the image. (After seems to result in a overexposed look vs the original image)
If using as a stand alone you should run this immediately after stretching. It should not give you an overexposed look. Can you share the image you are having issues with? frank@setiastro.com
Very impressive, it sharpens hot young blue star clusters in galaxies the way it should, with BlurX even if they are resolved /partly resolved, it merges them in to structures that look like worms. This is evident on a lot of images on Astrobin. Even John Hayes mentioned this and was struggling with his latest image.
Over sharpening leading to wormy appwarances happens way too much. Heck even my latest image of ic342 has it becuase of blurx (i processed it before cosmic clarity was ready). I need to reprocess it now with cosmic clarity.
Overall this is an amazing contribution to the image processing arsenal. The non stellar sharpening is truly impressive.
Curious if the star sharpening was trained on reflector images as well as refractor ones: the diffraction spikes on bright stars are turning into strings of pearls. Also seeing a little dimming in the centers of bright stars.
It was partially trained on reflector images. I only have a few masters from my newt. If youhave master lights from a reflector i would love to add them to the modelling. frank@setiastro.com
Great work…stars on the edges which are my problem are rounded with your script…terrific.
For some reason I can only use version 2 of BXT and the rounding it does is not as good as version 4…so perhaps by following up BXT ver2 with your new sharpening script the stars on the edges will be even better taken care of…thank you.
good to know! maybe BXT for non stellar features, and Cosmic Clarity for the star mask!
I have actually found cases where BXT was hallucinating more artifacts in non-stellar regions than CosmicClarity. If you own BXT it is worth running it vs CosmicClarity and just using whichever gives you the best results. Can't hurt. BXT will probably win 95% of the time, but my free tool is there and only going to get better :)
@@setiv2 ….by the way Frank, do you use version 4 or 2 of BXT?…for some reason I cannot get ver 4 to work.
Version 4
Really great work man! The only difference I could see between yours and bxt was that bxts stars were a little less saturated in my opinion (other than little rounder)
But really an amazing tool! 👍
Thanks! I am still getting more data to train, and honestly i like the stars having a little more color. Bxt really can leave a bunch oh white points I am not a fan of. I really need to double my data again and maybe even use another neural net layer, but for a first production version i think it us pretty darn good, and the price is right :)
Hi Frank. Looks great. I am having trouble getting the program to download on my M1 MacBook Pro. Seems there are some incompatible files?? The PI link works, but not program or models.
Shoot me an email and we can figure it out. I know for a fact othet mac users have downloaded and used cosmic clarity
Hi Frank, I tried to use it on my macbook as well, everything works, exept that there occur some errors when running the program: “Error reading image”, “ValueError: not enough values to unpack (expected 4, got 3)” and “Failed to execute script due to unhandled exception”.
Any ideas on how to solve this?
I’ve put a tif file in the input folder before running the program.
@@duncanmatse4345 can you email me? frank@setiastro.com
Thanks so much for this effort. your tool should be used before noise reduction. I’m curious why you show examples after using a noise reduction tool.
Because some people like hitting their images with a bit more sharpening sometime after removing the noise, especially at the end of their processing. The ai was also trained on both noisey and noiseless concolved models
Charity is doing a fanstatic job, but you seem to be putting round circles in the center of stars compared to BXT!
Can you give me an example? I havent seen that on any of my data. frank@setiastro.com
Can you elaborate?
For being for free more then good enough
Great job .
Hopefully I’ll learn how to use it.
I did the up date but now I getting a error message . The output file was not found after 120 sec. Do you want to keep waiting.
Thanks Steve
Right. So instead of just erroring out it asks if you want to wait longer. That way on slower systems you can wait for the output file so PI can load it back in.
@@setiv2ok got fixed. Just needed to redownload and install. System is not slow I ‘m slow. LOL
Question
Do I use your program after my final work as I use
Astap,graxpert,siril,and Ps
You would use it during your siril phase, probably right after your stretch it to be nonlinear
@@setiv2
I do stretch after I remove the stars With starnet ++
Do I use your on star mask or do it after I put it together stars on Ps
Hi, I downloaded a few times because I haven´t managed to open the program. I double click in the file "setiastrocosmeticlarity" and a little back window appears and then desappears. The same for the file "setiastrocosmeticlarity_denoise". So as I said, I uninstalled the program and install it again. Same result. What could be wrong ?
What operating system are you using? Where did you extract the folder to? Did you m9ve anything out of the folder once you extracted it?
@@setiv2 Good news, it is working. I may had done a wrong manipulation. I went to check also the Cuiv The Lazy video on your program. Seems ok. Many thanks, sorry if I scared you.
Hi! There is no chance to download windows version. Google indicated that quota for this file gone and recommend to tell Owner
There isnt a quota limit. Probably it was too many people trying to access at once. Just give it another try.
@@setiv2 thanks, I tried many times, but Google indicate that the file is bloked for a 24 hours
drive.google.com/drive/folders/1z4Nh9M3Jz6JT2pZrYV2Y-IP-R5ouBVjS?usp=drive_link
Same for me. I can't tell if I am unzipping the files correctly or running into security issues on my Mac. I know it's bone-headed, but an idiot's guide to installation on a Mac would be quite helpful. I have BlurX and Graxpert, but really wanted to give this script a try. Thanks.
New Version 5.3.1 is out. I bet it had to do with the hidden file in the input folder. Give the new version a try.
@@setiv2 Still no luck. I running MacOS 16.4.1. I think I will wait for the next OS update and try again. In PI, the script starts, I get a security warning which I allow and then the script just spins. Totally befuddled at this point.
@@michaeljarvis4612 the script calls the folder where cosmicclarity is. So if cosmicclarity isnt working the script wont. Make sure your cosmicclarity is no in a write orotected folder and that you went to privacy and security and enabked it to run
I did everything according to the instructions, but after starting the Cosmic Clarity console appeared and disappeared in a few seconds, later this error appeared "Timeout waiting for file: C:/Users/arrow/Desktop/Latest Models - Updated 9.27.2024\output\b_clone_sharpened .tif
Failed to find the output file after waiting" but the output file is there in the folder
Can you make sure the input and output folders are empty prior to running it. Give it another run and see that it says. Specifically if the console opens and startd sharpening
@@setiv2 Yes, the folders are empty...I will contact you via "contact" from your page and send you a video via Google Drive
Mmmm, its not producing output for me, got the very latest files, tried reinstalling several time, it goes thru the process but outputs nothing....standalone version on win11
Can you share the file you are trying it on? Also where do you have the program saved?
@@setiv2 Previous version was working great, deleted all files and reinstalled latest, its on a dedicated astro ssd
@@HmSGhengis can you run it in standalone mode and take q screen shot of the terminal when it finishes? I want to see everything the terminal printed
@@setiv2 Sorted,, FYI it was run in standalone mode, works with .tiff not the .fits that nina is generating
@@HmSGhengis that is correct. PNG, JPG, TIFF are all acceptable
Let me fiddle with it some more. I am sure its a user error on my end.
I like cosmic clearity on brighter star because it kept the star like glowing or sparkling as how we see under atmosphere and that how I like to see it on my photos too
Some detractors would say "this isn't really Deconvolution, it's just sharpening..."
(Not me mind you)
But what would you say to those who do?
Nice tool.
It was trained on reversing convolution... so it is absolutely deconvolution.
Then they would have to say the same for blurx. Anything with a neural net is trained how to reverse something, whether it is aberrations, convolution, contrast enhancement, whatever.
Or graxpert is really just large scale contrast normalization and not gradient removal too in the strictest sense, since we dont know what the true gradient is
And to be fair my title screen says AI based sharpening :)
@@setiv2 oh for sure. I looked at the root of the definition of the word deconvolution: "the improvement of resolution of images or other data by a mathematical algorithm..."
🤷
Good enough for me!
Beautiful
You are not compared to blur exterminator, but to me it is as good as that for my astrophotography level I do definitely will save me a lot of headache and money.
Good job thank you so much.
Trying to get it to work but I get this error message in PI.
*** PixInsight API Error: MetaFileFormat: TIFF: Invalid or nonexistent directory: C:/Users/***/Documents/PixInsight Database/Latest Models - Updated 9.27.2024/input/masterLight_BIN_1_6248x4176_EXPOSURE_600_00s_FILTER_Ha_mono_drizzle_2x_autocrop_integration_autocrop.tiff
*** Error [000]: C:/Program Files/PixInsight/src/scripts/SetiAstroCosmicClarity.js, line 372: Error: Failed to save image as C:/Users/***/Documents/PixInsight Database/Latest Models - Updated 9.27.2024\input\masterLight_BIN_1_6248x4176_EXPOSURE_600_00s_FILTER_Ha_mono_drizzle_2x_autocrop_integration_autocrop.tiff
You need to download the cosmic clarity program first, then use the wrench icon in the script to tell the script where the program is. Be sure to watch my first video where I talk about that. I have since updated the instructions on my website too to remind people to do that.
@@setiv2 Got it. I misunderstood your instructions. I thought the app was for the stand-alone users.
@@NevadaDesertSkies nope everyone needs it. I have updated better onstructions on my website now too.