26 replies to this topic

[martinclayden]

I've attempted to download both Cosmic Clarity and Statistical Stretch. Unfortunately both of these potentially very useful utilities are detected as possible Viruses, by Google, Norton AV and Windows Defender.

 

The specific virus report by Norton AV is heur.advml.b

Has anybody else seen this, is there a fix, can the warnings be ignored?

Martin

Edited by martinclayden, 27 October 2024 - 08:48 AM.

[Phil Perry]

Where are you getting them from? Is it a reputable distribution site? I would be leery of ignoring AV warnings, especially if multiple utilities are flagging it. Best to go with the authors' site(s), rather than some random collection that may have been compromised.

[martinclayden]

I'm getting both tools from the Seti Astro site. I messaged Frank (the author) and he suggested whitelisting the apps or even sending me the Python code to run locally. While this would work, it still leaves a doubt/risk imo. Hence I'm asking the forum for an opinion.

Edited by martinclayden, 27 October 2024 - 09:03 AM.

[Skywatchr]

I've attempted to download both Cosmic Clarity and Statistical Stretch. Unfortunately both of these potentially very useful utilities are detected as possible Viruses, by Google, Norton AV and Windows Defender.

 

The specific virus report by Norton AV is heur.advml.b

Has anybody else seen this, is there a fix, can the warnings be ignored?

Martin

https://search.brave...c0e29c7d456cba3

 

Conclusion

Heur.AdvML.B is a heuristic detection designed to identify potential malware using advanced machine learning technology. While it may not always be accurate, it’s essential to verify the file name and path, exclude it from future scans if harmless, and maintain good antivirus software and system hygiene practices to minimize the risk of detection.

 

Emphasis added by me.

 

I would whitelist it.  Your anti-malware software will still detect something anywhere/everywhere else on your system should it be something "malicious"..

 

And...

 

https://learn.micros...on-on-a-newly-c

 

There seems to be a bit of a "bug" in the detection routines.

[SteveCollyer]

It's fine.

I've used Windows Defender along with Malwarebytes and the Seti Astro files are clean.

 

Would I would suggest is getting rid of Norton AV as soon as you can.  It's ridiculously clunky, and sucks up a ton of PC performance. 

You may need to use https://www.revounin...eware-download/ or similar to properly remove Norton.

Edited by SteveCollyer, 28 October 2024 - 06:30 AM.

[cpl43uk]

Sadly I have tried to load Cosmic Clarity programmes (both Mac version and PC-mode using Parallels) and whilst they download into the folder structures as requested (after authorising them), along with v5 of the models, when I try to run them (I hit submit button with a 32-bit tif file sitting in the Input folder taken from SiriL) the flash screen flashes past at speed - no obvious processing taking place - and nothing in the output file. This is true if I just use the test image sitting in the input (having deleted it from output folder).

 

Not at all sure what the problem might be....

[Tapio]

Has anyone been successful in downloading Cosmic Clarity lately?

I have tried but everytime it says something like too many users have downloaded, try again in 24 hours.

[RodgerDodger008]

Yep from his website and yes you get virus warnings, if you trust the source ignore. Cuiv said it was fine so that was good enough for me. Not tried it much yet.

[Tapio]

Well I tried with Edge this time and it started to download - not with Firefox

[RodgerDodger008]

What I can say about this software as a standalone application at least is that it seems incredibly slow. I probably do not have the fastest pc out there but I ran a comet image through neat image a hour ago as a control image, put the same image in the in folder of cosmic clarity denoise and it is still going. It would have to be amazing for me to be will to wait this long. But I will compare the two the results. 

 

The issue with the slowness is not so much that, it is you need to set the level of denoise or sharpening, then wait a hour at least before you can tell if you overcooked it or not. At least in standalone this is probably not for me.

[kg7]

Super buggy so far.  It downloads just fine, and it does run some of the time.  It seems like windows has difficulty reading the input and output folders, as if they're constantly being scanned, which makes it difficult to load files up or look at results the ~20% of the time that it actually runs.  80% it just crashes immediately.

 

When it does work, it definitely makes a difference.  Even with default settings stars are smaller and details are sharper.  Following Cuiv's advice I'm running it on prestretched images, and I find that I often have to restretch them (back down) after Cosmic Clarity runs because it lightens them too much.  That could just be a settings issue.  I'm running the V5.2 models released 10/31/24.

 

I'm still not totally convinced it's not malware, either intentionally or not, because my entire file system is buggy after running it the first time.  Folders don't open, or don't close correctly.  System hangs after multiple reboots, all kinds of sudden problems.  Taking my system entirely offline seems to fix the problem, which is worrying.

Edited by kg7, 31 October 2024 - 11:58 PM.

[RodgerDodger008]

Maybe thats your anti virus software kg7, runs fine on mine no scanning etc but I only use windows defender and windows firewall. Gave up on norton and macafee a while ago. But it is incredibly slow.

You may need to mark it as trusted software somehow for your anti virus to stop freaking out.

Edited by RodgerDodger008, 01 November 2024 - 12:20 AM.

[SteveCollyer]

The entire Sharpen/Denoise process takes about 10-12 minutes or so on my PC.  My system just has a basic Intel GPU as well.

[TDPerry]

Has anyone been successful in downloading Cosmic Clarity lately?

I have tried but everytime it says something like too many users have downloaded, try again in 24 hours.

That happens when it is hosted on a Google Drive.  They only offer so much bandwidth/downloads per day to files hosted on there.

It is one reason that the Meade files (documents and programs) I was able to obtain after their site went silly are hosted on my Google Drive for linking to on my site as they don't get downloaded much.  The stuff like PIPP and similar are actually hosted on my VPS instance and do not have those restrictions.

I'm in the process of downloading the 5.3.1 versions and making them available for direct download from my resources area on my site.  I still have about 50GB of space free, so I guess I can devote a 1/10 of that to hosting those 3 files.  I can't promise that the latest versions will always be there since I don't use it and will probably forget to check regularly.

Edited by TDPerry, 01 November 2024 - 03:10 PM.

[kg7]

Updated review, working much better after resolving some user error.  I would edit my previous post if I was allowed to do so.

 

I solved my file access problems by fixing my OneDrive backups, which were apparently creating parallel problems unrelated to Cosmic Clarity.  Files seem to work just fine now.

 

Regarding the reports that it's slow, it seems fine to me.  Denoising is always kind of slow, but the non-PI interface gives you a window with a percentage countdown so it's very tolerable.  Takes maybe 30 seconds per image on my system.  

 

The standalone interface isn't great, because you have to manually move your file into the input folder, run CosmicClarity, then navigate to the output folder, copy the result and paste it back into the input folder and run CosmicClarity Denoise.  This does allow you to play with the settings for each step, but once you have those dialed in it would be nice to run them both sequentially from the same window.

 

And the results are definitely promising.  As an example, this was my entry into the monthly imaging challenge, using my old workflow:

 

 

 

And here are some zoomed in crops of the core of M33, the original on the left (from the above) and on the right is a version I reprocessed using CC and CCDenoise (between background extraction and stretching).

 

 

You'll have to excuse difference in color balance, brightness, etc because this was an entirely new reprocess of the same data.  I just manually cropped the centers of each image so they're not a pixel perfect match but they're pretty close.  I think the improvements in star size and shape are pretty obvious, the dust lanes are at least nominally improved, and I think the denoising is better than what I previously got out of Graxpert's 3.0.2 AI denoise.  The original image was about a 50% crop from a C8 hyperstar system with an ASI2600mc.

 

If I really stare at these two images, I can see some of the same problems that BXT has, in terms of a seemingly random determination of what is a tiny star to delete as background noise and what is noise to enhance into a tiny star, but it seems much less aggressive about making things up than BXT is.  I will continue to play with the settings to see if I can't improve my results.

 

But even this first pass, with just the default settings, makes it pretty clear that CC and CCdenoise are definitely superior to my old workflow of deconvolution in Siril and denoising in Graxpert.  And it's hard to argue with free.  Thanks, dev!

Edited by kg7, 02 November 2024 - 12:39 AM.

[RodgerDodger008]

Thanks kg, ok i will try and solve my slowness issue. Not sure if it is to do with the file type im using png, the fact that it is a eaa live stacked and pre stretched image or my pc. But glad it you have found it to be fast that means I can solve that issue.

[kg7]

More Cosmic Clarity testing this afternoon.  Data are 3.5 hours of M27 taken with an l-enhance filter on a C8 at the native f/10.  Tracking and focus were not very good, so this is data that I normally wouldn't have bothered to process multiple times, but I'm trying see what CC can salvage.  5 minute subs were first stacked in Siril and then background extracted with Graxpert.  

 

Then my normal workflow would be to denoise in Graxpert, and process normally (color calibration, star extraction, stretching, recombination).  I did no deconvolution. This is a 100% zoom crop of the nebula to show as many problems as possible.

 

Next I took that same data after graxpert background extraction and applied the Cosmic Clarity default settings for both sharpening and denoising (instead of Graxpert denoising) and then repeated the exact same processing steps.  CC applies a slight stretch and color saturation, so I tried to balance those out to match the original image above.

Notice that there are many more stars visible, and some of them aren't visible at this stretch in the original fuzzy image at all (but they are all in fact present, as they're visible in other processing attempts on this data, so it's not like CC made them up).  I'm also not very happy with the way the nebula interior was sharpened.  This is probably user error on my part as I figure out how to use this tool.  The stars also acquired far too much color and had to be desaturated.

 

At least with this dataset, I can see the potential of Cosmic Clarity to improve dodgy data but I'm not very happy with the results this time.  Better?  Yes.  Good?  Not yet.

Edited by kg7, 02 November 2024 - 04:48 PM.

[RodgerDodger008]

KG I also experimented but toned down the default settings quite a bit, if you do that I think you will be happier. More stars definitely appeared for me but when comparing to the orginal they were all there. Could not find details anywhere that were not in the ogirinal.

Edited by RodgerDodger008, 02 November 2024 - 04:51 PM.

[darkcloud]

Just tried the PI scripts.  I get an error from both the denoise and the sharpen scripts that indicate they can't find the file.   I think they have assumed a filetype of .tiff?

[kg7]

Just tried the PI scripts.  I get an error from both the denoise and the sharpen scripts that indicate they can't find the file.   I think they have assumed a filetype of .tiff?

Yes, the documentation says it needs to be a 32 bit tiff file.  For the standalone version I'm using outside of PI, if you don't have a tiff in the input folder the script takes about half a second to start up and then just aborts without doing anything.

[kg7]

A friend who shoots with a DSLR and camera lens sent me his most recent M33 for advice.  He was unhappy that he forgot his Bahtinov mask this night, and his focus wasn't quite right.  This is a heavily cropped in view of a wide angle shot of Triangulum:

 

I ran the Cosmic Clarity sharpen and denoise (still free!) standalone utilities with default settings.  This is run on a jpeg (reconverted back to tiff for CC) after he did his processing to what he considered a final image.  I think it makes a pretty dramatic difference, even with images like this one at a large pixel scale and mediocre focus.

 

To my eyes the deconvolution looks as good as what BXT does, without the artificial rerounding that often shows up in BXT.  I have no idea what BXT is doing under the hood but it sure looks like it's star-masking and resynthesizing the larger stars that don't deconvolve very well.  That means fake data, and Cosmic Clarity is just doing the deconvolution part and is thus a bit more faithful to what's really in the sky.  

 

The more I play with it, the more convinced I become that everyone who's not using BlurXterminator should be using Cosmic Clarity in its place.  I have found the results to be notably better than what I was getting out of AstroSharp/AstroClean, the only other free utility that I know about that does the same thing.

Edited by kg7, 08 November 2024 - 03:57 PM.

[RodgerDodger008]

Good review, btw I have also used png files on the standalone and it accepted those too.

[MikeGH]

Hi,

regarding Cosmic Clarity, I have downloaded and extracted files as instructed but cannot save any of my 32-bit TIFF images in the input file. I am told I do not have administrator rights / permissions.  Does anyone know what I might be doing wrong, please? I can run the clarity program on the test image that came installed with the downloaded program.

 

Regards,

Mike

[flamidey]

You will find excellent support on Seti Astro's Discord channel over here https://discord.gg/tyZPC3hj

[MikeGH]

Thanks very much.  In fact I had extracted the program files into my Program files file which was a mistake.  I've left them in downloads and it works well.  In fact Cosmic Clarity and noise reduction works very well, but from the few trials I have made I cannot get such good results with satellite removal.  I need to test a few more images and probably I am doing something wrong. So far the trails are still there but are black rather than white.

Still for free software it's amazing and probably constantly improving.