No matter what improvements are made, some people will always complain. People buy BXT to fix their stars. It now does a better job of fixing stars. End of. If you don't like it and want to be a purist, you do you. This is the best version of BXT we've ever had. I just tested it out on some old data and it blew me away. Well done.
Well, you cant be a "purist" if you have bad stars to start with. Stars are points of light and not stretched in real life. So with using BXT you get closer to reality.
@@matej.mlakarAbsolutely! Most people I see complaining are people who are literally jealous with some sort of "gate keeping" complex haha. They all be like:"I bought my super expensive equipment and spend countless sleepless nights for a single picture to be ahead of everyone else out there! I perfected my tilt issues and it took me weeks, now others get similar results as me, with less expensive equipment, and they dont have to spend weeks doing what I did, mimimi" :D Oh and they call the results "fake", yet use manual sharpening, deconvolution, star and noise reduction tools and use clone stamps. But when when you can do all that with a single click? Hey, thats cheating! :D Even with BXT, its "bad data in, bad data out". And BXT brings your images closer to reality, just as you said. :) It's almost like amateur Astrophotography is a hobby, not a competition, eh? :D
Excellent summary as usual. I liked the extra info at the end. I would have never thought about using correct only and then combine channels and run BX again. Thanks Sascha!
WOW. First of all, awesome video on this update. Second, I just took my absolute worst set of subs I've ever taken and stacked them. The stars weren't even egg shaped, they were little tridents lol. The result is 95% flawless. Unreal.
Thanks Sascha, great explanation of this amazing upgrade to an already essential AP tool. I'm exclusively shooting OSC so the workflow is certainly more straightforward for myself than for some others. Quickly ran some NGC7000 data through it and it's remarkable how much better it looks even at such an early stage. I suspect I'll be reprocessing old data for the foreseeable future! Clear skies everyone.
Thank you very much for the heads up - really nice to see v2 addressing some of the issues with v1. I did notice similar issues in my images too especially with over sharpening for some objects and the dark rings appearing usually around galaxies. Now it would be about time to re-visit some images :)
I do the same as we all do. It’s the journey of exploration and pride you feel when you are able to see your first images. It’s true you can do it the easy way, but it cheapen our experiences. I will continue to fight the insects and ice for my own experience…
Great presentation as usual! One correction though, you stated near the end that you didn't have to do a manual psf anymore, but that is not quite the case. The manual psf option is still available for cases where there are not many stars in the image (long focal length) In these cases Russ still recommends doing a manual psf calculation.
Thanks a lot - great information. It is just that until now there was a general recommendation to always manually define the psf, this is not the case anymore. But good to know that there is still a situation where it is needed!
Great introduction to the new great piece of software! I've question in the movie you said that the new BXT should be used with colour images, but what to do with LRGB data. I should apply BXT to RGB stack and separately to L stack or do something different?
Hi Sascha. Great news. Something is not clear for me. When I have the Sii, Ha and Oiii channels, do I first have to combine them with LRGBcombination before using BlurXT2? And do I have to split channels again before processing the narrow band channels further? Like Foraxx etc. How do I split them again in Sii, Ha and Oiii?
That is a very good question which also is not 100% clear to me. This is one of these questions which I will send to Russ and come back with an answer!
I completed the upgrade and at first use, selected AI version 4, and ran the process. The next time I opened BX, I was asked to select AI 4 again. How do I make it stick?
Not celebrating just yet - I’ve been using a Bill Blanshan workflow where I stretch before doing LRGB Color Combination & his Narrow Band Normalization. Some adjustment in this will be required to Color combine linear, then BXT, then stretch (with Bill’s stretch or GHS), then NBN.
I was definitely excited. Got home, performed the update, pointed to to new version using the AI button then ran it on my new data of the Jellyfish nebula and fail. Tried to restart, nothing. Pointed back to the last version, works perfectly. Hoping to try to get it worked out. 😢
Thank you. Actually, on your post on Astrobin, someone else had the same issue. Seems Russ was aware and has a fix on his FAQ page. Seems it was an issue with a dll file that was outdated in Pixinsight. Funny because I just downloaded the latest version. Anyway, I downloaded the file he gave, placed it the Pixinsight bin folder and it worked.
So, what was FWHM of good image without Blur XT? My guess it was about 4. I run similar evaluation. Original image 1.8” (only Lum) after Blur XT v1, I got 1.5”, v2 gave 1.35”. Some improvement, but I could not clearly spot it with naked eyes. Bottom line, excellent tool for lousy pictures. Work on acquired data quality.
If there would be something like a "Astrophotography-Drillinstructor" - it would be you. I literally see you standing in front of me, shouting in my face: "It this pathetic attempt of an astrophoto the best you can do? Back to your telescope and work on acquired data quality!!!!!!!!" 🤣
Thank you Sascha! Have you tried this on moon pics yet? I've used your technique of sharpening lunar images with BXT to great effect. But now the issue is that these are always already stretched coming out of Autostakkert.
I understood - no, will not happen - BXT will not merge stars, as from how deconvolution works it has rather the tendency to break it apart than to merge
wow, comes as a free upgrade for people that have already paid for it. That is not very common anymore. Thanks for the highlight, I might have missed that.
I tried it on a few different Lum integrations from targets I've already processed with the previous version, and in every case, though the non-stellar sharpening is significantly increased/improved, stars aren't reduced as much using the same settings as before. Anyone else seeing that?
First of all, I would say non-stellar sharpening is improved - not increased. It was probably over sharpened before in V1. When it comes to stars it is the same story - the are much better corrected, but rather improved than reduced. Which makes sense as BXT is not a star reduction software. But obviously you always increase the star sharpening to achieve the same reduction effect as before.
@@viewintospace Ok, I'll accept more so improved over increased, but also some of both. I don't have the images handy, but one comparison I did was on NGC 2146 and while the the center dust lane was definitely improved though actually wider/thicker than the V1 result, sharpening in other areas of the galaxy was 'increased' in the sense that those areas were made less blurry than V1. Regarding the stars, my comment was more an observation than a criticism. I thought you had seen a reduction in FWHM where I saw the opposite. On another note, I know Russ has said this is a single application tool, not one you use iteratively, but I'm curious why not, in theory at least. I've recently experimented with multiple iterations at less aggressive settings with positive results. BTW, thanks for a great video!!
No, he should not (and he will not btw...). APP is great for stacking but not a processing software at all, and Photoshop might be ok for post processing but also is not equipped at all for adequate Astrophoto processing. I know PixInsight UI is horrific, but still - there is (presently) no way around it.
Interesting as well because GradientXterminator was one of Russ's first plugins and is still only available in Photoshop. I was hoping he would eventually bring it over to PI, but it hasn't happened yet. And with graXpert working so well now maybe it's not necessary. But part of me still thinks Russ will still tackle gradients in the PI world with an amazing future release!!
@@Mr77pro I think you should tell this part of you to stop thinking... 😉 On one side there is GraXpert, but there is also the MARS project of PI. So what we really do not need is a 3rd Gradient Removal Option.
This is going too far. What is the point in buying expensive gear and spending months getting it set up properly, learning to guide properly etc etc when some numpty can grab a cheap scope, take a few unguided blurry images and then use AI to make them look great? Version 1 was fine, it basically made deconvolution, an existing process, a lot easier. This version may not "make up" data but that's not the point, it is certainly transforming bad data into better data to an extent that the final version bears no resemblance to the original data. Very depressing.
People said similar things when we transitioned into digital from film. If anything this introduces more people into the hobby, as it's complexities get solved through software. The people who make spectacular images will co continue to do so.
Here we go again..... 😵💫 We had the same discussions with BXT1..... let's not start again. BXT2 makes actually nebulas more realistic now, as they were oversharpened with BXT1. And about the stars - can we agree that stars are ALWAYS round? So if it makes stars round, it puts it in its correct shape, nothing more.... At the end, it is the same discussions we have with the Seestar and similar devices - more and more, people can create astrophotos easier and with less expensive gear. It sucks for people two spent a fortune on their equipment, but it general it's a good thing!
I agree with midnight lightning on this one. I think it's perfectly okay to have anomalies in your image and I believe this type of image manipulation is just going a little too far. I think it's nice to have tools that help with aberrations but this is almost getting to a point where all images are going to look the same and then it's just a matter of how other people colorize their data afterwards. Doing normal sharpening is fairly basic and it uses your existing data, but we have no idea what is actually happening here. If AI is making determinations on what should an image should look like then that is just false, I would rather just use the data that I have and know that it's real. I know that using software like topaz denoise, you can see new structures being created and I'm beginning to start leaning away from doing any type of AI related processing now and just stick to natural imaging. I am all for technology but this is taking things a little too far just like some of the AI generated Astro images that are being put out on social media right now, which is a joke. I do understand where Sasha is coming from, I don't want to be a hater of new technology but at the same time I am very concerned with software creating things based on What it thinks it should be. I am perfectly fine with aberrations in my image and it only challenges me more to find a way to get rid of them myself versus allowing software to do it for me.
@@billblanshan3021 I see your point Bill! Just thought of a mountain climber reaching the top by climbing the vertical wall. The other person takes a hiking path to the top and the third takes a cabletrain… All three stand finally on the same mountain but everyone is happy - some for their own best effort they spent, some simply for enjoying the view…,
@billblanshan3021 @midnightlightning1 Today Adam Block answered this question perfectly on Facebook: "How is an image that has elongated and blurry stars not a fake image? That isn't really part of the scene and the information is distorted. Literally fake. You might be tempted to say that the elongation and blur is something that is understandable and happens in a regular way (bad tracking, optics, focus...etc). Well..if you can have a reality and distort it... you can, if you are clever enough- do the inverse. The clever part is the hard part in many cases. The point is, there is quite a bit of information in your elongated and blurred stars." - Adam Block, PixInsight for Beginners Forum, Facebook
No matter what improvements are made, some people will always complain. People buy BXT to fix their stars. It now does a better job of fixing stars. End of. If you don't like it and want to be a purist, you do you. This is the best version of BXT we've ever had. I just tested it out on some old data and it blew me away. Well done.
Agree 100% 👍🏻
Well, you cant be a "purist" if you have bad stars to start with. Stars are points of light and not stretched in real life. So with using BXT you get closer to reality.
@@matej.mlakarAbsolutely! Most people I see complaining are people who are literally jealous with some sort of "gate keeping" complex haha.
They all be like:"I bought my super expensive equipment and spend countless sleepless nights for a single picture to be ahead of everyone else out there! I perfected my tilt issues and it took me weeks, now others get similar results as me, with less expensive equipment, and they dont have to spend weeks doing what I did, mimimi" :D
Oh and they call the results "fake", yet use manual sharpening, deconvolution, star and noise reduction tools and use clone stamps. But when when you can do all that with a single click? Hey, thats cheating! :D
Even with BXT, its "bad data in, bad data out". And BXT brings your images closer to reality, just as you said. :)
It's almost like amateur Astrophotography is a hobby, not a competition, eh? :D
Awesome, thanks for the video and the info about 2.0 being released
Great job Shasha! Thanks for the video!
Excellent summary as usual. I liked the extra info at the end. I would have never thought about using correct only and then combine channels and run BX again. Thanks Sascha!
Thanks!
Thanks a lot - very much appreciated!!!!! 🙏🏻😊
WOW. First of all, awesome video on this update. Second, I just took my absolute worst set of subs I've ever taken and stacked them. The stars weren't even egg shaped, they were little tridents lol. The result is 95% flawless. Unreal.
Hello Very interesting and accurate presentation . thanks for your time Well done
RC Astro is goated with the sauce.
Thanks Sascha, great explanation of this amazing upgrade to an already essential AP tool. I'm exclusively shooting OSC so the workflow is certainly more straightforward for myself than for some others. Quickly ran some NGC7000 data through it and it's remarkable how much better it looks even at such an early stage. I suspect I'll be reprocessing old data for the foreseeable future! Clear skies everyone.
Thank you very much for the heads up - really nice to see v2 addressing some of the issues with v1.
I did notice similar issues in my images too especially with over sharpening for some objects and the dark rings appearing usually around galaxies.
Now it would be about time to re-visit some images :)
Given how cloudy it is outside, it is the perfect time to redo all your images 😁
I do the same as we all do. It’s the journey of exploration and pride you feel when you are able to see your first images. It’s true you can do it the easy way, but it cheapen our experiences. I will continue to fight the insects and ice for my own experience…
Great video.
Great presentation as usual! One correction though, you stated near the end that you didn't have to do a manual psf anymore, but that is not quite the case. The manual psf option is still available for cases where there are not many stars in the image (long focal length) In these cases Russ still recommends doing a manual psf calculation.
Thanks a lot - great information. It is just that until now there was a general recommendation to always manually define the psf, this is not the case anymore. But good to know that there is still a situation where it is needed!
Great introduction to the new great piece of software! I've question in the movie you said that the new BXT should be used with colour images, but what to do with LRGB data. I should apply BXT to RGB stack and separately to L stack or do something different?
Yes, I would do it on the combined RGB and on the LUM separately.
I am blown away that BXT is handling star trails of stars that are from either tllt or incorrect spacing!
Hi Sascha. Great news. Something is not clear for me. When I have the Sii, Ha and Oiii channels, do I first have to combine them with LRGBcombination before using BlurXT2? And do I have to split channels again before processing the narrow band channels further? Like Foraxx etc. How do I split them again in Sii, Ha and Oiii?
That is a very good question which also is not 100% clear to me. This is one of these questions which I will send to Russ and come back with an answer!
I completed the upgrade and at first use, selected AI version 4, and ran the process. The next time I opened BX, I was asked to select AI 4 again. How do I make it stick?
It should stick on its own - noting you have to do.
@@viewintospace What I had to do was rename BlurXTerminator.2.pb to BlurXTerminator.2.pb.old and AI ver 4 stuck when the process was opened.
Not celebrating just yet - I’ve been using a Bill Blanshan workflow where I stretch before doing LRGB Color Combination & his Narrow Band Normalization. Some adjustment in this will be required to Color combine linear, then BXT, then stretch (with Bill’s stretch or GHS), then NBN.
So you already figured it out how to modify - so you can celebrate 😊
Wonder if you could colour combine, run bxt then separate the channels and use your old method from there?
I was definitely excited. Got home, performed the update, pointed to to new version using the AI button then ran it on my new data of the Jellyfish nebula and fail. Tried to restart, nothing. Pointed back to the last version, works perfectly. Hoping to try to get it worked out. 😢
Wish you luck to solve the issue!!
Thank you. Actually, on your post on Astrobin, someone else had the same issue. Seems Russ was aware and has a fix on his FAQ page. Seems it was an issue with a dll file that was outdated in Pixinsight. Funny because I just downloaded the latest version. Anyway, I downloaded the file he gave, placed it the Pixinsight bin folder and it worked.
@@markmruczek4391 Happy to hear!
So, what was FWHM of good image without Blur XT? My guess it was about 4.
I run similar evaluation. Original image 1.8” (only Lum) after Blur XT v1, I got 1.5”, v2 gave 1.35”. Some improvement, but I could not clearly spot it with naked eyes.
Bottom line, excellent tool for lousy pictures. Work on acquired data quality.
If there would be something like a "Astrophotography-Drillinstructor" - it would be you. I literally see you standing in front of me, shouting in my face: "It this pathetic attempt of an astrophoto the best you can do? Back to your telescope and work on acquired data quality!!!!!!!!" 🤣
@@viewintospace This is actually your image in mirror. However, you eventually came to many conclusions that I told you. Work on quality of data!
I never used bxt cause to me it always had poop results but this new version might be interesting
Thank you Sascha! Have you tried this on moon pics yet? I've used your technique of sharpening lunar images with BXT to great effect. But now the issue is that these are always already stretched coming out of Autostakkert.
I did not try it yet but I have the same concern that you have.... 😨
How about double stars, are they merged into one perfect star or does BXT AI distinguish 🤔
It does distinguish - no problem. The issue was that BXT V1 make double stars out of eggy single stars - that does not happen now anymore.
Cool, but what I mean is the other way around, that double stars are merged into one star.
I understood - no, will not happen - BXT will not merge stars, as from how deconvolution works it has rather the tendency to break it apart than to merge
wow, comes as a free upgrade for people that have already paid for it. That is not very common anymore. Thanks for the highlight, I might have missed that.
I tried it on a few different Lum integrations from targets I've already processed with the previous version, and in every case, though the non-stellar sharpening is significantly increased/improved, stars aren't reduced as much using the same settings as before. Anyone else seeing that?
First of all, I would say non-stellar sharpening is improved - not increased. It was probably over sharpened before in V1. When it comes to stars it is the same story - the are much better corrected, but rather improved than reduced. Which makes sense as BXT is not a star reduction software. But obviously you always increase the star sharpening to achieve the same reduction effect as before.
@@viewintospace Ok, I'll accept more so improved over increased, but also some of both. I don't have the images handy, but one comparison I did was on NGC 2146 and while the the center dust lane was definitely improved though actually wider/thicker than the V1 result, sharpening in other areas of the galaxy was 'increased' in the sense that those areas were made less blurry than V1. Regarding the stars, my comment was more an observation than a criticism. I thought you had seen a reduction in FWHM where I saw the opposite. On another note, I know Russ has said this is a single application tool, not one you use iteratively, but I'm curious why not, in theory at least. I've recently experimented with multiple iterations at less aggressive settings with positive results.
BTW, thanks for a great video!!
RC astro must not have updated in the repository because i still have AI version 2 on 1.2.1. Unless you got a early copy by chance
Yes, I was beta testing - the software should be out soon!
@@viewintospace omitting this in your "announcement" is a significant omission
Yes, it was not planned that way.... But lets say "external factors" forced me to publish....
Just got update from Russ: Release is at 16.00 UTC
@@viewintospace excellent! Just imaged the crab nebula last night so im ready to see how it does
Hoe geraakt ik aan die pdf?
You will find it on my Patreon account with many other support documents. Link is in the description.
This is AP easy mode now. No point in getting expensive optics and crazy flat fields lol.
When it comes to round stars you are right. But when it comes to detailed nebulae, the equipment quality will still be relevant!
He should really make this available for APP or photoshop on Mac 😢
No, he should not (and he will not btw...). APP is great for stacking but not a processing software at all, and Photoshop might be ok for post processing but also is not equipped at all for adequate Astrophoto processing. I know PixInsight UI is horrific, but still - there is (presently) no way around it.
Interesting as well because GradientXterminator was one of Russ's first plugins and is still only available in Photoshop. I was hoping he would eventually bring it over to PI, but it hasn't happened yet. And with graXpert working so well now maybe it's not necessary. But part of me still thinks Russ will still tackle gradients in the PI world with an amazing future release!!
@@Mr77pro I think you should tell this part of you to stop thinking... 😉 On one side there is GraXpert, but there is also the MARS project of PI. So what we really do not need is a 3rd Gradient Removal Option.
@@viewintospace Yes, this is certainly true. So that part of me is done thinking that!🤣🤣
The coma corrector manufacturers hate this trick. Just kidding, I'm aware that better raw images lead to better results
This is going too far. What is the point in buying expensive gear and spending months getting it set up properly, learning to guide properly etc etc when some numpty can grab a cheap scope, take a few unguided blurry images and then use AI to make them look great? Version 1 was fine, it basically made deconvolution, an existing process, a lot easier. This version may not "make up" data but that's not the point, it is certainly transforming bad data into better data to an extent that the final version bears no resemblance to the original data. Very depressing.
People said similar things when we transitioned into digital from film. If anything this introduces more people into the hobby, as it's complexities get solved through software. The people who make spectacular images will co continue to do so.
Here we go again..... 😵💫 We had the same discussions with BXT1..... let's not start again. BXT2 makes actually nebulas more realistic now, as they were oversharpened with BXT1. And about the stars - can we agree that stars are ALWAYS round? So if it makes stars round, it puts it in its correct shape, nothing more.... At the end, it is the same discussions we have with the Seestar and similar devices - more and more, people can create astrophotos easier and with less expensive gear. It sucks for people two spent a fortune on their equipment, but it general it's a good thing!
I agree with midnight lightning on this one. I think it's perfectly okay to have anomalies in your image and I believe this type of image manipulation is just going a little too far. I think it's nice to have tools that help with aberrations but this is almost getting to a point where all images are going to look the same and then it's just a matter of how other people colorize their data afterwards. Doing normal sharpening is fairly basic and it uses your existing data, but we have no idea what is actually happening here. If AI is making determinations on what should an image should look like then that is just false, I would rather just use the data that I have and know that it's real. I know that using software like topaz denoise, you can see new structures being created and I'm beginning to start leaning away from doing any type of AI related processing now and just stick to natural imaging. I am all for technology but this is taking things a little too far just like some of the AI generated Astro images that are being put out on social media right now, which is a joke. I do understand where Sasha is coming from, I don't want to be a hater of new technology but at the same time I am very concerned with software creating things based on What it thinks it should be. I am perfectly fine with aberrations in my image and it only challenges me more to find a way to get rid of them myself versus allowing software to do it for me.
@@billblanshan3021 I see your point Bill! Just thought of a mountain climber reaching the top by climbing the vertical wall. The other person takes a hiking path to the top and the third takes a cabletrain… All three stand finally on the same mountain but everyone is happy - some for their own best effort they spent, some simply for enjoying the view…,
@billblanshan3021 @midnightlightning1
Today Adam Block answered this question perfectly on Facebook:
"How is an image that has elongated and blurry stars not a fake image? That isn't really part of the scene and the information is distorted. Literally fake. You might be tempted to say that the elongation and blur is something that is understandable and happens in a regular way (bad tracking, optics, focus...etc). Well..if you can have a reality and distort it... you can, if you are clever enough- do the inverse. The clever part is the hard part in many cases. The point is, there is quite a bit of information in your elongated and blurred stars." - Adam Block, PixInsight for Beginners Forum, Facebook