Folks, I have never met Adam Block.......I became a member of Adam Block Studios close to two years ago.. This was the investment that paid more dividends than all of my gear purchases. I paid far less for membership to the studios than the cost of my last guide camera. For the record I do not know Adam Block although I feel like I do from watching ALL of his videos.
Well... as I recall you posted a video called the "BEST TOOL EVER" ... I chose to avoid that since things are changing and things don't always age well. lol Thanks for watching. It isn't polished... but best I could put together in just a few days- both processing and making a video.
@@AdamBlock Oh mate, you've done a wonderful job on the video - You're far too modest! 🙂 RE: "BEST TOOL EVER" well, it might be time for me to make a "BEST TOOL EVER PT2" LOL! :-D Clear skies!
I joined Adam Block Studios several months back and went through all the Pixinsight videos. Gold star for me. It’s a lot of information! But weather (pre-smoke) has been a no go until May new moon. I got a small set of images to work with on M27. So for the past 3 weeks I’ve been going back through the videos to get a better idea of what to do. Was still learning Nina and didn’t get any useful flats. I’m at the deconvolution section, which refers me here. Obviously I’m buying the tool. I just want to thank Adam for all the work he’s put into this project. You said Blur-exterminator is the wow factor. Adam, you are the Wow Factor!
I saw this new tool before I watched your video. I instantly purchased it before trying it myself, because Russ's tools are the ones I use the most, they are unbelieveable good. At first I was a bit shocked when the price showed up, but it is really worth the money, it is what you need.
this is like falling in love with Astronomy all over again...this is really blowing my mind..and I will certainly to reconsider doing PXS purchase after doing a trial of it a while back. This tool seems to be worth every penny for sure. Thanks a ton Adam.
Hi Adam, thank you very much for this review. I was blown away by Noise Xterminator's simplicity and efficiency. Blur Xterminator has that same feel. Looks really promising!
Thanks so much for your continued focus on Russ Croman’s work! His new modules are forcing me to go back and revisit all my old data sets. And now I have to start over all again :). And having watched, WOW!!! Bought it as soon as I saw this!
Holy Moly !!! In essence, it is the tools that constantly daydreams; for a beginner like myself, deconvolution is really challenging. It has been quite satisfying to see an expert's review. I'll attempt it in my few sessions. Ty Adam !
I simply say "Thank you for such a complete introduction of BlurXTerminator". Impact is clear and I already started to use and see impossibles now possible.
Absolutely wow! I just got into astrophotography last year and boy did I time my entry well. Incredible improvements in telescope designs, go-to mounts, acquisition, CMOS cameras and post processing. Russell's tools are incredible. Thanks for taking the time to provide detailed explanations in all of your videos! Dr B from Manitoba, Canada
The BlurXterminator tool along with StarXTerminator will manage my biggest flaws in imaging. I'm about to tear down my CGX because at 30s, I have to reject 50% of my subs for star eccentricity alone. I'll still fix it, but I plan on getting this tool and fixing some of my favorite images I have taken in the past. I have already had great results with his star removal tool that makes it feel like I'm almost cheating. Also Adam's "Fundimentals" program has allowed me to use PixInsight far more effectively and plan on moving onto his other program. This last year has been gamechanging for me.
Hey Adam! This was a lengthy video but I didn't notice until it was over as the wealth of information you have shared about this new tool is very generous. Russel is really becoming the AI whisperer when it comes to training the neural net to do impossible feats like this. Your deconvolution videos on your site were very helpful but also very time consuming as deconv can be. This is like magic! I am happy to have learned the process the manual way to appreciate what Russ has created here. Thanks for the video! CS!
Amazing stuff Adam. I bought it the minute I saw your endorsement. Amazingly I brought up an image of M51 last year shot at roughly 733 mm with my Stellarvue triplet and with the BlurXterminator it is allowing me to zoom in and present a larger image that actually holds up. Cool stuff. Thanks for the info and video!
Thank you for this very informative tutorial. I tried it right away and the results are incredible. Moreover, it works even with my latest Mars images when you set the PSF manually.
Minute 42: The differences are easily visible in your post. This looks like a spectacular advance. Than's for sharing a detailed insight into how to use it.
Perfect timing! I have been trying to get my head around Deconvolution and failed, even using EZDeconv hasn’t been very successful, and now this comes along. Thanks for breaking the news.
Narrow band? Where would we use this in our workflow? I’d like to remove stars at the beginning, like your rosette color modifier video. All three masters? Ha? First step Once combined?
Another excellent and informative video Adam! Just this morning I thought to myself, 'I wonder if Russ is working on a sharpening program, that would be so cool'.
Wow! Excellent video, Adam! The difference of what you were able to do with your M4 data, compared to when you originally processed it during your first time using PixInsight, is dramatic - and really speaks to the effectiveness of this tool!
I hereby dub thee Russ Croman the best tool creator in the galaxy (probably). Thanks for the in depth look at this tool I'm excited to go back through my stuff!
Wow. Extraordinary results. I was hoping at some point Russ would do something like this but this exceeded all my expectations. Thank you for being so quick in releasing this tutorial. It’s going to transform everyone’s images and make beautiful images from amateur equipment and even more stunning images from high end equipment.
Silly question. You said best time to apply is just after dbe and color correction when you first make an rgb image. Imaging LRGB am I to do LRGB combine use decon then split the Lum for further processing? Idk maybe I have been going about this all wrong. Do I not need to process LUM separately? I have been since I began.
Fantastic work! AI models are certainly shining in these application areas at the moment. In terms of workload, I would normally use EZ-Decon on the integrated mono channels and then combine into RGB. With BXT in the workload, would I no longer use EZ-Decon and just apply BXT to the combined RGB or LRGB image? Thanks!
I'm going to download it right now. I had to go back through all of my old images when NoiseXterminator came out, and then again, to a lesser degree with StarXterminator AI11, but it looks like another 3-4 months of re-does is going to be in order.
Amazing topic. Thanks for those detailed explanation and examples, Adam. And of course a huge thank you to Russel for making this state of the art technology available to the astrophotography community!
What!! This is an amazing step forward. I've been trying EZ Decon and also applying your decon steps Adam, but always end up with unwanted ringing or other problems. I'll be revisiting a recently completed nebula image and stepping through it again by adding this tool into the fold. Thanks for bringing this to our attention Adam - your AB Studios lessons have re-energized my astrophotography goals.
Thanks Adam. I got it when Russ put it out there and was already using it but this provided a great deal of context and info. Agree it would be nice to display or log the PSF being used but trial and error works for now. Thanks again
The results you show are astounding. Beyond WOW. thanks to Russ the tools, and to you for the clear introduction, description and explanations. I look forward to trying this out.
This will vastly improve “effective seeing” for those of us with larger oversampled scopes for our average seeing conditions. My 9.25 just got a new lease on life.
Many thanks Adam. I am now trying this out. VERY useful to know that NR should not be done first. I would always assume you should reduce noise before sharpening (lest you sharpen the noise too!). So very happy to stand corrected!
58:49 This reminds me very much of wavelet sharpening of images of Mars. There is a lower limit on size, and oversharpening causes artifacts of that size.
I know I already commented once and asked a question sorry :) I am a member of you site if that matters :) My second question is what impact might this have one relaxing any subframe selection parameters as they relate to eccentricity...? Knowing that we have potentially more correction available for fixing eccentricity, might we allow more eccentric star though the subframe selector now? Thereby giving us access to subs we might have rejected before?
the change at 12:00 makes me really wonder how it could pull out that much information. For example If i just focus at the top of the biggest pillar, the fine details are incredible.
This is mind-blowing :) I can't wait to implement this in my typical narrowband workflow. I typically often use star exterminator on each filter first and then stretch just the nebulosity. With the intention of re-introducing a separately captured RBG stars dataset. Given this, I assume I'd run BlurX first on each filter right after integration, then StarX and continue my workflow? And of course run BlurX on the RBG dataset as well before re-introduction. Thank you so much for getting this video out so quickly :)
It's probably best to run BlurXTerminator on color images rather than individual channels. The network is trained to look at all channels together, and do things like correct for unequal FWHM values between the channels. It can't do this if it doesn't see all of the channels simultaneously.
@@rrcroman Thank you so much Russell for your reply, this process is really amazing, great work! Since the process has to be run on linear data I suppose it’s not possible to use it in a narrowband workflow where one typically stretches individual channels separately?
For NB, I would 1) channel combine linear to SHO, 2) calibrate color using SPCC NB mode or ColorCalibration with nebula as white reference (both are described in the SPCC doc), 3) run BXT. After BXT decon, if other palette/stretching choices are desired, extract back to separate channels and have at it.
Extremly cool, Adam. Thank you and Russ for that. Just one question: Do I use the Blur XTerminator on RGB and additional on the luminance with LRGB data or only on one of the two? CS Christian
Good question. This is uncharted territory.. and good ABS tutorial material. :) I did have a little experience with this... since the sharpening can be local to some degree... there might be differences between the L and RGB in a way that creates an interesting blend of LRGB later. You might find (as I did) that faint stars and things in the RGB were sharpened/brightend to a degree that brough more color to these things when creating the LRGB than you might expect. I had some crazy red galaxies in the background. So the safer route is probably make the LRGB first... then BXT to avoid some weirdness.
@@AdamBlock Thank you so much for the detailed answer. I'll try than all methods first and will wait patiently for your experiences in the upcoming tutorial at ABS.😀😅
I have now tried BlurX-Terminator twice. Perhaps for reasons totally not connected to BXT, my small stars get "dots" in the center. That's a type of artifact that I would see in conventional deconvolution. That's why I made the association. I am now using a smaller sensor camera (ASI294) instead of my ASI2600s, which had been shredding my corner stars. Just my results so far, I left BXT at its default values. What other things can cause dotted stars?
Many thanks for the video Adam. Very comprehensive review of what looks like a great new tool. You mention your video about fully processing M51. I always have trouble processing galaxies in PI so would find this useful. Which of your packages is this included in? Thanks
You implied that local Deconvolution is dependent on local PSF, which lead you to conclude that BXT won’t work iteratively - you can’t fix star shapes first and then run stellar and non-stellar sharpening as a separate step because the first step will have modified the PSFs for all stars. So it means we have only one shot to sharpen the image and correct stars, correct? Also where in the process would you apply NoiseXterminator? Still in linear phase? Applying it before BXT would also modify PSF so definitely it can’t be used before BXT. For narrowband images: apply BXT and NXT on each channel first or create a channel combination and then apply sharpening and denoising on the colour image (after colour calibration and SCNR)?
Thanks for analysis! I've just tested it on some of my own data: extremely impressive on good initial data, ok to good on not so good data (tracking error, pinched optics). Better on mono channels than on OSC data. But a huge time saver compared to the complete manual process which leads, at least in my case, to worse results... You can't stop progress I guess... but for people who have been through the tedious deconvolution and denoising learning processes in Pixinsight this is a bit of a slap in the face ;)
I would be interested to learn if there is a Batch-Mode Operation for BlurXterminator. Because I want to sharpen my 44 Comet-Aligned Images prior to let StarXterminator do its job. with the initial state of my RGB-fotos, about 50% of the stars were not removed, and this in a completely irregular manor. So I presume a lack of sharpnes beeing the origing of it. Thanks for any kind of hint. KVa
That's not specifically covered, but it does mention that noise reduction should not be done before deconvolution: it can alter or destroy the very low-contrast information in the image needed for proper deconvolution.
@@rrcroman the Narrow band. Where would I implement this step using narrow band , and the color modifier workflow? You know, where Star Removal is one of the first steps along with linear fit before actually combining everything ? Rosette example. So excited!!!!
You will want to do the color calibration of your data (SPCC) before BXT. Otherwise you do not want to do any stretching or noise reduction before using it.
Very interesting. And now also the name Deconvolution started to make more sense, as in the past it seemt to me like it should be called Convolution and vice versa :D
Extraordinary Adam. For a first reaction, it's excellent. I've added BlurXTerminator to my workflow now but am enduring the "dilemma of choice" phenomenon. Which of my images to I start with? I have a good idea after seeing your video.
Adam, I was following along on your NGC1333 Fundamentals video, but started the process with SXT and then BXT. When I hit the Unsharp Mask procedure, I was wondering if that counted as "Double Decon"?
Unsharp mask is not deconvolution. It is blind to any sense of PSFs and it operates as a kernal filter across all pixels (at a desired scale). It is a contrast enhancing process like many others.
@adamblock Awesome video and so timely - though I must say, as a member of Adam block Studios, I can't wait to see what you unpackage further in your tutorials and can't recommend it enough! Whilst watching however, I wondered if the difference of visible/missing data (in the previewed finger area of your M1 vs HST's image) could be due to its expansion and rapidly changing structure as evidenced in your earlier time lapse, i.e. the 2 images were taken at different times?
Not so much of a difference in structure there... you will note I purposefully choose the outer envelope. Those structures are large and not really changing much while moving outward. At least not on these time scales..but you do have a good point!
Hi Adam, thank you for this very good walk through the blurexterminator. Do you think applying the bluexterminator on individual monochrome images is better than doing it on the combined colored image? I am thinking specifically on SHO processing
It is complicated. If you are matching star sizes..no- not a good idea of course. If you are doing this on non-stellar- you can do it on the individual images- but you should use the same (manual) PSF. However, when you do this- there is no correction for optical aberrations. This is the benefit of doing it on a full color image (whether starless or not).
As ever (and totally expected from Adam), an in depth, clear explanation - thanks! One question - I wonder if BXT could be used on science images as well?
@Adam Block For narrowband images from mono cameras there's an issue of star sizes being different. HA is usually the sharpest and SII usually has the most bloated stars. Also difference in seeing (due to weather or object's angle above horizon) through the session can make a difference. Would you recommend to do channel combination/pixel math or whatever you use to make colour NB images to use source images with original star sizes and leaving it to BlurXterminator to correct stars with colour fringes in the RGB image, or apply some other process to make sure stars are same size before combining mono masters into RGB image?
My answer... get real RGB stars. (I am serious). NB stars are kind of a silly thing. They are NOT narrowband objects- why treat them as such? This issue of star sizes... oh yeah, NOW people start to understand the intricacies of this issue. For years people would tell me I blab on and on about these technicalities... but no one was doing LRGB in the same way. Now with NB images of incommensurate stars and availability of easy deconvolution it all becomes important again. The bottom line is whether doing LRGB or in your special case of RGB- yeah the star sizes (PSFs ) need to be similar or weird halos and artifacts will be an issue. I would not rely on BXT to "fix" the halos or fringing for this. You could use BXT (maybe) to match PSFs. That is the long answer..the short answer is I do not know precisely- this isn't a permutation I have yet to deal with.
@@AdamBlock because RGB stars are so big compared to NB stars, especially HA stars, that I can barely see the object behind the dense field of stars. Also neither StarXterminator nor StarNet2 is not always able to remove/mask smallest stars so after replacing stars with RGB I still have little yellow/orange SHO or red HOO stars. I tried using MMT to smooth them out but it also blurs the details of the nebulosity. I tried both EZ Star Reduction and “Bill’s method” and they are not being able to pick up smallest stars. 😢 One thing I want to experiment with is unscreen HA stars and use them as a luminance layer for unscreened RGB stars which should reduce their sizes without creating dark rings around them as most algorithms do (both EZ Star Reduction methods, Bill’s method and BXT). And then rescreen them onto RGB image.
On my processing of that IC 405 image I had actually successfully employed Deconvolution, one of my first times doing so. Russell is making astrophotography processing almost too easy! Thanks for the video, Adam. -A
Awesome video and function, many thanks Adam. I noticed you have PI configured so that Images display over the top of Icons (minimised images). In my config the Icons show on top of the Images - how do I change it?
Folks, I have never met Adam Block.......I became a member of Adam Block Studios close to two years ago.. This was the investment that paid more dividends than all of my gear purchases. I paid far less for membership to the studios than the cost of my last guide camera. For the record I do not know Adam Block although I feel like I do from watching ALL of his videos.
Thanks for that unsolicited kindness!
@@AdamBlock You are more than welcome - it is deserved!
Dude! I feel the exact same way - even what's free on youtube has been an absolute game changer for me. thanks Adam!
Absolutely 💯 amazing 👏
Your channel is criminally underrated.
Help Me! LOL thanks for that...
Good grief this looks incredible, Russ's work is wild! - Thanks for sharing this and demonstrating so comprehensively Adam, you're the man! :-D
Well... as I recall you posted a video called the "BEST TOOL EVER" ... I chose to avoid that since things are changing and things don't always age well. lol Thanks for watching. It isn't polished... but best I could put together in just a few days- both processing and making a video.
@@AdamBlock Oh mate, you've done a wonderful job on the video - You're far too modest! 🙂
RE: "BEST TOOL EVER" well, it might be time for me to make a "BEST TOOL EVER PT2" LOL! :-D
Clear skies!
Sold! Thanks Russ and Adam! Oh btw Adam, you’re a great marketeer for Russ’ product 😀
WOW, is right!!! Thank you for the detailed explanation Mr. Block. I will be purchasing this tool...
I joined Adam Block Studios several months back and went through all the Pixinsight videos. Gold star for me. It’s a lot of information! But weather (pre-smoke) has been a no go until May new moon. I got a small set of images to work with on M27. So for the past 3 weeks I’ve been going back through the videos to get a better idea of what to do. Was still learning Nina and didn’t get any useful flats. I’m at the deconvolution section, which refers me here. Obviously I’m buying the tool. I just want to thank Adam for all the work he’s put into this project. You said Blur-exterminator is the wow factor. Adam, you are the Wow Factor!
Nice Adam! Thanks for being ahead of the curve
Thank you my friend.
Hello from UKRAINE! Excellent tool! Thank you for review!
Amazing tool and equally amazing presentation of its capabilities!
Thank you! Best I could do on short notice!
Just great. I recon this tool will be helpful for many people. Thanx Adam.
A very useful review of BXT thanks Adam. Your final four animated images are sensational!
Thank you!
This is extremely cool. Thanks for sharing this guide.
Thanks for you r kind words.This was all done in just a few days... so not polished.. .but hopefully useful.
It is truly incredible...just applied to some old TIFFS and WOW!!!!...thankyou RC and Adam for this 1st video
Oh yes, that's a very big WOW !
Thanks for your high quality video!
I saw this new tool before I watched your video. I instantly purchased it before trying it myself, because Russ's tools are the ones I use the most, they are unbelieveable good. At first I was a bit shocked when the price showed up, but it is really worth the money, it is what you need.
Yes, completely changed and massively simplified my workflow....
This looks absolutely amazing! Thanks again Adam for everything you do to keep our hobby moving forward!
Thanks for watching!
this is like falling in love with Astronomy all over again...this is really blowing my mind..and I will certainly to reconsider doing PXS purchase after doing a trial of it a while back. This tool seems to be worth every penny for sure. Thanks a ton Adam.
Thank you
Hi Adam, thank you very much for this review. I was blown away by Noise Xterminator's simplicity and efficiency.
Blur Xterminator has that same feel. Looks really promising!
Thanks so much for your continued focus on Russ Croman’s work! His new modules are forcing me to go back and revisit all my old data sets. And now I have to start over all again :). And having watched, WOW!!! Bought it as soon as I saw this!
Thank you Adam for the in-depth demonstration. I’ll be adding this to the other tools from Russ.
Holy Moly !!! In essence, it is the tools that constantly daydreams; for a beginner like myself, deconvolution is really challenging. It has been quite satisfying to see an expert's review. I'll attempt it in my few sessions. Ty Adam !
BXT seems really impressive ! Thanks for this video !
You are right… it is a wow…I did some polls on Facebook…. But here I found the answers… Your explanations are excellent…. Thank you so much
Thanks for watching!
Awesome video, again, thank you Adam and obviously another fantastic tool from Russell Croman
This is incredible, like magic!
thanks for sharing.
I simply say "Thank you for such a complete introduction of BlurXTerminator". Impact is clear and I already started to use and see impossibles now possible.
Russell has hit another one out of the park. You've done it again, too, with this demonstration of its powerful results. Thanks to both of you.
Absolutely a fantastic addition to Pixinsight. Thank for a very informative tutorial on this new method!
Wow! Thank you, Russel and Adam for dissecting it!
Well, I'll be downloading the trial and having a go as soon as Christmas/work is out of the way! Great video, thanks.
Absolutely wow! I just got into astrophotography last year and boy did I time my entry well. Incredible improvements in telescope designs, go-to mounts, acquisition, CMOS cameras and post processing. Russell's tools are incredible. Thanks for taking the time to provide detailed explanations in all of your videos! Dr B from Manitoba, Canada
The BlurXterminator tool along with StarXTerminator will manage my biggest flaws in imaging. I'm about to tear down my CGX because at 30s, I have to reject 50% of my subs for star eccentricity alone. I'll still fix it, but I plan on getting this tool and fixing some of my favorite images I have taken in the past. I have already had great results with his star removal tool that makes it feel like I'm almost cheating. Also Adam's "Fundimentals" program has allowed me to use PixInsight far more effectively and plan on moving onto his other program. This last year has been gamechanging for me.
Hey Adam! This was a lengthy video but I didn't notice until it was over as the wealth of information you have shared about this new tool is very generous. Russel is really becoming the AI whisperer when it comes to training the neural net to do impossible feats like this. Your deconvolution videos on your site were very helpful but also very time consuming as deconv can be. This is like magic! I am happy to have learned the process the manual way to appreciate what Russ has created here. Thanks for the video! CS!
Thanks Dave!
Really incredible. Thanks to Russ, and to you Adam for tutoring us as usual. [and 6770 is spectacular!]
Thanks for sharing! I'm already testing!
This is going straight into my Pixinsight workflow, thank you for doing this tutorial 👍🏻😀
Thank you for watching!
Amazing stuff Adam. I bought it the minute I saw your endorsement. Amazingly I brought up an image of M51 last year shot at roughly 733 mm with my Stellarvue triplet and with the BlurXterminator it is allowing me to zoom in and present a larger image that actually holds up. Cool stuff. Thanks for the info and video!
I "get" your statement re using BE in Linear state before extended processing but it still has an incredible performance on old finished TIFFS
Thank you for this very informative tutorial. I tried it right away and the results are incredible. Moreover, it works even with my latest Mars images when you set the PSF manually.
Another excellent tool, thanks for the demonstration and info 👍
Minute 42: The differences are easily visible in your post. This looks like a spectacular advance. Than's for sharing a detailed insight into how to use it.
Perfect timing! I have been trying to get my head around Deconvolution and failed, even using EZDeconv hasn’t been very successful, and now this comes along. Thanks for breaking the news.
Fantastic demonstration!
Excellent video, esp.regarding the background infos and your expectations! Got my BXT license immediately as I already use SXT and NXT very successful
Thanks for watching the video!
Narrow band?
Where would we use this in our workflow? I’d like to remove stars at the beginning, like your rosette color modifier video.
All three masters?
Ha?
First step Once combined?
As a subscriber, thanks for this video and all your other videos. Some of the best money I've spent on this hobby.
Another excellent and informative video Adam! Just this morning I thought to myself, 'I wonder if Russ is working on a sharpening program, that would be so cool'.
Great intro to another great tool by Russ. Thank you!
Wow. that is incredible! Thank you for explaining in such detail, especially the psf.
Wow! Excellent video, Adam! The difference of what you were able to do with your M4 data, compared to when you originally processed it during your first time using PixInsight, is dramatic - and really speaks to the effectiveness of this tool!
I hereby dub thee Russ Croman the best tool creator in the galaxy (probably). Thanks for the in depth look at this tool I'm excited to go back through my stuff!
Wow. Extraordinary results. I was hoping at some point Russ would do something like this but this exceeded all my expectations. Thank you for being so quick in releasing this tutorial. It’s going to transform everyone’s images and make beautiful images from amateur equipment and even more stunning images from high end equipment.
Silly question. You said best time to apply is just after dbe and color correction when you first make an rgb image. Imaging LRGB am I to do LRGB combine use decon then split the Lum for further processing? Idk maybe I have been going about this all wrong. Do I not need to process LUM separately? I have been since I began.
Fantastic work! AI models are certainly shining in these application areas at the moment. In terms of workload, I would normally use EZ-Decon on the integrated mono channels and then combine into RGB. With BXT in the workload, would I no longer use EZ-Decon and just apply BXT to the combined RGB or LRGB image? Thanks!
Wow Adam how cool.Just got it!!!
I'm going to download it right now. I had to go back through all of my old images when NoiseXterminator came out, and then again, to a lesser degree with StarXterminator AI11, but it looks like another 3-4 months of re-does is going to be in order.
Amazing topic. Thanks for those detailed explanation and examples, Adam. And of course a huge thank you to Russel for making this state of the art technology available to the astrophotography community!
Amazing. Thanks Russ and Adam. Would drizzling before running blurXterminator interfere with its function?
No... Drizzling is part of the integration of data. BXT operates on integrated images. So all is fine!
A great gift for Christmas!
Adam, Thanks for the video
your welcome
What!! This is an amazing step forward. I've been trying EZ Decon and also applying your decon steps Adam, but always end up with unwanted ringing or other problems. I'll be revisiting a recently completed nebula image and stepping through it again by adding this tool into the fold. Thanks for bringing this to our attention Adam - your AB Studios lessons have re-energized my astrophotography goals.
Speechless! What a fantastic tool...
Thanks Adam hope to see you at NEIAC.
Thanks Adam. I got it when Russ put it out there and was already using it but this provided a great deal of context and info.
Agree it would be nice to display or log the PSF being used but trial and error works for now.
Thanks again
Thanks for walking through the tool. It is impressive!
The results you show are astounding. Beyond WOW. thanks to Russ the tools, and to you for the clear introduction, description and explanations. I look forward to trying this out.
This will vastly improve “effective seeing” for those of us with larger oversampled scopes for our average seeing conditions. My 9.25 just got a new lease on life.
Many thanks Adam. I am now trying this out. VERY useful to know that NR should not be done first. I would always assume you should reduce noise before sharpening (lest you sharpen the noise too!). So very happy to stand corrected!
58:49 This reminds me very much of wavelet sharpening of images of Mars. There is a lower limit on size, and oversharpening causes artifacts of that size.
I know I already commented once and asked a question sorry :) I am a member of you site if that matters :) My second question is what impact might this have one relaxing any subframe selection parameters as they relate to eccentricity...? Knowing that we have potentially more correction available for fixing eccentricity, might we allow more eccentric star though the subframe selector now? Thereby giving us access to subs we might have rejected before?
the change at 12:00 makes me really wonder how it could pull out that much information.
For example If i just focus at the top of the biggest pillar, the fine details are incredible.
Is this even real life? What an awesome tool! Thanks Adam for explaining in detail how to use it and show the results with different types of DSOs.
This is mind-blowing :) I can't wait to implement this in my typical narrowband workflow. I typically often use star exterminator on each filter first and then stretch just the nebulosity. With the intention of re-introducing a separately captured RBG stars dataset. Given this, I assume I'd run BlurX first on each filter right after integration, then StarX and continue my workflow? And of course run BlurX on the RBG dataset as well before re-introduction. Thank you so much for getting this video out so quickly :)
It's probably best to run BlurXTerminator on color images rather than individual channels. The network is trained to look at all channels together, and do things like correct for unequal FWHM values between the channels. It can't do this if it doesn't see all of the channels simultaneously.
@@rrcroman Thank you so much Russell for your reply, this process is really amazing, great work! Since the process has to be run on linear data I suppose it’s not possible to use it in a narrowband workflow where one typically stretches individual channels separately?
For NB, I would 1) channel combine linear to SHO, 2) calibrate color using SPCC NB mode or ColorCalibration with nebula as white reference (both are described in the SPCC doc), 3) run BXT. After BXT decon, if other palette/stretching choices are desired, extract back to separate channels and have at it.
@@rrcroman This is excellent, thank you Russell
I wasnt all that impressed with NoiseTermintorX but i've tried blurXterminator and its a game changer, unreal.
You said WOW. I literally exclaimed "Holy $@&!". Look forward to seeing this integrated into your future Horizons workflows as well.
LOL I do plan on making some AdamBlockStudios.com specific videos on this...
Extremly cool, Adam. Thank you and Russ for that.
Just one question: Do I use the Blur XTerminator on RGB and additional on the luminance with LRGB data or only on one of the two?
CS Christian
Good question. This is uncharted territory.. and good ABS tutorial material. :) I did have a little experience with this... since the sharpening can be local to some degree... there might be differences between the L and RGB in a way that creates an interesting blend of LRGB later. You might find (as I did) that faint stars and things in the RGB were sharpened/brightend to a degree that brough more color to these things when creating the LRGB than you might expect. I had some crazy red galaxies in the background. So the safer route is probably make the LRGB first... then BXT to avoid some weirdness.
@@AdamBlock
Thank you so much for the detailed answer. I'll try than all methods first and will wait patiently for your experiences in the upcoming tutorial at ABS.😀😅
OMG! This is revolutionary!
It really is...
I have now tried BlurX-Terminator twice. Perhaps for reasons totally not connected to BXT, my small stars get "dots" in the center. That's a type of artifact that I would see in conventional deconvolution. That's why I made the association. I am now using a smaller sensor camera (ASI294) instead of my ASI2600s, which had been shredding my corner stars.
Just my results so far, I left BXT at its default values. What other things can cause dotted stars?
Many thanks for the video Adam. Very comprehensive review of what looks like a great new tool. You mention your video about fully processing M51. I always have trouble processing galaxies in PI so would find this useful. Which of your packages is this included in? Thanks
You implied that local Deconvolution is dependent on local PSF, which lead you to conclude that BXT won’t work iteratively - you can’t fix star shapes first and then run stellar and non-stellar sharpening as a separate step because the first step will have modified the PSFs for all stars. So it means we have only one shot to sharpen the image and correct stars, correct?
Also where in the process would you apply NoiseXterminator? Still in linear phase? Applying it before BXT would also modify PSF so definitely it can’t be used before BXT.
For narrowband images: apply BXT and NXT on each channel first or create a channel combination and then apply sharpening and denoising on the colour image (after colour calibration and SCNR)?
Thanks for analysis! I've just tested it on some of my own data: extremely impressive on good initial data, ok to good on not so good data (tracking error, pinched optics). Better on mono channels than on OSC data. But a huge time saver compared to the complete manual process which leads, at least in my case, to worse results... You can't stop progress I guess... but for people who have been through the tedious deconvolution and denoising learning processes in Pixinsight this is a bit of a slap in the face ;)
hear hear!
I would be interested to learn if there is a Batch-Mode Operation for BlurXterminator. Because I want to sharpen my 44 Comet-Aligned Images prior to let StarXterminator do its job. with the initial state of my RGB-fotos, about 50% of the stars were not removed, and this in a completely irregular manor. So I presume a lack of sharpnes beeing the origing of it. Thanks for any kind of hint. KVa
fantastic video. always a great teacher. thanks
Does the documentation cover how to balance the sharpening effect in blurXterminator with the detail enchantment of noiseXterminator?
That's not specifically covered, but it does mention that noise reduction should not be done before deconvolution: it can alter or destroy the very low-contrast information in the image needed for proper deconvolution.
@@rrcroman the
Narrow band.
Where would I implement this step using narrow band , and the color modifier workflow?
You know, where Star Removal is one of the first steps along with linear fit before actually combining everything ? Rosette example.
So excited!!!!
See my reply to Damian Vine's similar question.
Incredible video! Question, do you recommend running the BlurX immediately after DBE?
You will want to do the color calibration of your data (SPCC) before BXT. Otherwise you do not want to do any stretching or noise reduction before using it.
The de-ringing looks impressive... indeed no black eyes.
No one wants a black eye of any kind!
Unreal!!! Thanks for this video
Very interesting. And now also the name Deconvolution started to make more sense, as in the past it seemt to me like it should be called Convolution and vice versa :D
Extraordinary Adam. For a first reaction, it's excellent. I've added BlurXTerminator to my workflow now but am enduring the "dilemma of choice" phenomenon. Which of my images to I start with? I have a good idea after seeing your video.
Adam, I was following along on your NGC1333 Fundamentals video, but started the process with SXT and then BXT. When I hit the Unsharp Mask procedure, I was wondering if that counted as "Double Decon"?
Unsharp mask is not deconvolution. It is blind to any sense of PSFs and it operates as a kernal filter across all pixels (at a desired scale). It is a contrast enhancing process like many others.
Re-reading... but if you mean doing USM before BXT... yeah, this would be a no-no.
This is incredible stuff. Thx Adam! Just to make sure will this work if you are taking images with a cmos color camera (ie. without filters)?
Absolutely... as long as you meet the requirements of deconvolution (good sampling, bright signal...etc)
Thx, just got the trial!
@adamblock Awesome video and so timely - though I must say, as a member of Adam block Studios, I can't wait to see what you unpackage further in your tutorials and can't recommend it enough! Whilst watching however, I wondered if the difference of visible/missing data (in the previewed finger area of your M1 vs HST's image) could be due to its expansion and rapidly changing structure as evidenced in your earlier time lapse, i.e. the 2 images were taken at different times?
Not so much of a difference in structure there... you will note I purposefully choose the outer envelope. Those structures are large and not really changing much while moving outward. At least not on these time scales..but you do have a good point!
Hi Adam, thank you for this very good walk through the blurexterminator. Do you think applying the bluexterminator on individual monochrome images is better than doing it on the combined colored image? I am thinking specifically on SHO processing
It is complicated. If you are matching star sizes..no- not a good idea of course. If you are doing this on non-stellar- you can do it on the individual images- but you should use the same (manual) PSF. However, when you do this- there is no correction for optical aberrations. This is the benefit of doing it on a full color image (whether starless or not).
Thank you so much for your time and the effort you put into this!!
As ever (and totally expected from Adam), an in depth, clear explanation - thanks! One question - I wonder if BXT could be used on science images as well?
Thank you!
@Adam Block For narrowband images from mono cameras there's an issue of star sizes being different. HA is usually the sharpest and SII usually has the most bloated stars. Also difference in seeing (due to weather or object's angle above horizon) through the session can make a difference. Would you recommend to do channel combination/pixel math or whatever you use to make colour NB images to use source images with original star sizes and leaving it to BlurXterminator to correct stars with colour fringes in the RGB image, or apply some other process to make sure stars are same size before combining mono masters into RGB image?
My answer... get real RGB stars. (I am serious). NB stars are kind of a silly thing. They are NOT narrowband objects- why treat them as such? This issue of star sizes... oh yeah, NOW people start to understand the intricacies of this issue. For years people would tell me I blab on and on about these technicalities... but no one was doing LRGB in the same way. Now with NB images of incommensurate stars and availability of easy deconvolution it all becomes important again. The bottom line is whether doing LRGB or in your special case of RGB- yeah the star sizes (PSFs ) need to be similar or weird halos and artifacts will be an issue. I would not rely on BXT to "fix" the halos or fringing for this. You could use BXT (maybe) to match PSFs. That is the long answer..the short answer is I do not know precisely- this isn't a permutation I have yet to deal with.
@@AdamBlock because RGB stars are so big compared to NB stars, especially HA stars, that I can barely see the object behind the dense field of stars.
Also neither StarXterminator nor StarNet2 is not always able to remove/mask smallest stars so after replacing stars with RGB I still have little yellow/orange SHO or red HOO stars. I tried using MMT to smooth them out but it also blurs the details of the nebulosity.
I tried both EZ Star Reduction and “Bill’s method” and they are not being able to pick up smallest stars. 😢
One thing I want to experiment with is unscreen HA stars and use them as a luminance layer for unscreened RGB stars which should reduce their sizes without creating dark rings around them as most algorithms do (both EZ Star Reduction methods, Bill’s method and BXT). And then rescreen them onto RGB image.
On my processing of that IC 405 image I had actually successfully employed Deconvolution, one of my first times doing so.
Russell is making astrophotography processing almost too easy!
Thanks for the video, Adam.
-A
Thanks A -A
Awesome video and function, many thanks Adam. I noticed you have PI configured so that Images display over the top of Icons (minimised images). In my config the Icons show on top of the Images - how do I change it?