I paid almost no attention to spacing, and indeed I don't even think I was perfectly focused.
So let's sum up the benefits. So far, we no longer need to guide well, we no longer need flatteners and we don't need to be in focus.
That's a heck of an upgrade! 😀 All kidding aside, I did the upgrade also and I'm impressed.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
|
I’m waiting for version VI where all you need is one sub from a cell phone camera and no calibration frames.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
To be sure, it's still not magic. The top and bottom right in this image are still pretty bad. Bright corner stars also are only barely acceptable. This is a full autostretch, but even with a lighter stretch I think more attention to backfocus is needed on my part:
 |
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
The discussion here is interesting. The sensor in my refractor sits about 0.5-1.0 mm too far out from the optimum position so my images exhibit just a little bit of astigmatic error in the extreme corners of 42 mm circle so I definitely appreciate the ability of BXT to fix that--particularly since I can't walk out to the backyard to readjust the spacing. One of the clever things that Russ has done is to make the sharpening algorithm non-shift invariant, which means that it works differently in different regions of the image. Round, tight stars are a big deal in astronomical images, but the thing to keep in mind is that it is always better to start with an optically well aligned system in the first place--particularly if you want to maximize image detail while minimizing artifacts. Yes, BXT2 does a spectacular job of fixing a lot of ills but I've always believed that it is best to start with the best possible data beforeusing tools like BXT/BXT2 to deconvolve the results. It's the imaging analog to a fundamental rule of mechanics: Always avoid using software to fix bad mechanics. Fix what's wrong first and then use software to boost performance beyond the limit of the mechanical tolerances. For a lot of the examples show here, it looks like BXT2 is just being used to try to get back to what the optics should have been producing in the first place. If that's the best you can do...well, so be it; but my advice is to avoid using BXT2 to give up on trying to fix fundamental optical problems. You'll get even better results if you first dial in your optics and then use BXT2.
John
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
John Hayes:
You'll get even better results if you first dial in your optics and then use BXT2.
While I fully agree with this, I think this is hard enough that we will take all the assistance we can get. And then, as a secondary matter, question whether that comes from software or optical perfection  |
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Works really well.
I gave my M16 data another shot. It showed misshapen stars all over the field due to the less than great reducer as well as bad seeing due to the low altitude of the nebula. BXT2AI4 did really well and corrected the stars. 
 |
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
I confirm that it works well also on coma sometimes present in short and fast focals.
This example below is a small field extracted in the corner of a Samyang 135 image processed by BX with the "correct only" option.
 |
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
John Hayes:
The discussion here is interesting. The sensor in my refractor sits about 0.5-1.0 mm too far out from the optimum position so my images exhibit just a little bit of astigmatic error in the extreme corners of 42 mm circle so I definitely appreciate the ability of BXT to fix that--particularly since I can't walk out to the backyard to readjust the spacing. One of the clever things that Russ has done is to make the sharpening algorithm non-shift invariant, which means that it works differently in different regions of the image. Round, tight stars are a big deal in astronomical images, but the thing to keep in mind is that it is always better to start with an optically well aligned system in the first place--particularly if you want to maximize image detail while minimizing artifacts. Yes, BXT2 does a spectacular job of fixing a lot of ills but I've always believed that it is best to start with the best possible data beforeusing tools like BXT/BXT2 to deconvolve the results. It's the imaging analog to a fundamental rule of mechanics: Always avoid using software to fix bad mechanics. Fix what's wrong first and then use software to boost performance beyond the limit of the mechanical tolerances. For a lot of the examples show here, it looks like BXT2 is just being used to try to get back to what the optics should have been producing in the first place. If that's the best you can do...well, so be it; but my advice is to avoid using BXT2 to give up on trying to fix fundamental optical problems. You'll get even better results if you first dial in your optics and then use BXT2.
John
Wholeheartedly agree John, and Russ makes this abundantly clear on his website:
"While BlurXTerminator AI4 can correct a much wider range of optical aberrations, there is no substitute for proper equipment tuning. Better raw data will always make for better final images regardless of which tools are used. AI4 can correct some serious aberrations with ease, but there will always be a limit to how much badness any software tool can fix. Invest the time and effort to minimize collimation errors, focal plane tilt, imaging train flexure, focusing errors, and guiding errors. You will be rewarded with superior results." |
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
I had BXT version 1 setup to use GPU acceleration. Does anyone know if this should carry on working with version 2, or do I need to do some fiddling again? (I seem to remember there was a complicated setup for this when I first installed BXT)
Thanks
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Stuart Taylor:
I had BXT version 1 setup to use GPU acceleration. Does anyone know if this should carry on working with version 2, or do I need to do some fiddling again? (I seem to remember there was a complicated setup for this when I first installed BXT)
Thanks
It isn't about BXT that is using it, it's about PI making it available to third parties such as BXT. So no fear, AI4 will be working just as well using GPU without any additional fiddling with it.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
andrea tasselli:
Stuart Taylor:
I had BXT version 1 setup to use GPU acceleration. Does anyone know if this should carry on working with version 2, or do I need to do some fiddling again? (I seem to remember there was a complicated setup for this when I first installed BXT)
Thanks
It isn't about BXT that is using it, it's about PI making it available to third parties such as BXT. So no fear, AI4 will be working just as well using GPU without any additional fiddling with it.
Super! Thanks Andrea. I am not very technical, alas, so appreciate your advice.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
I confirm that it works well also on coma sometimes present in short and fast focals.
This example below is a small field extracted in the corner of a Samyang 135 image processed by BX with the "correct only" option.
It does extremely well on my Rokinon images too. Weirdly, though, StarXTerminator will still not fully remove bright stars in my Rokinon pics after application of BXT.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Wow!! this is really something.
Here is an extreme corner of a recent image showing some horribly eggy stars. Look what a run of 'correct only' does for it! (recommended workflow now is to run this before SPCC)
 |
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
I’ve run it on a few different luminance integrations from previous projects and though the non-stellar sharpening is increased/improved over the previous version, the stars aren’t reduced as much using the same settings as before. Are other seeing this?
Cheers,
Scott
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Scott Badger:
I’ve run it on a few different luminance integrations from previous projects and though the non-stellar sharpening is increased/improved over the previous version, the stars aren’t reduced as much using the same settings as before. Are other seeing this?
Cheers,
Scott
Yes!
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Scott Badger:
I’ve run it on a few different luminance integrations from previous projects and though the non-stellar sharpening is increased/improved over the previous version, the stars aren’t reduced as much using the same settings as before. Are other seeing this?
Cheers,
Scott
Yes indeed! I found I had to nudge the sharpen stars slider up a bit. Previously I think I had it at 0.12 now I use 0.2
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
John Hayes:
The discussion here is interesting. The sensor in my refractor sits about 0.5-1.0 mm too far out from the optimum position so my images exhibit just a little bit of astigmatic error in the extreme corners of 42 mm circle so I definitely appreciate the ability of BXT to fix that--particularly since I can't walk out to the backyard to readjust the spacing. One of the clever things that Russ has done is to make the sharpening algorithm non-shift invariant, which means that it works differently in different regions of the image. Round, tight stars are a big deal in astronomical images, but the thing to keep in mind is that it is always better to start with an optically well aligned system in the first place--particularly if you want to maximize image detail while minimizing artifacts. Yes, BXT2 does a spectacular job of fixing a lot of ills but I've always believed that it is best to start with the best possible data beforeusing tools like BXT/BXT2 to deconvolve the results. It's the imaging analog to a fundamental rule of mechanics: Always avoid using software to fix bad mechanics. Fix what's wrong first and then use software to boost performance beyond the limit of the mechanical tolerances. For a lot of the examples show here, it looks like BXT2 is just being used to try to get back to what the optics should have been producing in the first place. If that's the best you can do...well, so be it; but my advice is to avoid using BXT2 to give up on trying to fix fundamental optical problems. You'll get even better results if you first dial in your optics and then use BXT2.
John
***
Precisely this. I was just having this discussion with a friend where I tested multiple scenarios with data of various total integrations and also optical errors such as collimation and misalignment. Data that had the least errors and the longest integration had the best results using BXT. It wasn't even comparable. Data that had large error and/or shorter integration tended to enhance or affect areas that actually shouldn't have even been touched resulting in a largely different result. It really doesn't replace longer integration time, better seeing, guiding, focus, etc.. all it does is enhance. Good data will always be leagues ahead when paired with BXT than poor data paired with BXT
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
I'm in the process of setting up a FF camera with my old FSQ106N. And I have yet to do any tilt adjustment to it, so I thought I'd compare the raw and BXT2 versions to see the level of improvement. It is considerable. I will still manually adjust the tilt to make the raw system better, but this is great.
https://astrob.in/6kfdfj/D/ |
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
I am still a bit confused about how to use BXT2 with monochrome images.
Since it uses information that is in the different color channels, does that mean I need to recombine, run BXT2, then extract again?
And if the optical train needs to be the same to get this intra-channel information, I cannot refocus between filters?
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Michiel:
I am still a bit confused about how to use BXT2 with monochrome images.
Since it uses information that is in the different color channels, does that mean I need to recombine, run BXT2, then extract again?
And if the optical train needs to be the same to get this intra-channel information, I cannot refocus between filters?
If I'm not wrong, you should run the BXT2 on every channel only, if they were performed with different optics/cameras, in other cases use it on the combined image. Please correct me, if I'm wrong.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Michiel:
I am still a bit confused about how to use BXT2 with monochrome images.
Since it uses information that is in the different color channels, does that mean I need to recombine, run BXT2, then extract again?
And if the optical train needs to be the same to get this intra-channel information, I cannot refocus between filters?
I think you should combine and then run BX and I don't think any differences in the optical train, or in the seeing/focus, between channels matters. At least when I've tried running it on the channels separately, the result wasn't as good as combining first. Not sure what the 'official' recommendation is. You can also run BX on the Luminance only (like with traditional decon) and not see much loss in overall resolution in the final image.
Cheers,
Scott
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
Scott Badger:
You can also run BX on the Luminance only (like with traditional decon) and not see much loss in overall resolution in the final image.
But then why can you do that for the luminance channel, and not the other channels separately? BXT2 does not recognise whether it's a luminance, Ha, Red or Blue channel, does it?
It was recommended in the documentation to combine first, but it was always my impression (I haven't started mono processing yet, will soon) that each channel needs to be processed (decon, noise, star removal,... ,possibly stretching?) before recombining. There are many steps I would think to do before recombining that would intefere with the original signal and would therefore stop BXT2 from working properly.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
It seems to work on the separate channels, but I found the final result was better when doing it after combining. Not sure what the reasoning is, though. Regarding the other processes you mentioned, you wouldn't use traditional decon plus BX, just one or the other, and as I mentioned, typically traditional decon is used on the luminance only and not the RGB. Sometimes I apply noise reduction to the separate channels, and sometimes not until combined, but I don't know if it matters much. You definitely don't want to remove stars before combining since you'll want them for color calibration, and stretching should also be done after combining (and after decon/BX and color calibration).
Cheers,
Scott
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
John Hayes:
The discussion here is interesting. The sensor in my refractor sits about 0.5-1.0 mm too far out from the optimum position so my images exhibit just a little bit of astigmatic error in the extreme corners of 42 mm circle so I definitely appreciate the ability of BXT to fix that--particularly since I can't walk out to the backyard to readjust the spacing. One of the clever things that Russ has done is to make the sharpening algorithm non-shift invariant, which means that it works differently in different regions of the image. Round, tight stars are a big deal in astronomical images, but the thing to keep in mind is that it is always better to start with an optically well aligned system in the first place--particularly if you want to maximize image detail while minimizing artifacts. Yes, BXT2 does a spectacular job of fixing a lot of ills but I've always believed that it is best to start with the best possible data beforeusing tools like BXT/BXT2 to deconvolve the results. It's the imaging analog to a fundamental rule of mechanics: Always avoid using software to fix bad mechanics. Fix what's wrong first and then use software to boost performance beyond the limit of the mechanical tolerances. For a lot of the examples show here, it looks like BXT2 is just being used to try to get back to what the optics should have been producing in the first place. If that's the best you can do...well, so be it; but my advice is to avoid using BXT2 to give up on trying to fix fundamental optical problems. You'll get even better results if you first dial in your optics and then use BXT2.
John
Absolutely @John Hayes! With this new version I could now use a fr that I’m not happy with and will be returning for credit. Based on what I’m seeing here I wouldn’t need to worry about it and just image. Like you I prefer to start with the best I can before any software corrections are applied!
This said I’m going to try this out with some of my images from this FR just to see how much better it would be. Regardless of the results I’ll still return the FR but it will be an interesting project for this test.
Dale
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
John Hayes:
Round, tight stars are a big deal in astronomical images, but the thing to keep in mind is that it is always better to start with an optically well aligned system in the first place--particularly if you want to maximize image detail while minimizing artifacts.
I think beyond dispute, this is correct. But can it also be true that you get to a point of diminishing returns with increasing mechanical and optical perfection? That is, and particularly with newer and even better versions of software like this, will the corrections become good enough that going to the very highest levels of optical and mechanical alignment will not present an adequate return on investment, i.e., these improvements will no longer be detectable? Clearly, we are not there yet. While BX2 is very good, it does leave behind discernable artifacts that are evident on close inspection. But even now, I do think BX2 has made the demand for absolute mechanical and optical perfection a bit less important than it was.
|
You cannot like this item. Reason: "ANONYMOUS".
You cannot remove your like from this item.
Editing a post is only allowed within 24 hours after creating it.
You cannot Like this post because the topic is closed.
1.20
2.71
2.71
22.64
11.99
2.11
3.01
12.64
4.69
7.61
6.06
1.51
7.17
15.85
