GradientCorrection - PixInsight's New Secret Weapon? Tested VS GraXpert
ฝัง
- เผยแพร่เมื่อ 29 ก.พ. 2024
- Download the latest PixInsight and give it a go! :-)
Image shown as a reference for comparison 1 was from the supremely talented Adam Block -
www.astrobin.com/6ubf5c/
Check his channel here! - / @adamblock
Affiliate links - If you want to use these then it'd really help me out at no extra cost to yourself! :-) Thank you!!
These tools have changed the processing game for me, no joke.
BlurXTerminator -
www.rc-astro.com/software/bxt...
StarXTerminator -
www.rc-astro.com/software/sxt...
NoiseXTerminator -
www.rc-astro.com/software/nxt...
GradientXTerminator -
www.rc-astro.com/software/gxt...
StarShrink -
www.rc-astro.com/software/ss/...
First Light Optics -
www.firstlightoptics.com/tele...
365Astronomy -
www.365astronomy.com/askar-12...
HighPointScientific! :
www.highpointscientific.com/?...
or
bit.ly/3uzB4FJ
My 3d Print store -
My Ebay store - www.ebay.co.uk/usr/bargain_bu...
My Etsy store - www.etsy.com/shop/lukomatico?...
/ lukomatico
Gaming channel : / @lukomaticogames (will be reborn soon!)
Topaz Labs Affiliate Link! - topazlabs.com/ref/1133/
/ lukomatico
Amazon Affiliate Links! - Anything you buy through one of these links will give me a small commission at no extra cost to yourself!
Amazon UK - amzn.to/3EOCgpV
Amazon US - amzn.to/3quGPlD - วิทยาศาสตร์และเทคโนโลยี
![THE TOYS x NONT TANONT - ดอกไม้ที่รอฝน (spring) [Official MV]](http://i.ytimg.com/vi/pPa1d5cC8M4/mqdefault.jpg)
I saw this PI update in my inbox and thought I'd wait for you or Cuiv to produce a video on it versus Graxpert. Didn't have to wait long 😄
Appreciate your producing this considering you’re under the weather. Like others who commented, I just saw this in my inbox this morning and was about to investigate the Pixinsight channel….and, bam! there was your video! You’re always on top of this quickly evolving hobby and it’s much appreciated (saves us a lot of work). Hopefully I speak for everyone else at this channel when I say, take a break, get to bed and get better. We’ll all be here when you’re up to speed physically,
Thank you ever so much Joe, as always!!
So glad to see this update and comparison from you. Thanks so much, Luke! Awesome video and very helpful :)
Thanks so much Sarah!! :-D
Much appreciate your helpful comparison, Luke! Many thanks!
My pleasure!
Wow, you are definitely battling some serious gradient! Thanks for demonstrating it, knowing it will work on some of these means it will be a cakewalk for what some people see ❤. Showing the intricate difference side-by-side on the faint nebula is really what matters you chose great targets for that! I think all of these tools at this point can remove the negative grade, but the real battle is which one keeps the most “good” data!
Thank you ever so much mate, I appreciate your support!! 👍👍
As soon as I got the Pixinsight notice I wondered about a comparison between the new gradient tool and Graxpert. Then, in no time, I got your notice of a comparison. Good job. I'm convinced. Sorry about the weather and you being under the weather. But, you've made the best of it 😊
Glad it was helpful, Mike! Thanks ever so much my friend 👍
Awesome Luke! Pleiades image, GC overcorrected near the "arms" of nebulosity in the center (creating those dark lanes)! I think changing the thresholds or smoothness may help!
Thanks so much mate!! :-D good eye!!
Hey Luke! This looks very interesting! I have tested it on my VDB-152 as well and had similar results but some others are a bit mixed. I suppose as this tool evolves, it will get better over time and as the MARS project builds up more data to work with. The most interesting thing about this is that it is not a tool based in AI! Man you've had it hard this year with illnesses :-( Hope the next thing you catch is a comet :-D Clear Skies mate!
Great comparison Luke. Every time a new tool comes out I start to reprocess previous images. Gives me something to do on cloudy nights! Gradients have always driven me crazy. Gradient Expert is pretty good but sometimes too aggressive. I’ll try the new tool in PI. Finally had a couple of clear nights (after more than 2 months of clouds). Imaged Thor’s Helmut and the Headphone Nebula. Stunning! Hope you are feeling better. Dr B from Manitoba, Canada 🇨🇦
Well, this is exciting. Thanks, Luke.
(Still extremely cloudy here in Savannah too)
Hope it clears soon Pat!!
Thanks! There’s not a chance my own comparisons would have caught the differences yours did. Cloudy here in SoCal too but someday I’ll put this to use.
That's really kind of you my friend, thank you!! :-D
Graxpert saved my Horse Head, that cursed star actually gave me a reflection of the Artimus-C Pro sensor on the back side of the filter. Had a nice square around Alnitak wich Graxpert was able to remore.
The latest GradientXPert AI was trained on seperate channel data so after cropping the very first step was to apply to each channel before anything else. Also with narrowband manually setting sample points made a huge difference. That said the new Gradient Correction tool is simply better and ridiculously fast. Yes the last year in Pixinsight and 3rd party tools and scripts has radically improved end results to the point we can deal with rubbish weather by reprocessing old data. Great work and yes ABE and DBE are past their used by date. Thanks for the full comparison tests.💫👍
Thanks for the comparison Luke! It looks like either tool will do a really good job and with some fine tuning I doubt that I could pick a winner. As I'm used to Graxpert I will probably just continue to use that tool, but I will give PI's GC tool a spin as I think the fine tuning may be a little easier having watched their instructional videos. As always a great comparison with a range of targets being corrected.
Thanks for that Luke, these comparison videos are very useful. I use GraXpert standalone and I was glad to see it compared well with an internal PI tool. I'll buy PI one day I am sure, but just happy to play with freebies for moment. Hopefully you get some clear skies soon!
Hi John, I am a newbie since Jan 2024 (Seestar S50) and spent a couple of weeks with Siril and GIMP with good results. My only issue with Siril is you can only have one image open at a time. However it is free and worth starting out with. Many of the tools are applicable to other programs too. But I decided to invest the £255 - UK in Pixinsight (and I also went with BlurX, NoiseX and StarX). PI may seem expensive but over all there are many other free tools available for PI and once you get past the three weeks initial familiarization it is a fantastic piece of software that quite honestly is a bargain (no nonsense monthly rental). I would recommend the jump to PI. I compared stacking in Siril with PI Batch Weighted stacking in PI, PI produced a lower noise image and the autocrop is awesome.
Was looking forward to this, I saw the Pixinsight video and knew you would have a comparison video up tout suite👍To my eye GradientCorrection just has the edge on this particular image but it is pretty close. I can feel another round of 'reprocessing' coming on to test this out on my images. It's great to have another option for gradient removal as it's one of the most annoying parts of image processing. I think it's fair to say that the likes of GraXpert has probably prompted the PI team to come up with their own version which can only be a good thing.
Agreed mate! I can't wait for the MARS project to come out which should replace them all in theory, I do hope it's user friendly though 😅
Hi Luke great comparison video mate, the improvement in software over the last few years is unbelievable and it just keeps getting better I think that the software we use to edit our images is now more important then the equipment we use, hope your on the mend mate , will be good to get some clear skies so we can all see the stars again
I think you're right Tony! Past a certain point of equipment quality the software and processing matters more than the gear!
Thanks for the comparison, Luke! I noticed on the Lobster Claw image in the lower left-hand corner GraXpert again removed what looked like some nebulosity, but that was retained by PI GradientCorrection. (Mind you, my monitor is not very good and I'm watching on TH-cam, so I'm not certain what I was seeing was correct.) It might just be because the tool is new and I've only watched two of the videos quickly, but the PI's GradientCorrection seems a bit complicated to me. I think they should have gotten an AI engine to calculate the optimal settings for PI's GradientCorrection prior to running it on an image! 🤣 Appreciate you doing this comparison.
Hey Greg! - I think that would be a great compromise, having AI to analyse the image to correct the input on the tool! 👍
@@lukomatico I think so too, Luke! But I have to say, the PI folks seem to have gone out of their way to emphasize the 'natural intelligence' of the GradientCorrection process so they may be averse to even thinking about AI! 🙃
Well done Luke, top notch comparison, and it looks like PI have a valid competitor to GraXpert, and simple to use too…👍🏻
Thanks mate!! 👍
I agree with you, Pixinsight’s native tool looks better to me. Great video! You sold me. Clear Skies!
Great comparison mate I've tried Graxpert but couldn't get the thing to work which is probably due to my old PC lol, hope you get well soon mate clear skies
Thanks so much mate!!
I think GraXpert is awesome for what it is - a completely free tool that can be used standalone. But, I've noticed that it more often than not add artifacts to my images so I've used it with caution and instead stuck with ABE. This new feature in Pixinsight looks really promising. Thanks for the comparison you did. I was really curious about the actual side-by-side differences myself.
Nice comparison !!!
Glad you enjoyed it!!
I literally bought graxpert like three days ago not knowing pixinsight was getting ready to launch something else LOL If I had known, I would have obviously waited
Thanks for the ‘heads up’ Luke. Might be worthwhile saying that even if you have the latest version of PI you need to update it as, confusingly, the update still has the same version number (all be it a different build number).
Good shout!! Thanks mate 👍👍
I had a little play and found Gradient correction better for complex gradients with adjustments made after viewing the videos from Pixinsight's youtube page and Sasha's youtube page (View into Space) with the fine tuning you get some amazing results with Gradient Correction. Also I had a play with using both on one image and it seemed to work well. Graxpert still has its used but I reckon Gradient Correction will be a winner.
I have installed the program in the latest version of pixinsight, and the image looks very similar to the dbe, but later this image gives me problems sometimes with the autostretch, and other times after doing colorcalibration if I do a scnr to eliminate the green noise it makes me the purple image, then I used a duplicate image from the graexpert but with photometriccolorcalibration, and from then on it didn't do any more strange things to me
Hi Luke, really interesting comparison. I haven't tried gradient correction yet but if it's even better than graxpert it must be good ! Crazy to think just a couple of years ago we were all happy with DBE or ABE. Probably the one to top them all will be the conclusion of the Mars project. If its not serious questions need to be asked in the pixinsight office 😂
Absolutely agree Paul! I have serious concerns that the MARS tool will not be user friendly though if every other tool in pixinsight is anything to go by 😅 it's like they deliberately find the most awkward way to place, phrase and control things - unreal 😂
Hey Luke, mice job as always!
I don't think it invalidates your results in any way, just an observationIn: in SPCC you have selected "background neutralisation" with sample coordinates but in the video you are not selecting starless background sample. If the coordinates you were working with are coming from an old processing then potentially your background may be sub-optimal. Worth checking.
Thanks for the tips mate! All these images were freshly solved for this video, I don't know if that helps 👍 Cheers!
Extermination on the stars could have given us a better view on how well these different tools preserved the nebulosity in your image but nice comparison. I stick with the APP gradient removal tool as I never found Graxpert and other tools consistent enough.
Thanks!
Thank you ever so much my friend!! That's really kind of you 👍👍
Very difficult targets to nuke gradients!! Thanks for the video dude!
Thank you very much for this instructional video, Luke. I had always done background extraction before SPCC, I see in this video that you did it the other way around, is your method correct procedure?
Thank you for watching!! :-) RE: correct procedure, I've tried both ways and don't find a massive difference really so I couldn't really say - great question though!
cheers Luke :)
Cheers mate! Hope you've been well 👍
I had a hunch something would be coming out soon for gradient. Sort of thought it would be RC Astro. I wasn’t a fan of Graxpert, so real happy to see this now.
Neat.
Thanks! Currently I use ABE in PI, would this be in addition to that or instead of that (on an average image without a ton of gradient)?
Cheers for watching! - it would entirely replace ABE or DBE in my estimations :-) Clear skies!
I had good results from both the new Gradient Correction and Graxpert. I have found that if I use GXP with smoothing on 0 it can remove more nebulosity/slightly reduce the intensity of the nebulosity than when using smoothing on 1. So buy default I start with smoothing on 1. Also the new GC does not show the background removed which would be useful.
Great tip mate!! Thanks 👍
I don't use PixInsight but I am thinking about it, mostly because I can't use Blurexterminator on Affinity Photo.
For gradient removal I currently use RC Grad Exterminator ( which ain't cheap ) have you ever done a comparison with this and Graexpert?
This weather in the UK is unreal at the moment, I know it's never that good but I can't remember so few clear nights as I've seen over the last six months. Same goes for colds and flue, they seem to go on for months. Anyway, take care Luke, hopefully things will improve soon.
Thanks mate!
I've never tried gradient xterminator before so can't really say there, sorry!
Hope this weather improves soon for us all, whew it's been bad
If the settings were adjusted on Graxpert to not be as aggressive , the results may be the same? - nice eye opener Luke, im sure we will all use both now and compare.
Very likely mate yeah, it'd take a bit of fiddling but I'm sure they could be brought closer :-)
Wow looks like someone had early access. Great to have choice innit. I tried it late last night the new tool on default setting removed less nebulousity I think
Thank you mate! No early access here haha, just insomnia 😂
How to get gradient correction tool? Is it Gradient XT? Thanks
Its within pixinsight itself on the latest updated version 👍 sorry about the late reply!!
The true test, for this is going to be eliminating the gradients in broadband mosaics
I would suggest locking the channels in the STF after applying SPCC. Otherwise you are stretching the channels individually, thereby screwing up the color correction and most likely introducing the green cast you were getting.
Good shout! I tried linked and unlinked off camera though and it didn't make a difference on these once SPCC was complete, I'll keep it in mind though, could be just this data I chose!
Imagine taking flats like usual but the next morning you go to calibrate, register, and stack your complete data set. And suppose the flats don’t perfectly correct out the dust donuts. Can GC correct for those?
It can't correct for that kind of gradient unfortunately! great question though my friend!
Why did you run SPCC first?
Good question! - I ran SPCC to make sure the images being shown were well colour calibrated before attempting a comparison of the gradient removal, - without doing that it would slightly bias the results towards GraXpert. Cheers!
the new tool looks very capable, and seems to better than graxpert in just about every scenario, that last lobster claw example you showed was the first time i saw something to the contrary but i wonder if that would’ve been different with some fiddling with the sliders. excellent for those times when graxpert absolutely shits the bed, which feels like it’s been happening more to me lately. to be honest as nice as this new tool is i’m just waiting for the MARS project to conclude!! i honestly think that will be such a game changer, i think it’ll completely revolutionize amateur astrophotography and everybody who uses it will have leagues better images. hopefully that comes soon and you get some clear skies when it does!
Hey mate! - Totally agree really, all very fair! :-D I too can't wait for the MARS Project to finally come out, should be a real game-changer I'm hoping!
If you’re going to compare to tools like DBE, you should probably try at least a little bit to use DBE properly. That was sloppy.
That said… there’s something to be said for a tool that does an excellent job with less time and less fuss.
I think I did a good job of giving all the tools about the same level of effort, -not much!
Besides, there's a near zero chance DBE would top these tools anyway even with a perfect, laborious application.
I agree though, the less fuss the better in my opinion!
Clear skies!
There’s a good chance DBE could have given you some of the best correction of the lot if configured properly, because it allows you to be highly precise about what is sampled and how.
That’s the thing, though: DBE is a tool for a certain sort of job where the fuss is warranted. It gave you a poor result in this case because you specifically used it incorrectly, haphazardly dumping samples on areas of relatively stronger nebulosity.
Not a big deal at the end of the day, but if you’re going to educate with a comparison video like this, it makes sense to compare things properly. DBE can still produce superior results in the right use cases.
Extermination on the stars could have given us a better view on how well these different tools preserved the nebulosity in your image but nice comparison. I stick with the APP gradient removal tool as I never found Graxpert and other tools consistent enough.