Statistical Stretch Script: My new script to dynamically calculate & perform astro image stretching
ฝัง
- เผยแพร่เมื่อ 20 ก.ย. 2024
- Where you can find my scripts: www.setiastro....
Repository URL for pixinsight: raw.githubuser...
My newest script will dynamically determine how to stretch your astro image based on the images statistics. That way it will hold the background to a user specified level, increase the brightness of the true signal and prevent the image from being over stretched. A very common thing seen if anyone has ran STF on a starless noiseless image!
If you have questions or comments please leave below! They are all really appreciated!
Looking for astronomy gear?
Use my links to Agena Astro or High Point Scientific and support the channel!
agenaastro.com...
www.highpoints...
![HIWWHEE | OHM [Official MV]](http://i.ytimg.com/vi/OJJLtTLqLag/mqdefault.jpg)
If you havent upgraded to PixInsight build 1605 (which came out in March) I highly recommend you do that as all scripts are going to require signatures going forward.
If you still run into a parsing error or signature error you can download the zip file directly from my website and then just extract the .js file. You can then go to Scripts -> Feature Scripts -> Add Script and select the folder you extracted the .js file too.
It still has an error
run --execute-mode=auto "C:/Program Files/PixInsight/src/scripts/StatStretchV1prod/statisticalstretch.js"
Processing script file: C:/Program Files/PixInsight/src/scripts/StatStretchV1prod/statisticalstretch.js
*** Error: Signature verification failed for 'StatisticalStretch': Unknown code signing identity 'FranklinMarek': C:/Program Files/PixInsight/src/scripts/StatStretchV1prod/statisticalstretch.js
@@TevisC please remove the .xsgn file. Seems your pixinsight authorship library wasnt updated recently.
I'm running the latest version of PI 1.8.9-2, and It still won't load for me even after removing the .xsgn file.
Can you send what error you are getting? You can also email me which is a big help since we can send screenshots too
@@setiv2 Disregard my above post, for whatever reason it loaded and is working now - thank you.
Been using PI for more than 10 years and used about a dozen quick stretch methods of all shapes and sizes and then this simple to use script comes along and does a better job out the box than the lot! Super work, looking forward to your next innovation! Many thanks.
That is great to hear! I appreciate it!
I just added your scripts to PI yesterday and both “Statistical Stretch” and “NB to RGB Star Combination Tool” are fantastic! Statistical Stretch gives me better results than EZ Stretch and it’s much easier to use than GHS. Your NB to RGB is great too! Thank you.
I've been using ez soft stretch because it was super quick and I was too lazy for GHS, then I found this, which gives me the best of both. Extremely good stretch, quickly and with ease.
Thank you so much, this will probably improve my images by quite a lot with the ability to stay lazy
I never had great initial stretches with GHS. I hope this gets you 90% there then you can use curves or ghs to finish your artistic tweaks!
That is freeking amazing. Great video on it, thanks for developing this terrific tool.
Very nice. And the “under the hood” description, I think, is the best way to learn how to really use a script. So awesome. I’d love to see this kind of walkthrough for other commonly used scripts/processes.
My first trial worked great! Thank you for your hard work and contribution. I will enjoy this script a lot.
Awesome!
I finally got a chance to play with this script on a recent M101 image that I took. This is a great script, I really like it a lot! Thanks for your effort.
This script is great. It will be my go to way to stretch images from now on. After stacking and cropping, graXpert, blurX, noiseX and then your stretch script! Excellent job! Thanks
Great to hear!!
That's a realy nice script. I usually start my stretch with Masked Stretch, and this did a great job much faster and easier. Thanks
Great to hear!
Downloaded this script last night and it worked great on my LRGB M33 image. I always have trouble stretching galaxy images, but this worked so well! One area of improvement would be to have an interactive preview of some kind. When you launch the script, you cannot undo/redo changes on the image whilst the script dialog box is open. As a workaround, I create a process icon using the triangle and adjust the settings in this window whilst testing on some previews. It's quicker than re-launching the script. Minor comment aside, thanks for making a great script and making in available free of charge Franklin. Subscribed to the channel and sent a donation via your website :)
Wow thanks so much! Real time preview is beyond my coding experience right now (I am not a coder and these are the first times i have done anything in javascript and learning as I go). That is a future goal for sure. I do know my math though, while the dialog is still open just change the target median to whatever new value. It re does the calculations and adjusts the image to your new target value. No need to exit out.
@@setiv2 Ah that's cool! So if I have performed stretches by adjusting the median, and also change sigma boost, does that compound with the existing stretch(es) or does the script work to give the same final result irrespective of how many times I executed it? As you say, this would avoid having to close and reopen the script. Nice work on your first script too, I'll be checking out the Star Stretch one next :)
It does compound. If you start with no optional sigma boost then all the math is completely reversable. I would suggest executing it until you find the median you like, then click the sigma boost for one final execute. I would also suggest my NB to RGB stars script as we come into nebula season. No need to shoot seperate RGB stars anymore :)
Looking forward to trying to this out asap.
It is certainly easy to use and fast. I will have to play around some more with the adjustment sliders but I redid an old IC1805 using .25/.1 and it turned out well after some curves to blacken the background. EdgeHD 11/6200MM. Thank you. Great explanations!
Brilliant. I’ll be using this script in my workflow.
Awesome Script Seti! Thank's a lot for the work!
Cannot wait to try this out. Thank you so much.
For many years I’ve struggled with STF and Hyperbolic stretch. Then you came along to save the day! Thanks so much for this. I’m going to redo all my images. Can’t wait for Nebula season as well to stretch new objects. Clear skies and thanks again for this game changer!
You are so welcome! That is really great to hear!
Really great script and really well explained thanks I will be using going forwards
Great to hear! Always open to suggestions and improvements too.
Thanks you all the great videos you make! I'll try this script! Thanks!
Thanks so much for the script, this is perfect and exactly what I was searching for.
Amazing tool, very innovative! Thanks for this Frank.
Love it! Thank you Franklin.
This is really awesome!
this is amazing... really appreciate your efforts!!
Glad you found my scripts!!
@@setiv2 Not as glad as i am :)
Great work! Thank you. Looking forward to test it out.
Great script, thank you
This is a very cool little tool you have there, thank you for sharing!
Very nice. I will try this on some M101 data I am currently integrating. Thanks.
Sacré boulot encore….merci beaucoup .
looking forward to try it.
Cool stuff! I will for sure test it out!
Looking forward to any feedback from your tests!
Very nice! You have some cool scripts. Thank you!
Thanks! This should fit in right in, in time for nebula season with my other scripts.
I like this a lot. Thank you. Maybe you can do a video that shows a comparison between your model and GHS.
GHS is just a different tool honestly. What I started doing is this as my initial stretch and then GHS to do a final tweak to highlight whatever I want to focus on.
Great work, I love your approach and explaination.
Some questions to better handle the script
Do you recommend to stretch starless image or not necessarily ?
My osc images looks greenish after the script, shall i always colorcalibrate it first ?
No option to color balance like ez stretch ?
With OSC with an IR cut filter you will always want to run SPCC before any stretching. If you are using narrowband filters you would extract and work all the channels seperately. You dont want your stretching solution to do the color balancing. Stretching should preserve the color balance. I always remove my stars and stretch the starless and stars seperately (hey I got a script for star stretching too :) ). General workflow I follow is. GraXpert (or other gradient removal) -> BlurX -> SPCC if RGB -> StarX -> NoiseX -> Stretch.
Great work
Very impressive! Thanks!
Adding the script as an instance icon bakes the settings into the process. Calling it globally via the script window displays default values but upon executing, the settings that were initially set in the instance are used.
Example: open process, change Target Mean to 0.1. Save process icon. Double click icon and execute globally (adding a view string to circumvent the error message of missing view). The script loads and displays 0.25 (i.e. default values). But executing uses the initial values that were set in the icon instance, in this case 0.1.
You cant run this script globally. You can also just double click any parameter to vhange the value when you double click the saved icon. Sure there arent sliders but it does the exact same thing as the dialog.
Nice work, I have to try that !!
how did you gain 100 subscribers overnight 😭😭😭
Would love something like this in Siril 😅
I have never used Siril. Can you code within it?
@@setiv2 you can create scripts, but perhaps not quite to the level in PI. Building a script in Siril seems intuitive enough, but after I spent several hours over the last few days trying to figure out how to attempt your statistical stretch script in Siril and made little progress... Welp 😂
Very nice but i have got a problem of resolution. I can't read the explications above. I have the latest version of Pix ang working with my PC in 4K. Kinds regards
Is the instruction box text too small, cutoff, showing a scroll bar? You can also email me directly with a screenshot of what you are seeing. That may be more helpful
Awsome! Could this be integrated in Siril too?
I have never used Siril. If Siril allows for javascripting I dont see why not. There isnt anything too exotic in this script. You would have tp find someone familiar with Siril though that I could work with.
Parsing repository update information: Parsing Signature element (line=1 offset=0): Unknown code signing identity 'FranklinMarek'
If you run into a parsing error or signature error you can download the zip file directly from my website and then just extract the .js file. You can then go to Scripts -> Feature Scripts -> Add Script and select the folder you extracted the .js file too
@@setiv2 run --execute-mode=auto "C:/Program Files/PixInsight/src/scripts/StatStretchV1prod/statisticalstretch.js"
Processing script file: C:/Program Files/PixInsight/src/scripts/StatStretchV1prod/statisticalstretch.js
*** Error: Signature verification failed for 'StatisticalStretch': Unknown code signing identity 'FranklinMarek': C:/Program Files/PixInsight/src/scripts/StatStretchV1prod/statisticalstretch.js
And there is no .xsgn file in that folder? You can also email me which helps since we can share screenshots to troubleshoot
@@setiv2 will it still get an update notification when installed via the zip package?
@@MarcelloBarile no. You need the repository to get the automatic updates. What you need to do is upgrade pixinsight to build 1605 (the latest build that came out in march). That will solve all your signing errors.
I think there is an issue with the repo, I get this error:
Unknown code signing identity: 'FranklinMarek'
You need to update pixinsight to the latest build 1605 that came out in march. You get it from the pixinsight downl9ads page
Do you stretch each Mono channel individually with the tool? Or do you clean the channels up, combine and use your tool?
That really depends on your target. Nebula you can stretch seperately until you like how they look, for galaxies you probably want to combine and color calibrate prior to stretching
too strong. Congratulations😉
Can you confirm that your install works on Linux Mint. I'm having the same signature problem and deleting the signature file doesn't fix it. I'd really like to give this a try, but can't if I can't install it.
The operating system shouldnt matter as it is coded in pixinsights javascript environment. The signature problem shouldnt show up at all if there is no .xsgn file as pixinsight wouldnt have anything to check then. Best i can recommend is download the zip of the script. Extract it. Delete the .xsgn file making sure there is only one .js file in the folder and then go scripts -> feature scripts -> add script. Point it to the folder with the sole .js file in it.
I have just updated my repository with all your scripts and have quite a bit of data to experiment with. I have an OSC camera with 2 x NB filters (Ha/Oiii and Sii/Oiii) and using DBExtract to get to RGB (with an optional SHO image). Would you recommend using this script on the individual RGB images, or after recombination to SHO? Many thanks for all you do.
You would use this at anyplace in your workflow where you go from linear to nonlinear. This should be your first stretch. Usually you stretch the RGB data seperately from your NB data prior to combining them
@@setiv2Thank you.
It looks like all these images are monochrome in nature. What does this stretch look like for OSC images?
I perform a stretch on a color image in the video as well. Go to timestamp 13:05
This is super! For an RGB image, would it be best to separate the channels first and then run a Linear Fit process to equal out the colors before doing this Statistical Stretch?
No, linear fit is a super outdated method. Run SPCC (use my new script Find Background to get your ROI for the backgroundneutralization section of SPCC) then just run Statistical Stretch in the Linked mode.
@@setiv2 Thank you very much for the quick reply. I tried your Find Background, Stretching Stars and Stat Stretch. All three worked incredible.
No idea why but this script just turns everything green. STF does not do this
You need to color calibrate first. My guess is you were doing an unlinked STF. This script necessarily preserves color calibration.
@@setiv2 Ok won't work for me then as colour calibration keeps failing.
I was unable to perform any stretching using your script - it says in red in the console:
*** Error: Invalid character in expression:
BlackPoint = iif((med($T) - 2.7*sdev($T))
Very weird. What version of statistical stretch are you using? What version and build of PI do you have?
@@setiv2 I downloaded it from your website yesterday Statistical Stretch V1.5. PixInsight 1.8, I do not know how to find out the build.
@@setiv2 Statistical stretch V1.5. PixInsight 1.8, build I don't know.
your repository link is not working in pixinsight. Is there an alternative URL? thanks and great job on the scripts
If you havent upgraded to PixInsight build 1605 (which came out in March) I highly recommend you do that as all scripts are going to require signatures going forward. You do this from the PixInsight download site. This will fix the error if you are trying to use the repository. Also make sure you copy the full URL. I recommend going to my website and copying the url from there.
If you still run into a parsing error or signature error you can download the zip file directly from my website and then just extract the .js files (delete the .xsgn file). You can then go to Scripts -> Feature Scripts -> Add Script and select the folder you extracted the .js file too.
Let me know if this helps
Frank
Great video. When I perform a stretch, the result is a white screen. What am I doing wrong?
Is your STF still on?
Do you give me permission to show your script to the French community out there , on my TH-cam channel ?
Absolutely!
Just tried this on M51 from a ZWO-ASI533MC-PRO OSC. It produces a heavy, dark green cast. Any ideas why or solution?
Did you color calibrate first using SPCC or BackgroundNeutralization -> ColorCalibration?
@@setiv2 just tried SPCC first and problem solved. Initial image looks good now. Will play with other parameters now to find optimal setting for. M51. Thanks!
The stretch preserves whatever colorbalance you have in your image. Glad you got it balanced with spcc and stretched!
Thank you for sharing. It appears your scripts do not work. See error message below:
StarStretch2.1signed/star_stretch_v2.1.js
*** Error: Signature verification failed for 'StarStretch': Unknown code signing identity 'FranklinMarek': /Users/karelyade4ever/Desktop/StarStretch2.1signed/star_stretch_v2.1.js
run --execute-mode=auto "/Users/karelyade4ever/Desktop/StatStretchV1prod/statisticalstretch.js"
Processing script file: /Users/karelyade4ever/Desktop/StatStretchV1prod/statisticalstretch.js
*** Error: Signature verification failed for 'StatisticalStretch': Unknown code signing identity 'FranklinMarek': /Users/karelyade4ever/Desktop/StatStretchV1prod/statisticalstretch.js
If you run into a signature error your pixinsight authorship list wasnt updated yet. Please remove the .xsgn files. If you downloaded the zip file only extract the .js file and delete the .xsgn file. If you downloaded from the repository you will have to find the .xsgn files in C:/ProgramFiles/PixInsight/src/scripts
Perfect. Will try that tomorrow. Again, thank you. Can wait to try your script.
@@setiv2 It work! Thank you.
@@karelmachado8535 great to hear!
Hi, nice. How can i resolve this error please : Error: Signature verification failed for 'StatisticalStretch': Unknown code signing identity 'FranklinMarek'
Seems like your PixInsight isnt updated. You will need to go into C:/ProgramFiles/PixInsight/src/scripts. Find the StatisticalStretch.xsgn file and delete that, Do Not delete the StatisticalStretch.js file though
@@setiv2
My PixInsight version 1.8.9-2 build 1601 and not have the StatisticalStretch.xsgn under C:/ProgramFiles/PixInsight/src/scripts
@@setiv2 Solved, i installed manually without xsgn file
Glad you got it working! I havent had anyone else say they ran into signature errors. PixInsight did push out a new author list when you check for updates that your computer may have missed.
@@setiv2 github link is broken for me.
Did not work--both methods!
What error are you receiving? Also do you have the latest pixinsight build 1605? That is going to be needed as all scripts are started requiring signatures.