StarTools 1.9 Beta

[decay]

I didn't realize finishing the new multi-lingual was still an open item too.

Oh, not to be misunderstood ... 99.99 percent of the work is finished. There are only some few missing translations due to ongoing work on SVDecon and some fixed texts which we discovered too late, texts in progress bars etc.. We are of course used to see the English texts and so it’s easy to overlook some of them ...

Anyway, for me the most important part is not so much the UI with the texts on the controls, but more important all the excellent hidden documentation and help texts that are available for all modules and controls. I guess most Spanish or German speaking people doing AP will at least have basic skills in English language, enough to work with the ST UI itself. But in order to understand how all this works, it is important to understand the documentation and the help texts. And these are written in sophisticated English language (at least as far as I can judge ) and that may be a problem for some non native speakers.

So with this excellent translation work Jochen and Carles did, I hope that ST will be accessible to a wider audience.

Best regards, Dietmar

[Mike in Rancho]

... almost forgot. No thread is complete without the ai starless version:

This made my morning when I read it.

[alacant]

ringing around scattered super tiny stars against nebulosity.

Hi
Problem
I've noticed that too. Have you found a workaround or set of values (maybe in shrink?) to offset the ringing?
This is an example. Yes, it's enlarged beyond normal but it shows the effect I think you're describing; small stars against nebulosity.

Workaround
We find that it can be fixed by increasing the colour radius in Shrink. The problem is getting all the small stars in the mask. Stars which are included in the mask for Shrink merge well with the background. The smaller stars which are not masked remain with the dark ring around them. Playing around with sensitivity, even on the most sensitive setting, the smaller stars will not be detected.

Processed file here: https://photos.app.goo.gl/kUyvc2na1mJqApFX8

Cheers,
Steve

p1.png (223.57 KiB) Viewed 117958 times

[Mike in Rancho]

Hi
Problem
I've noticed that too. Have you found a workaround or set of values (maybe in shrink?) to offset the ringing?
This is an example. Yes, it's enlarged beyond normal but it shows the effect I think you're describing; small stars against nebulosity.

Workaround
We find that it can be fixed by increasing the colour radius in Shrink. The problem is getting all the small stars in the mask. Stars which are included in the mask for Shrink merge well with the background. The smaller stars which are not masked remain with the dark ring around them. Playing around with sensitivity, even on the most sensitive setting, the smaller stars will not be detected.

Cheers,
Steve

Hey Steve,

Yeah that's pretty much the post-SVD culprit there. One of them anyway. I do use Shrink to touch this up when I am ready for a final version, usually at the very end. For interim images and sometimes even finals, I'll just leave things be. Depends how obnoxious it looks, and occasionally those little dark rings sort of kind of help little stars pop out. Not that it's proper, but sometimes doesn't look bad.

What I do though is just a fully manual mask, no actual shrinking iterations, just deringing pixels usually 1.5 to 2.0, and I zero out the color tame as I don't want to alter any of that up.

It's a fair bit of elbow grease to accomplish, as one has to be zoomed in and then just pan across and down the image hunting for all the little rings. Also, these stars are so tiny that the mask controls aren't quite up to the task. I find using the drag-circle option quickest, though that's still awkward. What would be best would be a variable sized circle and you'd just apply it with a click. Like toggle single pixel, but that's too small.

I have had difficulty getting all these tiny stars included in a mask using the various automask settings. It still often misses them, or worse ends up including non-stellar features. Really I don't even want all tiny stars masked, just the leftover ringing ones. Reason being that the deringing actually imparts a bit of a fade/blur around the target mask. Perfect for remedying the ringing, but I don't want to fade and blur other features or the tiny stars that don't need it. Thus, fully custom mask.

It's really not too big of a deal, just a little task to do. Though some heavily nebulous images can be more work.

[admin]

Small update with some bug fixes, UI fixes and getting rid of (hopefully) the last "UNKNOWN" log entries.

[hixx]

Hi Ivo,
I still see the 1.9.559 beta 3 version from 9/25 on download page.....
Regards,
Jochen

[Mike in Rancho]

Small update with some bug fixes, UI fixes and getting rid of (hopefully) the last "UNKNOWN" log entries.

Hi Ivo,
I still see the 1.9.559 beta 3 version from 9/25 on download page.....
Regards,
Jochen

I got the new one downloaded and ran a bunch of tests. Maybe some kind of browser cache thingy, perhaps force a page reload or delete history, something like that?

[admin]

Small update again (1.9.561 beta 5). Chasing some odd bug that passes file dialog clicks onto StarTools on some OSs.

For any macOS folks running Ventura and later (e.g. @hixx?), I also added a new way of unquarantining that I'd like to roll out if it works reliably;

• control + click on StarTools.app. This should bring up a menu, from which you will want to select "New Terminal at Folder". A Terminal window will open.
• now type 'sh u' (space in between 'sh' and 'u') and press tab. This should auto-complete a command "sh unquarantine_ventura.sh". Press "return" to execute it.
• The script will tell you what it will be doing (which you can verify by opening the file with a text editor). If that's ok with you, enter your password so it can complete the unquarantining.
• It should return to the command prompt without errors. StarTools has now been unquarantined. You can close the window, and you can now open StarTools.app like any other application.

[hixx]

Hi,
thanks for the straightforward procedure
the unquarantine process worked for me with following changes:
1) The "Terminal in Folder command is hidden in a sub menu. In the menu that opens upon Ctrl-click, You'll need to select "Dienste" (services) first to open the sub menu
2) sh u didn't auto-complete, I had to copy/paste the full command sh unquarantine_ventura.sh into Terminal window
3) executing the command required me to give the admin PW (should be the admin login PW of your Mac)

Clear skies!
Jochen

[Mike in Rancho]

Guy's new SVD User Notes got me thinking about deringing again. This of course is always dangerous.

We corrected the line about the apod/deringing support mask no longer being modifiable in Mask. There are probably good reasons for that I'm sure, with the way SVD now handles automated box and outline sizing. But as everyone has noted sometimes things get left out, other times perhaps included where not wanted, when it comes to deringing. Maybe moreso with weak datasets. And hence a desire to be able to help deringing out in a targeted manner.

But then it struck me that starfishy sampling, while better for PSF modeling, may not be the best for deringing support.

Would split masks be a viable solution? An automated apodization mask, possibly unmodifiable so that we can't mess up SVD doing its deconvolution thing using starfishy shapes, but then a deringing support mask, which the user can customize in Mask, and maybe with a starting point of normal circular auto star creation.

That would allow us to add missed stars, touch up errors like half-selected stars, and wouldn't accidentally include unwanted regions for deringing (i.e. fuzzing up a bit) if caught up in a big or blended starfish area.