white pixels after Decon

[admin]

Yes, 314 is fine.No white dots Thanks.
Is it possible to use the alpha with my licence? I copied and renamed the 1 3 5 289 and copied the 314 to the same folder but I still get the evaluation copy.
Thanks

Great! Your workflow + data was a great help. Yes, you should be able to use your license just fine; simply copy it in the folder where 1.4.314 resides and all should be good (also see original StarTools welcome email, which states the same thing).

[alacant]

Sorry. It's back on 4 312. Same data:
crop
autodev
wipe dark anomaly 8
develop 90%
contrast
sharp
HDR
Decon

and the whie pixels are back:(

[admin]

Grrrr...

Can you try the latest version (1.4.315) in the download section?

This Tracking/sequence independence thing is tricky - admittedly it's a monster of my own making.

No other software (or sane software developer) would let or encourage you to apply deconvolution after so many other operations (stretching, sharpening, HDR). This is since, normally, it requires linear, unadulterated data (you're trying to reverse a point spread function that is evenly applied across the image - modify the image and that point spread function becomes unrecoverable as it will start varying all over the place).

StarTools tries to temporarily reverse all other operations that were applied since the data was in that virgin state, apply deconvolution, then reapply all the operations again as if deconvolution had always been in place; changing the past to influence the future so to speak. Getting just one small thing wrong and the whole thing falls apart...

Let me know how you go!

[alacant]

This Tracking/sequence independence thing is tricky - admittedly it's a monster of my own making.

No other software (or sane software developer) would let or encourage you to apply deconvolution after so many other operations (stretching, sharpening, HDR). This is since, normally, it requires linear, unadulterated data (you're trying to reverse a point spread function that is evenly applied across the image - modify the image and that point spread function becomes unrecoverable as it will start varying all over the place).

StarTools tries to temporarily reverse all other operations that were applied since the data was in that virgin state, apply deconvolution, then reapply all the operations again as if deconvolution had always been in place; changing the past to influence the future so to speak. Getting just one small thing wrong and the whole thing falls apart...

Ah, OK. So Decon should be applied before Autodev-ing? In your quickstart guide it is applied after. Just following that ATM.
Anyway, I tried it with the original wf and no white pixels with 315. From now, I'll Decon at the beginning (?). Maybe add this to the quick-start guide?
Thanks for your help.

[admin]

Ah, OK. So Decon should be applied before Autodev-ing? In your quickstart guide it is applied after. Just following that ATM.
Anyway, I tried it with the original wf and no white pixels with 315. From now, I'll Decon at the beginning (?). Maybe add this to the quick-start guide?
Thanks for your help.

Actually, no. In StarTools you're allowed to perform deconvolution any time while Tracking is engaged. It's just that you were really pushing the Tracking engine with doing so many things to the data before actually applying deconvolution (it should've been able to cope with that).

That said, probably the best moment to apply deconvolution is after your final global stretch (e.g. after Wipe and then AutoDev/Develop). This is because decon can dig out detail that other modules can pick up on (Sharp, HDR). Also, the mask that Decon creates can be re-used in Sharp (which is convenient) so that it won't affect or further bloat any stars.

Does this make sense?

[alacant]

Hi Ivo
Yes, thanks, it does help. I think that as a coder, a particular workflow may seem obvious. It's less so for end users. Your tool is wonderful for beginners as it gets us started quickly. I think my fault is learning by my errors and asking stupid questions along the way. Maybe include e.g. the Decon explanation in the quickstart guide? It's certainly helped me get my wf a bit more organised. Still a steep mountain to climb to coders level of course but glad to know it's all in place if/when I ever get there. Thanks again.

[admin]

Hi Ivo
Yes, thanks, it does help. I think that as a coder, a particular workflow may seem obvious. It's less so for end users.

You're absolutely right, that's why I endeavoured to make workflow sequence less of a 'thing' than it otherwise is for people. If you're "in the know" and using something like PixInsight then you would know that deconvolution needs to be done to on linear data. You would know in order to visualise your operations on linear data, that you would do a temporary screen stretch first, etc.

In StarTools, all this stuff is abstracted away from the user by design. All user intentions are supposed to (unless you have a glitch like the one you pointed out ) be made sense of by the software, which puts things in the right order by going back and forth in time.

The other perk you then get is that, because StarTools by design needs to keep tabs on signal evolution, we can use that information to track noise propagation as the image was processed. This in turn allows for extremely targeted noise reduction. It's a whole extra dimension of data that can be used as an input to the noise reduction routines. This defacto gives better results than a noise reduction routine - no matter how clever - that does not have the same extra information available to it.

Your tool is wonderful for beginners as it gets us started quickly.

This is great to hear and exactly my intention.
The funny (and wonderful) thing is that StarTools attracts a lot of beginners, after which this emboldens them to try other processing software for a while (usually PixInsight - which I can heartily recommend, if even just for its excellent pre-processing), after which they tend to return to StarTools for their post-processing with a renewed appreciation for its signal-preservation and Tracking capabilities.

I think my fault is learning by my errors and asking stupid questions along the way.

If anything that's a virtue! I firmly believe the only stupid question is the one that never got asked. Please never feel you should excuse yourself for wanting to learn.