StarTools

Hi all,

A bit of a simple picture. The subs are from last year sometime in November. I had only 23 of them, 3 minutes each. It is taken with a Askar Fra 500 and an unmodded Canon DSLR 800D. No filters.
I kinda of like the outcome but there are some halo like artifacts around the brightest stars. Actually it does not bother me that much but has anyone an idea to reduce this? It is sometimes a recurring problem, i think it comes merely from deconvulation. On this image i forecedly used it to get pinpoint stars.

Glad to hear your opinions

Very cool Freddy! That's a deep and busy star field.

Maybe a slight tint of magenta to the background?

Yes I agree, some stars seem to have a shadow artifact, maybe some form of ringing. And perhaps the cores seem a bit flat. Overall I really like the look though.

Anything unusual in the workflow? What ST version did you use? I've had similar issues (and worse), especially since the deringing change. A couple other recent threads mentioned stuff like this too. For SVD, did you turn on intra-iteration + centroid tracking? That seems to be the acknowledged first step towards reducing such artifacting. Then deringing, though for me I can only take it so far, but can follow up with a no-shrink Shrink to boost the deringing.

There's also the possibility of a dud or two in the star samples, as Ivo mentioned.

Mike in Rancho wrote: ↑Sun May 07, 2023 8:38 pm What ST version did you use? I've had similar issues (and worse), especially since the deringing change. A couple other recent threads mentioned stuff like this too. For SVD, did you turn on intra-iteration + centroid tracking? That seems to be the acknowledged first step towards reducing such artifacting. Then deringing, though for me I can only take it so far, but can follow up with a no-shrink Shrink to boost the deringing.

There's also the possibility of a dud or two in the star samples, as Ivo mentioned.

I use the v1.9 so guess the beta version. I use only this lately, i like it more since some things imho seem to work beter then in the previous version.

What do you mean by : SVD, did you turn on intra-iteration etc...i am missing here something....???

I tried after looking in the 1.8 manual to cut back the halo via the shrink module, did not work though...

Mike

ok ok, found out what you mean, but no i did not try this before ...

Investigating and playing with it right now. Will it be somuch better, well i have yet to find out.

Hi Freddy,

fmeireso wrote: ↑Mon May 08, 2023 11:44 am What do you mean by : SVD, did you turn on intra-iteration etc...i am missing here something....???

please have a look at this thread
viewtopic.php?f=4&t=2821

and at Ivo's post here
viewtopic.php?p=14344#p14344

Finding suitable samples seems to be crucial and in some cases it may be better to choose a synthetic profile.

Best regards, Dietmar.

Thanks Dietmar,

Interesting reading stuff.
I encountered yesterday also the strange artifacts in the stars (black eyed artifact) which again could be solved by centroid tracking setting. And on some occasion not. Indeed in many cases just syntethic profile works as good. I think in testing yesterday only one stack look abit better with SVD.

It is however an interesting feature, i did not really know about it i confess, so at least again i have learned something

CS
Freddy

fmeireso wrote: ↑Tue May 09, 2023 8:31 am Thanks Dietmar,

Interesting reading stuff.
I encountered yesterday also the strange artifacts in the stars (black eyed artifact) which again could be solved by centroid tracking setting. And on some occasion not. Indeed in many cases just syntethic profile works as good. I think in testing yesterday only one stack look abit better with SVD.

It is however an interesting feature, i did not really know about it i confess, so at least again i have learned something

CS
Freddy

I should probably create a warning popup message for when the user hits "keep", was using samples, but has not switched on the centroid tracking mode. It takes a lot longer to calculate, but the result is definitely worth it.