BlurX Clipping Star Cores!?!?

S

Serpentarii

Member
Hello,

I just was shocked when i realised that BlurXterminator Clips the star cores. Yes I understand that deconvolution "pushing" together the pixle values and therefore they will add up to higher values in the center, but shouldn't then the image be not just converted to a higher bit range to avoid clipping? I am loosing all the color information in those stars. Or am I missing something?

It is not clipping btw when i use the "luminanc only" option but then I donte get the cromatic correction :-(

Curios if anyone else finds this devestating...

Cheers, Robert
 
For my BXT it works fine, no problem, have you contacted RC Astro directly? Russell is the most appropriate to resolve this issue.
 
So, I'm trying the process I linked to above for the first time myself, and while I haven't done anything with the resulting stars alone yet, I had a very odd result from GradientCorrection when I applied it to the whole image. It's not clear in the pdf when background extraction should occur in the 'new' workflow, but I applied it after the BX process described and I'm curious what's going on if anyone has some thoughts. FWIW, I got the same result with both default settings, and the settings I used successfully after the 'old' workflow BX. *** POST EDITED BY THE FORUM ADMINISTRATOR - FORUM RULES VIOLATION

Cheers,
Scott

Capture.JPG
 
Last edited by a moderator:
Have you looked at the protection mask generated by structure protection? It is often an obvious generator of this sort of result. I would never run with structure protection unless you are sure you need it. If you do run with structure protection always generate and examine the mask; you may have to iterate several adjustments of the protection parameters until you are happy that the mask is only protecting the regions you want protected.
 
Structure protection does not generate this kind of artifact under normal working conditions. Can you please upload the image to a cloud service so we can investigate what happens in this case?
 
Sure. It'll be a little bit before I can get to it, but in the meantime I wanted to add that after posting I tried GC on the image prior to the recip_e/BX iterations (see pdf linked to above) and it worked as expected. So, it seems to be an issue related to those tools and application.

Also, sorry for the rules violation. It's a little confusing that software incorporated into Pixinsight's Processes, and therefore intended, I assume, to be used as part of a Pixinsight workflow, is off limits

Cheers,
Scott
 
Hi Scott,

Thank you for uploading the image. The PixelMath operations you are referring to can alter the image in ways that make it problematic for GradientCorrection's algorithms. GC must always be applied to the linear image before any alteration.

As for the other issues, they should be discussed privately with the administrators.
 
By the developers of PI.

Happy to discuss it further by pm or in another thread, but now we’re really straying from the original (albeit unspeakable) subject…..

Cheers,
Scott
 
You must log in or register to reply here.
Back
Top