What's going on here?

[groberts]

Not wanting to miss some clear skies last night I shot 3-hours @ 300sec LRGB subs of the M95/96/106 area of Leo - it's a bit of a stretch with my WO GT81 but really just experimenting until some decent size DSOs return (or I get a bigger scope!).  

Seeing was so so, guiding likewise, with all individual subs OK in all wavelengths (except for a few aircraft trails which were rejected - I live between Gatwick & Heathrow!)  LRG subs all stacked fine in DSS but there's something wrong with the Blue subs when stacked, which seems to be confined to the top left area (see attached image) - they look OK individually. 

Any ideas what's going on and what I can do?

Graham

 

[carastro]

That's a bit odd, especially as it in not on the right hand side.

You say you have checked the individual subs and they are OK.   I am inclined to feel it is a software stacking error, but no idea what or why. 

I'd be tempted to stack a few at a time, and see if it only happens with one stack, and the stacks that work, stack the stacks.

No real idea what is going on though. 

Carole 

[groberts]

Thanks Carole, it is indeed perplexing but think I've seen it before and am also inclined to think it's stacking but why only Blue and why now, when it's been OK with other similar images recently? 

[almcl]

Pretty sure I had something similar happen a little while ago.

Not absolutely certain but I *think* changing the star detection threshold in DSS sorted it. If it wasn't that, then deselecting  or selecting individual stars for the master frame was the solution.

[wimvb]

I believe that this happens when DSS detects stars in only one side of the image. Try adjusting the star detection threshold so that more stars are detected and used.

[groberts]

1 hour ago, wimvb said:

I believe that this happens when DSS detects stars in only one side of the image. Try adjusting the star detection threshold so that more stars are detected and used.

I have tried changing the detection to 5% and 25%, neither of which made any difference.  

What is strange is that (a) it is only the blue stacking and (b) the blue subs look OK individually.

I've been using the default 10% star detection for years without any problem, so not really sure what's happening here?

Graham   

[groberts]

First thanks for all the comments.

After posting on the DSS forum group too, it turns out that the Stacking Parameter should be set as Bilinear not Automatic as I had it and bingo, it works!  In fact, as a result the whole stacked image is much better.

Regards, Graham

 

 

[carastro]

Glad you solved it, but still don't understand why it should work on some subs and not others.

Carole 

[groberts]

18 hours ago, carastro said:

Glad you solved it, but still don't understand why it should work on some subs and not others.

Carole 

I've now gone back and re-stacked all the other wavelengths using Bilinear and on close inspection can now see that the new re-stacked LRG wavelengths are also better than before - stars are tighter and better saturation etc. See examples - Original automatic R-stack top image, Bilinear R-stack lower image .

In conclusion therefore, it's clear that using Bilinear is, at least in this case, the better way to go.  Perhaps because blue are darker subs the star detection in general is more difficult  for the software to detect and using automatic stacking the problem becomes much more pronounced?

Anyhow, in a funny way it's been an interesting problem and long term, a useful outcome for future processing.

Graham