livestacking histogram jumps around more than before?

Somewhere to ask questions about the best way to use SharpCap
Forum rules


If you have a problem or question, please check the FAQ to see if it already has an answer : https://www.sharpcap.co.uk/sharpcap-faqs
dmccallie
Posts: 19
Joined: Thu May 09, 2019 6:46 pm

Re: livestacking histogram jumps around more than before?

#21

Post by dmccallie »

Robin,
I sent the link via a private message to admin.
Let me know if you didn't get it!

And thanks again for all the attention, and for your amazing software.

David McCallie
Powell Observatory
Astronomical Society of Kansas City
User avatar
admin
Site Admin
Posts: 13179
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: livestacking histogram jumps around more than before?

#22

Post by admin »

Hi David,

ok, those images that you sent were incredibly useful - I think I understand the problem now.

What is happening is that the brightest pixels in those images aren't from stars or any 'real' thing in the image - they are hot pixels. Live stacking thinks about the brightness of everything in the image as a fraction of the brightness of the brightest part of the stack. If the brightest part of the stack is a star, then that ratio stays the same as more frames are added (to within a tiny fraction) and everything works. If the brightest part is a hot pixel *and* you have any drift, then with each added frame the hot pixel becomes less important - because it is being smeared out over various pixels in the stack due to the alignment. That means that the brightness of the nebula/galaxy/whatever relative to the brightest part is increasing gradually, leading to histogram shift.

If you let enough frames pass then the hot pixels will be smeared out enough that one or more stars will become the brightest thing - that will stop the shifting. Alternatively, as you mentioned in your PM, increase the brightness of the input frames (even just turning up the gain will help).

At my end, I will look at a more robust way of scaling the brightness - probably taking something like the 99.99th centile or something as the 'white point' for the stack rather than the brightest pixel.

cheers,

Robin
dmccallie
Posts: 19
Joined: Thu May 09, 2019 6:46 pm

Re: livestacking histogram jumps around more than before?

#23

Post by dmccallie »

Hi Robin,

Thanks for the quick turn-around! I’m glad that the “jumping histogram” was not entirely a figment of our imagination(s).

Would capturing dark frames each night help reduce the histogram-scaling impact of the hot pixels? Would darks be likely to filter out the hot pixels? We have skipped using dark frames when livestacking for public display because the benefit is so limited.

I might have thought that using the alternate stacking mechanism (sigma rejection) would have helped, but as I recall during our testing, that made no difference.

In the mean time, we’ll scale up the intensity of the image and see how much that helps.

Thanks again.

—david
User avatar
admin
Site Admin
Posts: 13179
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: livestacking histogram jumps around more than before?

#24

Post by admin »

Hi David,

yes, darks should help, since when you subtract dark frames SharpCap will automatically detect hot pixels in the dark frame and avoid using data from those pixels - that should largely solve the problem. Sigma clipping also scales to the brightest pixel in the stack, so has the same problem. I think the right fix will be a code update from me - still thinking about the right way to implement it though, so it will not make this week's update.

cheers,

Robin
Post Reply