Go Back   IceInSpace > Equipment > Software and Computers

Reply
 
Thread Tools Rate Thread
  #1  
Old 08-02-2019, 01:21 PM
Stonius's Avatar
Stonius (Markus)
Registered User

Stonius is offline
 
Join Date: Mar 2015
Location: Melbourne
Posts: 1,495
Technical question re bias / Vs Superbias

Hi all,

I wrote a really long post before, but deleted it.

Basically, what I'm trying to figure out is if you take a dark and bias integration, then run them through calibration, should the statistics on the resultant frame match those same values mathematically subtracted in a spreadsheet?

Being integrations, we are examining the non-random elements of the noise.

In experimenting with noise reduction I'm not sure if I should use a bias integration to play with or a superbias.

I'm finding frames calibrated with the superbias are *higher in both mean levels and StdDev, which I'm guessing is a result of the Superbias being consistently lower in mean and StdDev levels. But that seems counter-intuitive - I thought the point of a superbias was that it got rid of more noise?

Markus

PS, I'm aware that biases with the 1600 are not real biases. I achieved them by integrating 2.5s exposures, doing a linear regression and scaling the frame to the y crossing in pixinsight.
Reply With Quote
  #2  
Old 08-02-2019, 04:19 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Bias/dark calibration doesn't remove shot noise. It removes unwanted signal, i.e. bias pedestal and thermal noise. In fact, it adds a little shot noise that remains in the master calibration files.

The superbias process takes a poor quality master bias and "improves" it by doing multiscale noise reduction. The superbias will add less shot noise during calibration than the poor master bias, but it won't be quite as effective at removing unwanted signal. I think this remaining unwanted signal is probably what you're measuring.

I'm not a big fan of superbias. I'd much rather build a good quality master bias based on a large number of bias frames.

Cheers,
Rick.
Reply With Quote
  #3  
Old 08-02-2019, 04:40 PM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
I agree that a good Master Bias is sufficient. The 1600s are very clean down that end in my experience. I prefer eyeballing the histograms to spreadsheets. And I have a library of bias Masters I have acquired over the years, some at different cooling setpoints that I favour, and there is almost no difference that I can detect. Even stretched to hell. Btw why shoot bias at 2.5sec? SGP shoots them at 0 seconds, basically dumps the idle sensor state. Maybe I have been doing it wrong.
Reply With Quote
  #4  
Old 08-02-2019, 09:50 PM
Stonius's Avatar
Stonius (Markus)
Registered User

Stonius is offline
 
Join Date: Mar 2015
Location: Melbourne
Posts: 1,495
Quote:
Originally Posted by RickS View Post
I'm not a big fan of superbias. I'd much rather build a good quality master bias based on a large number of bias frames.

Cheers,
Rick.
Given what you've said, I'm starting to think the same. My bias is baes on 250 frames. Is that what you mean by 'large number of frames'? or do you go even more?

Give the ease with which biases are captured, I can't imagine why anyone would bother making a substandard bias?

Cheers & thanks

Markus
Reply With Quote
  #5  
Old 08-02-2019, 09:52 PM
Ukastronomer (Jeremy)
Feel free to edit my imag

Ukastronomer is offline
 
Join Date: Jan 2019
Location: Llandysul, WALES, UK
Posts: 1,381
I do wish you would all speak English, or that I understood what you were talking about, goes over my head
Reply With Quote
  #6  
Old 08-02-2019, 10:01 PM
Stonius's Avatar
Stonius (Markus)
Registered User

Stonius is offline
 
Join Date: Mar 2015
Location: Melbourne
Posts: 1,495
Quote:
Originally Posted by glend View Post
Btw why shoot bias at 2.5sec? SGP shoots them at 0 seconds, basically dumps the idle sensor state. Maybe I have been doing it wrong.
It's to do with the way the 1600 operates at exposures below about 2.3s. ZWO have mentioned this in their forums (see attached graph).

Post #25 on this thread sums it up.

The only way to get a reasonable bias at the equivalent to zero exposure is to shoot slightly longer frames at >2.3s, figure out the slope of the line and extrapolate what the value would be if those exposures actually were 0s, then scale the difference in PI.

Best

Markus
Attached Thumbnails
Click for full-size image (ADU of Bias Frames V Exposure Length.png)
12.5 KB42 views
Reply With Quote
  #7  
Old 08-02-2019, 10:10 PM
Stonius's Avatar
Stonius (Markus)
Registered User

Stonius is offline
 
Join Date: Mar 2015
Location: Melbourne
Posts: 1,495
Quote:
Originally Posted by Ukastronomer View Post
I do wish you would all speak English, or that I understood what you were talking about, goes over my head
My friend, I know exactly how you feel. I felt exactly the same way until I fell down the wormhole that is astrophotography 8 months ago or so.

It's really just a question regarding how best to reduce noise - a lot of astrophotography is about reducing noise because the signals we want to capture are so darned faint.

:-)

Markus
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 07:04 PM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Testar
Advertisement
Bintel
Advertisement