#1  
Old 24-05-2013, 06:38 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Using HDR to avoid saturated stars

I often take a bunch of short subs along with my "normal" ones and then use HDRComposition in PixInsight to combine the integrated "short" and "long" images. This is an attempt to avoid those stars we've all seen which have a thin ring of colour surrounding a bleached white core where the star has saturated the CCD sensor.

When processing my recent image of Omega Centauri (http://www.iceinspace.com.au/forum/s...d.php?t=107621) I used this HDR technique and also did some 3D plots showing the star profiles with and without HDR. On the off chance that anybody is interested, here they are...

The first and third plots are two sample areas from an integration of 5 minute luminance exposures. The second and fourth plots are the same areas from a HDR combination of two integrations, the first of 0.5 second subs and the second of 5 minute subs. It's pretty easy to tell that the star profiles in the HDR version look much better!

Cheers,
Rick.
Attached Thumbnails
Click for full-size image (integration_L_Preview01_3dplot.jpg)
173.1 KB70 views
Click for full-size image (HDR_L_Preview01_3dplot.jpg)
180.3 KB64 views
Click for full-size image (integration_L_Preview02_3dplot.jpg)
192.9 KB54 views
Click for full-size image (HDR_L_Preview02_3dplot.jpg)
195.6 KB58 views
Reply With Quote
  #2  
Old 24-05-2013, 07:15 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Fascinating Rick. It looks like the process overcomes star "fattening" due to increasing CCD non-linearity at high signal levels, as well as outright clipping when the ABG cuts in. Is this a reasonable observation? Regards ray
Reply With Quote
  #3  
Old 24-05-2013, 07:30 PM
bojan's Avatar
bojan
amateur

bojan is offline
 
Join Date: Jul 2006
Location: Mt Waverley, VIC
Posts: 6,932
I am using this (or equivalent) technique all the time - when adjusting curves in DPP.
They are always steep at low intensities, while at the top they are more flattened but never completely flat (so, they are logarithmic-like).
Reply With Quote
  #4  
Old 24-05-2013, 08:03 PM
RobF's Avatar
RobF (Rob)
Mostly harmless...

RobF is offline
 
Join Date: Jul 2008
Location: Brisbane, Australia
Posts: 5,716
Good post Rick. I've never gotten around to seriously trying this except ad hoc for objects with really bright stars in the field (e.g. M42). Make a lot of sense and good of you to show some detail in the process.
Reply With Quote
  #5  
Old 24-05-2013, 08:30 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Quote:
Originally Posted by Shiraz View Post
Fascinating Rick. It looks like the process overcomes star "fattening" due to increasing CCD non-linearity at high signal levels, as well as outright clipping when the ABG cuts in. Is this a reasonable observation? Regards ray
Ray, I don't think it helps a lot with the "fattening" but it does replace the truncated top of the star profiles. Probably a second best strategy compared to your use of short subs with a very low noise camera.

Quote:
Originally Posted by bojan View Post
I am using this (or equivalent) technique all the time - when adjusting curves in DPP.
They are always steep at low intensities, while at the top they are more flattened but never completely flat (so, they are logarithmic-like).
Interesting, Bojan. I've never played with DPP. So you're doing something similar with a non-linear stretch?

Quote:
Originally Posted by RobF View Post
Good post Rick. I've never gotten around to seriously trying this except ad hoc for objects with really bright stars in the field (e.g. M42). Make a lot of sense and good of you to show some detail in the process.
Thanks, Rob. 3dplot makes it easy to experiment with these things and visualize the results. It's so hard to come by data when you don't have a permanent obs and live in the city so I feel obliged to try to make the best of what I can get!

Cheers,
Rick.
Reply With Quote
  #6  
Old 24-05-2013, 08:40 PM
bojan's Avatar
bojan
amateur

bojan is offline
 
Join Date: Jul 2006
Location: Mt Waverley, VIC
Posts: 6,932
Quote:
Originally Posted by RickS View Post
Interesting, Bojan. I've never played with DPP. So you're doing something similar with a non-linear stretch?
Yes.. I developed this method to deal with LP, but it takes care of saturation as well see here:
http://www.iceinspace.com.au/forum/s...4&postcount=37


The 3D diagram is a nice way to visualise the result, though..
Reply With Quote
  #7  
Old 24-05-2013, 08:51 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Thanks, Bojan.
Reply With Quote
  #8  
Old 24-05-2013, 09:19 PM
DavidTrap's Avatar
DavidTrap (David)
Really just a beginner

DavidTrap is offline
 
Join Date: Jun 2009
Location: Brisbane
Posts: 3,032
Looks convincing.

Do you believe you need the ultra short subs? Why not 5 or 10 sec? Very short subs can cause uneven illumination from shutter errors?

DT
Reply With Quote
  #9  
Old 24-05-2013, 10:05 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Quote:
Originally Posted by DavidTrap View Post
Looks convincing.

Do you believe you need the ultra short subs? Why not 5 or 10 sec? Very short subs can cause uneven illumination from shutter errors?

DT
David,

It doesn't take very long to push a bright star to saturation. Often a few seconds is more than enough. I usually watch the star profiles on Maxim to pick a good exposure time.

Uneven illumination hasn't been a problem for me. The HDR process only takes the very brightest bits from the short exposure stack - just the star cores typically. The HDRComposition process has an option (on by default) that shows you the mask that was used. The short stack often looks pretty noisy and horrible but as long as the star cores are OK it doesn't matter.

Cheers,
Rick.
Reply With Quote
  #10  
Old 24-05-2013, 10:35 PM
DavidTrap's Avatar
DavidTrap (David)
Really just a beginner

DavidTrap is offline
 
Join Date: Jun 2009
Location: Brisbane
Posts: 3,032
Ta for that!
DT
Reply With Quote
  #11  
Old 25-05-2013, 08:02 AM
naskies's Avatar
naskies (Dave)
Registered User

naskies is offline
 
Join Date: Jul 2011
Location: Brisbane
Posts: 1,865
Thanks Rick! I find my STF-8300 saturates stars very, very quickly so this will be a very useful trick to learn. I've been meaning to get started with PI so this is a great reason to do so
Reply With Quote
  #12  
Old 25-05-2013, 08:42 AM
RobF's Avatar
RobF (Rob)
Mostly harmless...

RobF is offline
 
Join Date: Jul 2008
Location: Brisbane, Australia
Posts: 5,716
I tend to have the 3D plot of the brightest star coming up as info in the Maxim camera window, and as Dave says, doesn't take long for an 8300 chip to start turn stars into "mesas"!

Guess I'm usually more focused (no pun intended) on getting round stars than perfect profiles, but it shouldn't actually take that much more work as you say, particularly if you're running an automated profile for data collection. Was reminded last new moon how important it is to the run set to go to maximise every minute of dark sky when you're a city astronomer.

Another PI option is masked stretch. Can be very gentle on data, but prone to a little ringing around very bright stars on nebs sometimes.
Of course you can go crazy with starmasks and morphological transforms to "smooth" the sharp edges on your star profiles too if you're going to be artificial about it.
Reply With Quote
  #13  
Old 25-05-2013, 12:42 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Definitely worth a try, chaps. Unless you have a camera with really slow downloads (I'm looking at you, STL11K!) it doesn't take very long to collect the extra data.
Reply With Quote
  #14  
Old 25-05-2013, 06:18 PM
Garbz (Chris)
Registered User

Garbz is offline
 
Join Date: May 2012
Location: Brisbane
Posts: 644
As a matter of interest (sorry for threadjacking) has anyone used the built in algorithms in Deep Sky Stacker to do this? DSS gives the option of using Entropy Weighted Average for specifically stacking images with different exposures to avoid burning out bright centres.

Any comparisons to this method here?
Reply With Quote
  #15  
Old 25-05-2013, 07:03 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Quote:
Originally Posted by Garbz View Post
As a matter of interest (sorry for threadjacking) has anyone used the built in algorithms in Deep Sky Stacker to do this? DSS gives the option of using Entropy Weighted Average for specifically stacking images with different exposures to avoid burning out bright centres.

Any comparisons to this method here?
Sounds similar. Hopefully someone who knows more about DSS than I do can comment.
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 03:17 AM.

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