#1  
Old 04-08-2012, 08:59 AM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Pixinsight - dark stars

I've just commenced on the long journey of learning how to drive PI.

I have a range of questions - which I may bore everyone with at some stages.

For starters, I have an odd experience.

Raw images when displayed in PI have a dark spot near the centre of all bright stars. In especially bright stars, this covers up to a third of the star.

I had thought it was a result of the minimum/maximum clipping process when calibrating images but its there in uncalibrated images too.

This is very odd - I haven't seen this with Maxim.

Images are taken with a QSI 583 using Maxim.

Has anyone else encountered this before?

Pete
Reply With Quote
  #2  
Old 04-08-2012, 09:01 AM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Dark stars

Here is a screen shot illustrating this effect

Weird -

Pete
Attached Thumbnails
Click for full-size image (Screen shot 2012-08-04 at 10.00.03 AM.jpg)
70.1 KB72 views
Reply With Quote
  #3  
Old 04-08-2012, 09:50 AM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
I'm using Maxim to capture from a couple of cameras (one KAF8300 based like your QSI) and the processing in PI I have never seen that.

Cheers,
Rick.
Reply With Quote
  #4  
Old 04-08-2012, 11:56 AM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Quote:
Originally Posted by RickS View Post
I'm using Maxim to capture from a couple of cameras (one KAF8300 based like your QSI) and the processing in PI I have never seen that.
Rick

can I ask you to check something for me please? If you open an image you have taken in Maxim that displays normally in PI and open the FITS header, can you tell me what you have recorded for the BZERO field.

Recently, I changed my default setting in Maxim to Save Date in Unsigned Format. I did this as the PI forums suggested that saving the data in signed format sets the range of ADU from 32768 to 65,000. This in turn causes dramas for flat fields.

Looking at my images, the unsigned setting results in this odd appearance while the signed data looks fine.

Cheers

Pete
Reply With Quote
  #5  
Old 04-08-2012, 12:23 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Pete,

My FITS files have BZERO=32768.0 and BSCALE=1.0 which means the data values are interpreted as unsigned 16-bit integers. I do not have the "Save Data in Unsigned Format" box ticked in Maxim. It seems do something other than what you'd expect...

You should be able to edit the FITS headers in your files to correct them. I think the data should be fine.

Cheers,
Rick.
Reply With Quote
  #6  
Old 04-08-2012, 12:42 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Quote:
Originally Posted by RickS View Post
Pete,

My FITS files have BZERO=32768.0 and BSCALE=1.0 which means the data values are interpreted as unsigned 16-bit integers. I do not have the "Save Data in Unsigned Format" box ticked in Maxim. It seems do something other than what you'd expect...


Cheers,
Rick.
Mmmmmmm - thats very helpful Rick. Thanks

Do you calibrate in PI?

I've been using the Batch Preprocessing script and can't get rid of a heavy gradient in blue. I had hoped that the signed/unsigned switch would do the trick.

More to ponder

Pete
Reply With Quote
  #7  
Old 04-08-2012, 12:56 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 pvelez View Post
Do you calibrate in PI?

I've been using the Batch Preprocessing script and can't get rid of a heavy gradient in blue. I had hoped that the signed/unsigned switch would do the trick.
When I first started using PI, I used to calibrate in Maxim. After that, I switched to calibrating in PI using ImageCalibration and ImageIntegration. These days I calibrate in PI using BPP.

I've not had problems with gradients apart from the usual ones from imaging under light polluted skies...

Cheers,
Rick.
Reply With Quote
  #8  
Old 04-08-2012, 01:37 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Thanks Rick

I'll have another look at this

Pete
Reply With Quote
  #9  
Old 04-08-2012, 02:38 PM
whzzz28's Avatar
whzzz28 (Nathan)
Registered User

whzzz28 is offline
 
Join Date: Dec 2011
Location: Brisbane
Posts: 348
Your using PI to stack correct (post processing)?

If so, when you open your stacked LRGB files three windows will open. The bottom window (first to open) is the one with all your data in it, the other two are the high and low pass rejected data.
I don't know what what constitutes as cutoff data for either low or high passes, but it should be in the post processing area somewhere, i believe it is the Sigma Low and Sigma High values under Image Integration (under Lights).
Mine are set to 4.00 and 3.00 respectively.

Maybe check them and make sure the values arn't funky.
I have seen something similar on my shots using PI to stack. I then ran the data through DSS to stack and it was fine
Reply With Quote
  #10  
Old 04-08-2012, 02:52 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 whzzz28 View Post
Your using PI to stack correct (post processing)?
Pete is having the problem with unstacked raw subs. Looks like Maxim with the "Save Data in Unsigned Format" option enabled writes the data in unsigned 16-bit but doesn't set BZERO in the FITS header as other applications would expect (the Maxim help file vaguely alludes to this).

PI thinks the data is 16-bit signed and so the values in the brighter areas of the image (with the most significant bit set) look like negative numbers and hence stars have dark cores.

Cheers,
Rick.
Reply With Quote
  #11  
Old 04-08-2012, 02:55 PM
whzzz28's Avatar
whzzz28 (Nathan)
Registered User

whzzz28 is offline
 
Join Date: Dec 2011
Location: Brisbane
Posts: 348
Quote:
Originally Posted by RickS View Post
Pete is having the problem with unstacked raw subs.
Ah sorry, my bad disregard the above.
Reply With Quote
  #12  
Old 04-08-2012, 03:11 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Thanks guys - and especially thanks Rick - that makes plenty of sense now.

So I'm back to unchecking the "Save Data in Unsigned Format" radio box and then see how things pan out.

Still weird if you ask me. Maxim doesn't mind - the images look fine in that program.

Pete
Reply With Quote
  #13  
Old 04-08-2012, 03:44 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
I've noticed all sorts of weird stuff happen when you transfer FITS files between different image processing apps. The PI guys argue, usually convincingly, that they are right and everybody else is wrong The basic problem is that the FITS file standards aren't complete and there is freedom to interpret them in different ways.

One example is the coordinate origin of images. I think that Maxim and CCDStack assume an upper left corner origin. PI defaults to lower left corner, but you can change this in the FITS Format preferences.

A bigger problem is that the FITS standard doesn't specify the black and white points for floating point values. PI assumes a black point of 0 and a white point of 1 but other apps use different ranges. Once again, PI lets you change this if you need to.

Cheers,
Rick.
Reply With Quote
  #14  
Old 04-08-2012, 04:50 PM
cfranks (Charles)
Registered User

cfranks is offline
 
Join Date: Feb 2008
Location: Tungkillo, South Australia
Posts: 599
There is a situation that I haven't been able to determine why, where MaximDL will automatically save a Screen Stretched image. Opening it in PI will give a completely overstretched image right down to the black star centres exactly like yours. If you take the image back to Maxim, open the screen stretch box, adjust the pointers to zero and maximum values, ie. unstretch it, and save it, PI will treat it as normal.

Charles
Reply With Quote
  #15  
Old 05-08-2012, 01:54 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Quote:
Originally Posted by cfranks View Post
There is a situation that I haven't been able to determine why, where MaximDL will automatically save a Screen Stretched image. Opening it in PI will give a completely overstretched image right down to the black star centres exactly like yours. If you take the image back to Maxim, open the screen stretch box, adjust the pointers to zero and maximum values, ie. unstretch it, and save it, PI will treat it as normal.

Charles
Thanks for the tip Charles - I'll give it a go.

Pete
Reply With Quote
  #16  
Old 05-08-2012, 01:58 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
I did some testing last night. I unchecked the "Save Data in Unsigned Format" and took 1 1/2 hours of the Jewel Box

Unsurprisingly - no odd stars. Here is a 180 second sub in G after calibration. The effects of the full moon are pretty obvious.

At least I know how to avoid this now - thanks to everyone for their suggestions.

As to my flats - that is another story - or at least another post

Pete
Attached Thumbnails
Click for full-size image (Screen shot 2012-08-05 at 2.56.29 PM.jpg)
95.3 KB41 views
Reply With Quote
  #17  
Old 05-08-2012, 02:03 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Good to see you've got one problem knocked on the head, Pete!
Reply With Quote
  #18  
Old 05-08-2012, 02:22 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Quote:
Originally Posted by RickS View Post
Good to see you've got one problem knocked on the head, Pete!
Thanks Rick - that's one problem out of the way.

Now to the myriad of others.

I do feel I'm starting to get the hang of PI - its a hell of a ride though

Pete
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 12:09 PM.

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