#21  
Old 07-08-2010, 02:00 PM
Terry B's Avatar
Terry B
Country living & viewing

Terry B is offline
 
Join Date: Mar 2006
Location: Armidale
Posts: 2,789
Pete
Looking at your unclabrated frame there is some vignetting in the left upper corner. It looks like frosting on the corner of the frame.
As for flats, they are tricky to get correct. I use CCDSoft rather than maxim but the process is the same.
I have a white board on the wall of my observatory and aim the scope at it. It is lit either by a light aiming at the opposite wall or just on an overcast day with the even illumination that that creates in my observatory.
If you use a T shirt you have to be carefull that the front of the scope is evenly lit and nothing partly shaddows the T shirt like a dew shield.
Reply With Quote
  #22  
Old 07-08-2010, 03:26 PM
pvelez's Avatar
pvelez (Pete)
Registered User

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

Yes, the vignetting was a function of the dew shield I used to take those shots. I've since upgraded to a light box and a RC8 which makes for a better match than the old LXD-75.

I've now experimented with Sky Flats Assistant which seems quite intuitive once you get it running. Now all I need to do is capture some lights and new flats and apply them. I tried to apply today's flats to some shots I took a month ago - urgh! not good. The old shots had already been calibrated with the dodgy flats so I ended up with a very nasty colour gradient. The right half of the screen has a delightful green hue while the centre left has a wonderful magenta look, fading to green at the extremities. I won't share those with you all - it could put you off your dinner.

Looking at the new flats, it seems they are similar to those I took but with a shorter rather than longer exposure time with ADU around 22,000 across the board.

So I'll let uyou all know how I go - probably tomorrow night

Pete
Reply With Quote
  #23  
Old 09-08-2010, 05:06 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Ok - it must be me.

I took a few test shots of an attractive group of galaxies in - Pavo I think - last night. I then fired up the light box (thanks Exfso) and took a series of flats using the Sky Flat Assistant plug in with Maxim. Took darks at the correct temperature and a bias or 2 for good measure so the darks would be scaled and applied to the flats.

All good.

I then used the Calibration Wizard to point to the relevant folder and made sure all the cal files were there. They were.

To the combine routine - open Stack, add folder and select Auto-Calibrate. Measure, align and combine files into R, G, B and L groups. Then colour combine.

Am I missing something?

I have an image with a very green hue in the background to the right of the target and about a third to the left. Down the middle there is a strong magenta tone.

So I conclude its my flats. I take a single image - through a green filter as its the most sensitive - looks like some vignetting to me as the centre of the image is noticeably brighter than the edges. I then apply the flat from the same filter - similar look in terms of brightness - wham! Maxim has overdone it. I have effectively an inversion of the light with too much brightness at the outside and not enough in the centre.

Now my weighting of colours is around 1.23-R 1-G and 1.89-B, so when I do a combine, I have too much green for all the image except the centre. Where there is insufficient green, its a delightful magenta.

Quite arty but a poor astrophoto. My missus thinks I'm like Michelangelo painting the ceiling in the Sistene Chapel - "Michael, when are you going to finish??". The cold nights never produce anything worth looking at let alone printing.

So.....how do I reconcile this?

1. Is my green filter sensitive to light at different intensities so it passes more light when the light is brighter than it should - so my flat is not really flat?

2. Or is there a problem with my calibration routine? I am no Maxim DL expert - just a novice.

3. Or is it not vignetting but rather something else rather odd? I had a SN f4 and had a similar issue - you can see that from the shots in the original post. I now use an RC8 with the SBIG 8300. Is this some other optical aberration eg do I need a flattener?

4. Or am I just lousy at flats? I reviewed each of the flats and found that they stayed well within the ADU range - coming in at around 20k - 22k ADU. The EL flats as well as the T-shirt flats were pretty good too.

5. Or maybe its just me. Eminently possible but in this case I suspect not.

If anyone can put me out of my misery, I'd be grateful.

My suspicion is that number 2 is the culprit. Perhaps I am double or triple calibrating so the images are being progressively lightened at the margins with each iteration. Luckily I have kept a back up of uncalibrated images from last night so I can try and manually calibrate them first and then stack, align and combine.

Incidentally, I read somewhere that the FITS header for images taken with Maxim DL include a field identifying whether calibration has been completed. For the life of me I can't find it. Am I making this up?

If I can get to the computer tonight, I'll post some images so all the above makes sense.

Cheers

Pete
Reply With Quote
  #24  
Old 09-08-2010, 05:57 PM
Bassnut's Avatar
Bassnut (Fred)
Narrowfield rules!

Bassnut is offline
 
Join Date: Nov 2006
Location: Torquay
Posts: 5,064
How many flats are you stacking for each filter, or are you takeing just one flat for each?.
Reply With Quote
  #25  
Old 09-08-2010, 05:58 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Stack of 10 Fred - median combine

Pete
Reply With Quote
  #26  
Old 09-08-2010, 06:36 PM
Bassnut's Avatar
Bassnut (Fred)
Narrowfield rules!

Bassnut is offline
 
Join Date: Nov 2006
Location: Torquay
Posts: 5,064
Well, its got me buggered. Can you post the flats, darks (stacked) and lights FITS on an FTP server for us to fiddle with?.
Reply With Quote
  #27  
Old 09-08-2010, 07:26 PM
pvelez's Avatar
pvelez (Pete)
Registered User

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

it would be my pleasure - though I don't have access to a FTP server. Is there a public-accessable site you know of I can use?

I'm testing a combine at the moment. Will post a sample shortly so you can see what's going on - or not as the case may be

Pete
Reply With Quote
  #28  
Old 09-08-2010, 07:51 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Maxim flats - the gory details

So here are a few shots to show you what I am wrestling with.

The first is an uncalibrated 5 minute image. It is a median combine of 10 images with a median combine. I have used the default auto-stretch in Maxim. Yes, I know there is heaps of dust about!

The second is a master flat taken through a green filter. Same deal with the stretch.

The third is after calibration.

Hope you can see my dilemma. Comments happily accepted.

Pete
Attached Thumbnails
Click for full-size image (Copy of Master_Flat Green 1_Green_1676x1266_Bin2x2_Temp-10C_ExpTime160ms.jpg)
196.4 KB32 views
Click for full-size image (Copy of NGC 6872-002G.jpg)
165.0 KB29 views
Click for full-size image (Copy of NGC 6872-002G_flat_applied.jpg)
157.8 KB32 views
Reply With Quote
  #29  
Old 09-08-2010, 07:55 PM
Bassnut's Avatar
Bassnut (Fred)
Narrowfield rules!

Bassnut is offline
 
Join Date: Nov 2006
Location: Torquay
Posts: 5,064
Here is the same problem as with the original image sequence, could you look at your post and confirm the order, as in what the 1st 2nd and 3rd images are.
Reply With Quote
  #30  
Old 09-08-2010, 08:06 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Sorry Fred

first is the master flat, second is the uncalibrated image, third is the "calibrated" image.

Pete
Reply With Quote
  #31  
Old 09-08-2010, 08:20 PM
Bassnut's Avatar
Bassnut (Fred)
Narrowfield rules!

Bassnut is offline
 
Join Date: Nov 2006
Location: Torquay
Posts: 5,064
OK, well, the flat looks reasonably normal, if not a bit mishaped, and the uncalibrated light looks like a fair match. The calibrated light looks sort of "over calibrated"". Ive seen a similar situation before, with a missmatch between light and flat filters. Are you absolutely sure the light and flat filters are the same?, check the filter allocation is correct in DL set up, or there isnt a mix up after capture. Check the FITs header filter data for each, DL auto calibrate relies on correct fits header data. Try manual flat selection for a light to confirm.
Reply With Quote
  #32  
Old 09-08-2010, 08:28 PM
Bassnut's Avatar
Bassnut (Fred)
Narrowfield rules!

Bassnut is offline
 
Join Date: Nov 2006
Location: Torquay
Posts: 5,064
Oh, and I hope you are saveing darks and flat stacks in DL with auto stretch DISABLED ?.
Reply With Quote
  #33  
Old 09-08-2010, 08:46 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Quote:
Originally Posted by Bassnut View Post
OK, well, the flat looks reasonably normal, if not a bit mishaped, and the uncalibrated light looks like a fair match. The calibrated light looks sort of "over calibrated"". Ive seen a similar situation before, with a missmatch between light and flat filters. Are you absolutely sure the light and flat filters are the same?, check the filter allocation is correct in DL set up, or there isnt a mix up after capture. Check the FITs header filter data for each, DL auto calibrate relies on correct fits header data. Try manual flat selection for a light to confirm.
Thanks Fred

agree that the light and flat look ok - its the calibration that is odd.

Using Sky Flat Assistant, I have been manually setting the filter before capturing the flat so it should match. The FITS header simply reads from Maxim DL so it should all correlate.

Pete
Reply With Quote
  #34  
Old 09-08-2010, 08:50 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Quote:
Originally Posted by Bassnut View Post
Oh, and I hope you are saveing darks and flat stacks in DL with auto stretch DISABLED ?.
Sounds an interesting one - how do I check this? I still have training wheels on with Maxim.

Looking at the FITS header it reads under CSSTRETCH "Range" / Initial display stretch mode.

CBLACK is 18574 and CWHITE is 23297

Not sure if that is good or bad

Pete
Reply With Quote
  #35  
Old 09-08-2010, 08:52 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
LRGB - not good

Here is a LRGB - curves and levels with a bit of a saturation boost to show the colour gradient.

(Only 1 image so the order is right!)

Pete
Attached Thumbnails
Click for full-size image (NGC-6872_calibrated.jpg)
117.8 KB34 views
Reply With Quote
  #36  
Old 10-08-2010, 05:14 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
I'm trying a new tack tonight - I'll take a few shots with another imaging program and see if they calibrate better. I believe I have CCDSoft lurking on my computer at home.

Any suggestions for Maxim are welcome. I am keen to use as few software packages as possible and until this arose Maxim looked the goods.

Pete
Reply With Quote
  #37  
Old 10-08-2010, 07:57 PM
jase (Jason)
Registered User

jase is offline
 
Join Date: Sep 2006
Location: Melbourne, Victoria
Posts: 3,916
Pete,

Make sure you've got everything checked in MaximDL as per the attached.
In particular if your scaling your darks, click on the dark frame group to determine what scaling you are doing. auto-scale is the conventional method, auto-optimize is RMS noise based similar to that used in CCDStack. You may find you'll obtain a better result with the latter - experiment. I typically build the bias and dark frame masters before processing the flats.

Looking at the images in post 28, your flats are over correcting the light frames hence the reason why you have a dark band down the center. You need to ensure your flats are accurate as possible. Ideally, what you should be left with is no system characteristics, i.e. donuts and huge gradient shifts, but more the typical sky glow and smoother gradients caused by light pollution. Looking at the raw light frame and the flat, I would suggest the flat isn't matching the system characteristics.
Attached Thumbnails
Click for full-size image (maximcal.jpg)
170.2 KB80 views
Reply With Quote
  #38  
Old 10-08-2010, 09:15 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Still no joy

Thanks for taking the time to post the screenshot.

My version of Maxim DL has a different layout for this screen but teh settings are the same.

I've tried to image with CCDSoft - same result

I've also used AIP4WIN to calibrate - again, same deal.

I'm now playing with shorter and longer exposures to get the ADU count away from the 20,000 threshold I've been using so far.

Will keep you posted

Pete
Reply With Quote
  #39  
Old 10-08-2010, 09:33 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Dumb question

If I am binnng at 2x, do I need to target the ADU for a flat at half what I target for 1x binning? So if my target is 20,000 ADU, do I reduce that to 10,000 if I am binning 2x?

Pete
Reply With Quote
  #40  
Old 10-08-2010, 10:13 PM
pvelez's Avatar
pvelez (Pete)
Registered User

pvelez is offline
 
Join Date: Aug 2006
Location: Sydney
Posts: 1,250
Progress - I think

OK, I think I'll leave you all n eace for a bit on this one

I've experimented with a range of settings. Sky Flat Assistant allows you to set the target ADU for your flat and then gives you an error margin. The default is 10% which I had been blindly using. Dropping it to 1% made the calibration a lot neater. Its still not perfect - I am sure there is some dark magic involved that I don't comrehend, but its a lot better.

Incidentally, I suspect that the key is not so much the target ADU setting as consistency between the images. I didn't see that much difference between 18,000, 19,000 and 20,000 ADU. But there was a big difference if the allowable error is 10% rather than 1% - even with a median combine.

Does that make sense?

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 10:09 PM.

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