Author Topic: Please, help processing NGC6960 (Solved/Closed)  (Read 11934 times)

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Please, help processing NGC6960 (Solved/Closed)
« on: 2012 August 28 15:45:34 »
Hi all,

I took a shot with my QHY8C (one-shot color CCD) of NGC6960 & I'm finding it extremely challenging to pull the nebular color.  The exposure is about 4.5hours from my light polluted location.  The image has been calibrated, debayered, star aligned, integrated, cropped & DBE.

http://dl.dropbox.com/u/8971416/NGC6960_integrationv1_dc.fit

Typically, what I do after DBE is:

1. BackgroundNeutralization which does not seem to have any effect on my color camera.  It just seems to work for Mono CCD.
2. ColorCalibration which does not seem to have any effect on my color camera.  It just seems to work for Mono CCD.
3. Convert the image from non-linear to linear using STF and Histogram tool
4. ACDNR
5. HDRMultiscalteTransform i.e. Typically a value of 4
6. Histogram Stretch
7. SCNR to remove any Green traces

Now, I've received advise that the Mono processing & color processing are pretty similar, but I'm finding that steps like #1 & #2 does not seem to make any difference in the image.  The background and colors of the stars come out pretty well balanced after debayered & DBE.  Now, if you have a different flow for Color or any tips or tricks that can help me extract the color of the nebula without overly saturating the image, I would love to learn them.

Thanks for your help,

Miquel Casas
Portland, OR
« Last Edit: 2012 September 19 20:51:10 by mcbbcn »

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4543
    • View Profile
Re: Please, help processing NGC6960
« Reply #1 on: 2012 August 28 23:09:33 »
i think you need some flats... and there's a lot of weird gradients left after your first DBE. maybe can you post the non-DBE version?


Offline gvanhau

  • PixInsight Old Hand
  • ****
  • Posts: 344
    • View Profile
Re: Please, help processing NGC6960
« Reply #2 on: 2012 August 29 21:54:20 »
Hello
I agree with PFILE, there are gratients and there is vigneting. There is also a big donut near the center star.
Did you use flat frames?
Either way, I see colors are there...
If you could post a image at an earlier  processing stage, it would be better.

Regards
Geert
Geert Vanhauwaert

Offline jacksonlu

  • Newcomer
  • Posts: 16
    • View Profile
Re: Please, help processing NGC6960
« Reply #3 on: 2012 August 30 00:33:49 »
I recommend just stack all of frames with dark frames only and see what happen on the quality. It seems with wrong process in beginning caused those gradients .  ;)

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Re: Please, help processing NGC6960
« Reply #4 on: 2012 September 02 10:52:12 »
Hi all,

Sorry about my late reply & thank you all for your comments.  I went back to do a new set of flats, just to discover that the inverter for the EL panel was fried, so I tried to fix it by replacing the transistor, but no luck.  I've put the order today for a new EL inverter, and as soon as I get it from Glowhut, I should be able to do my flats, and post a brand new spanking shot of NGC6960 with flats processed.  I'll be reporting on the new image soon.

Thank you all for your prompt answers.  This forum is great & this is a very special community.

Miquel Casas
Portland, OR

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Re: Please, help processing NGC6960
« Reply #5 on: 2012 September 07 23:05:42 »
Hi all,

I got my flat EL fixed, and I created a master flat, but I went processed the calibration with the master flat, something goes completely off with the color.  After I debayer, the subs look completely red, and they used to have the green/blue color of the LPS filter.  It used to be fine without flats, but I have to say that this is the first time I do flats with a color CCD (QHY8L), so I could very well doing something wrong.

So, I used MaximDL skyflats plug-in assistance to capture them.  I targette a 23,000 ADU value.

Here, you have an example of flat:

http://dl.dropbox.com/u/8971416/NGC6960/scope__1x1_skyflat0.fit

Here, you have an example of a flat after calibration of master bias:

http://dl.dropbox.com/u/8971416/NGC6960/scope__1x1_skyflat0_c.fit

I took 15 flat frames, and then I integrated using the multiplicative value as it says on the Pixinsight instructions for creating a master, and here is the result of the master file:

http://dl.dropbox.com/u/8971416/NGC6960/master_flat_frame_calibrated.fit

Just in case, here you have a link to the Master Dark & Master bias, but these files have been working well, and not affecting my color balance:

BIAS: http://dl.dropbox.com/u/8971416/NGC6960/Master_Bias%201_3328x2030_Bin1x1_Temp-15C_ExpTime4sx.fit

DARK: http://dl.dropbox.com/u/8971416/NGC6960/Master_Dark%201_3328x2030_Bin1x1_Temp-15C_ExpTime300s.fit

Here you have a sub after calibration:

http://dl.dropbox.com/u/8971416/NGC6960/_20120826_150s__-15degC_0028_000000951_c.fit

Here you have a sub before calibration:

http://dl.dropbox.com/u/8971416/NGC6960/_20120826_150s__-15degC_0028_000000951.FIT

Obviously, I can't get into processing without solving this puzzle about the flat frames.  I'm not sure why my flat frame is making my subs become red.  Obviously, flat frame is a division, so there is something going on with a color flat.  Do you apply any kind of special process to flats created with color ccd's?

Thank you,

Miquel Casas
Portland, OR
Then, I created the master flat

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4543
    • View Profile
Re: Please, help processing NGC6960
« Reply #6 on: 2012 September 09 00:45:21 »
were the master dark and master bias made in Maxim? the calibrated flat sub and the master flat look like they were "oversubtracted".

if i open the master bias, PI is happy with the file format. but if i open the master dark, it looks like f32 data in the i16 range... sort of.

if some of these masters came from Maxim i'd say that you should start over, creating the master bias and master dark in PI.

if those files came out of PI... well, i don't think they could have based on the content.

finally your uncalibrated light seems to not have very much signal in it. am i correct that the bayer pattern is RGGB?

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Re: Please, help processing NGC6960
« Reply #7 on: 2012 September 10 23:53:44 »
Hi,

The master bias and master dark were created in PI.

You're correct about the uncalibrated light, it's only 150 seconds, but I have about 4 hours of subs, so it'll have signal once I integrate them...if I can get pass the issue with the master flat.

I know my master dark and master bias work well, because I've used them to process other objects that did not need any flats (i.e. they were small DSO's which I could get away without using flats, and the master dark & master bias worked beautifully).  Now, the whole process of creating a master flat from a color CCD in PI seems quite complicated, and I don't know why.  This is what I learned so far, if I calibrate the lights with master bias, master dark, and master flat (without checking the "calibrate" option for master flat during the calibration of the lights), I get a pretty good image and pretty good color balance.  Now, if I calibrate my lights with my master dark, master bias, and master flats (but this time I check the "calibrate" option for the flats), then hell breaks lose.  You may say "Well, don't check the 'calibrate' option for the flats, but if I don't do that, then some of the dark dust still shows in the final integrated light.

Please, let me me know if you need more information.  Thanks for your help.

M.

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4543
    • View Profile
Re: Please, help processing NGC6960
« Reply #8 on: 2012 September 11 12:23:19 »
well, i still can't understand why my version of PI does not like the master dark. the fits reader wants to rescale it. if it was written out with PI the data should be f32, in the range 0..1, and as far as i can tell, it's not. i believe this is a problem.

at any rate, here's what i do (or what i did before the batchpreprocessing script).

master dark : just integrate all the dark frames per Vicent's tutorial
master bias : same.
master flat : first calibrate flat subs by subtracting a master bias. sometimes i used a master flat dark instead. as explained below this could sometimes be a different master bias than the above.

then while calibrating lights, i would include all 3 master frames, turn bias calibration off, turn dark calibration and optimization on, and turn flat calibration off.

this always yielded good results for me.

i am using a DSLR and sometimes i've had to use ISO100 for the flats (i always use ISO1600 for lights). because of this, it does not make sense to calibrate the master flat with the master bias during light calibration, since the ISO of the bias differs. so i would always pre-calibrate the master light.

what is your FITS reader set to? does it silently re-scale out-of-range data?

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Re: Please, help processing NGC6960
« Reply #9 on: 2012 September 11 21:28:13 »
Hi Pfile,

First of all, thanks for all your help.

Yes, my PI does not complain about my master dark...for whatever reason...

But, I'll experiment with your steps, and hopefully, I'll sort it out.

Best regards,

M.

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Re: Please, help processing NGC6960
« Reply #10 on: 2012 September 11 23:07:28 »
Hi Pfile again,

I decided to try something and I got the following results & I would love to have your opinion, because if if you find the image horribly wrong, I won't even go into the details about how I got here.

Basically, I went back to the drawing board, and I decided to tweak my flat one more time, and process just 25 minutes of subs, and these are the results I got...

1) The first image, right before DBE, obviously it has the hideous green of a very fancy LPS filter:

http://dl.dropbox.com/u/8971416/NGC6960/NGC6960_integration_v2.fit

2) Then, I decided to apply DBE.  At this point, I could see the background getting balance, and I started to appreciate sort of natural colors of red (alpha) and blue (oxygen) for the nebula:

http://dl.dropbox.com/u/8971416/NGC6960/NGC6960_integration_DBE.fit

3) Before, I went into further processing, I noticed that I had lots of green stars, so I applied SCNR for GREEN, and this is the image that came out:

http://dl.dropbox.com/u/8971416/NGC6960/NGC6960_integration_DBE_SCNRGreen.fit

...and the nebula is faint, because it's just 25 minutes from a Bortle 8 sky (Portland Downtown, OR)...

If you think the 25 min. image is OK, then I'll process all the subs, and see what happens and re-post again...but if you think there is something really wrong with the final result, then I'll look for another solution...

Thanks for all your help,

M.

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4543
    • View Profile
Re: Please, help processing NGC6960
« Reply #11 on: 2012 September 11 23:54:14 »
well, that does look a lot better. however, there are a lot of hot pixels visible in the integrated result. what integration method did you use and are your subs dithered? how many subs = 25 min?

also there seems to be something wrong with the alignment. or maybe the hot pixels are giving the impression of a mis-registered subs.

Offline gvanhau

  • PixInsight Old Hand
  • ****
  • Posts: 344
    • View Profile
Re: Please, help processing NGC6960
« Reply #12 on: 2012 September 12 09:21:49 »
Hello
this is not a help in your processing, but a hint about taking flats with QHY8L
(I also have this camera).
Remember that this camera uses different readout methods when capturing images, depending on the lenght of the exposure (at 1x1 bining).
So, if you thake pictures of less than 4 seconds, the camera takes two exposures (one for each half of the sensor) and so there is no amp glow on the picture. If you take esposures of more than 4 seconds, it takes only one exposure, and then there is a fixed amount of amp glow on the picture because the amp is on during the readout of the first half of the sensor.

Therefore, to be consistent with your lights, if light exposures are more than 4 seconds, then your flats shoud also be 4 seconds of exposure.

Regards
Geert


Geert Vanhauwaert

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Re: Please, help processing NGC6960
« Reply #13 on: 2012 September 12 18:07:06 »
well, that does look a lot better. however, there are a lot of hot pixels visible in the integrated result. what integration method did you use and are your subs dithered? how many subs = 25 min?

also there seems to be something wrong with the alignment. or maybe the hot pixels are giving the impression of a mis-registered subs.

Hi Pfile,

It was 10 subs of 2 1/2 min. each, so 25 min. in total.

I use MaximDL for Image Adquisition, and I dither my subs.

Regarding integration, I used for:

Image Integration
   -Combination: Average
   - Normalization: Additive
   - Weights: Noise Evaluation
   - Generate Integrated image: Checked
   - Evaluate Noise:  Checked

Pixel rejection (1):
   - Rejection Algorithm: Winsorized Sigma clipping
   - Normalization: Scale + zero offsett

Pixel rejection (2):
   - Sigma Low: 4
   -Sigma High: 3

Are these parameters OK? Would you suggest something different?

Thanks for your help,

M.

Offline mcbbcn

  • PixInsight Enthusiast
  • **
  • Posts: 78
    • View Profile
Re: Please, help processing NGC6960
« Reply #14 on: 2012 September 12 18:12:34 »
Hello
this is not a help in your processing, but a hint about taking flats with QHY8L
(I also have this camera).
Remember that this camera uses different readout methods when capturing images, depending on the lenght of the exposure (at 1x1 bining).
So, if you thake pictures of less than 4 seconds, the camera takes two exposures (one for each half of the sensor) and so there is no amp glow on the picture. If you take esposures of more than 4 seconds, it takes only one exposure, and then there is a fixed amount of amp glow on the picture because the amp is on during the readout of the first half of the sensor.

Therefore, to be consistent with your lights, if light exposures are more than 4 seconds, then your flats shoud also be 4 seconds of exposure.

Regards
Geert

Hi Geert,

It's nice to know that somebody has the same camera that I have.  I use an EL Panel for taking my flats and my MaximDL Sky Flat plugin calculates my ADU target which is 20,000.   In order to do so, it automatically shortens the exposure to have the right ADU, so I would assume that I would have to darken with some filters my EL panel to be able to get 4 seconds or over, right?

Also, I would be really curious to know your process for creating flats.  Do you do anything special with your flats to create them?  Do you use color flats?  Do you use an Light Pollution Filter?

On processing, what is your workflow?  I've noticed that Background calibration and color calibration does not seem to have any effect on my debayered image.  Some people say that the workflow between mono and color is very similar, but I have a mono CCD and color CCD and the workflow seems to be different in several steps.

Thanks for any insights that you can provide.

M.