Author Topic: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow  (Read 1061 times)

Offline jliu

  • Newcomer
  • Posts: 20
    • View Profile
There may be a possible bug in the image calibration process.  Please refer to the attached screen capture. Using the process in the prescribed way (master bias, master dark with optimize box checked, master flat) did not sufficiently remove the amp glow on the right side of the top window.

Instead, by manually doing it, first using image calibration with the master flat only to get the result in the middile window and secondly using pixelmath to subtract a scaled difference (i.e., 1.67 = 5 min/ 3 min)of the master dark and the master bias removed the amp glow.  I believe this is the way image calibration is supposed to work.  So, I am reporting it as a possible bug.

Jack

Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1296
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #1 on: 2019 February 23 21:55:23 »
I don't know that you can hold PI responsible for the foibles of your sensor :)  For sensors with bad amp glow it is usually best to use darks of exactly the same duration as your lights and you may need to disable dark optimization as well.

Cheers,
Rick.

Offline bulrichl

  • PixInsight Old Hand
  • ****
  • Posts: 438
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #2 on: 2019 February 24 02:37:03 »
You should not use Dark Frame Optimization with cameras that exhibit strong "amp glow".

The preferred workflow for image calibration in this case is:

Capturing of calibration files: take dark frames, flat frames and flat dark frames

Preparation of the masters:
integrate the dark frames -> MasterDark,
integrate the flat dark frames -> MasterFlatDark,
calibrate the individual flat frames with the MasterFlatDark,
integrate the calibrated flat frames -> MasterFlat (*)

Light frame calibration settings:
uncheck section 'Master Bias',
check section 'Master Dark', select the MasterDark, uncheck both 'Calibrate' and 'Optimize',
check section 'Master Flat', select the MasterFlat, uncheck 'Calibrate'

Bernd

-------

P.S. I forgot to list the integration of the calibrated flat frames, sorry.
« Last Edit: 2019 February 25 01:34:31 by bulrichl »

Offline jliu

  • Newcomer
  • Posts: 20
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #3 on: 2019 February 24 15:26:54 »
Prior to my discovering the two-step Image Calibration - Pixelmath process described in my original post above, I had unsuccessfully tried various combinations of bias / no bias, dark / no dark, calibrate/no calibrate, and optimize/no optimize, including, the suggestion of the second reply. The bias, dark, and flat masters were prepared as prescribed in http://www.pixinsight.com/tutorials/master-frames/index.html .  The two-step approach successfully removes the amplifier glow completely as well as correct for vignetting.

The camera in question uses the Sony ICX183M chip.  Remarkably I have no problems with using its sister OSC camera with the ICX183C chip.  I run the raw subs through IC as prescribed and then debayer.  The amp glow is completely removed.

Jack

Offline bulrichl

  • PixInsight Old Hand
  • ****
  • Posts: 438
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #4 on: 2019 February 25 05:55:43 »
It is always necessary to apply the dark current and bias correction first, and the flat field correction subsequently. It cannot work in reverse order. The flat field correction shall correct uneven field  (e. g. vignetting produced by the optics and dust in the light path causing shadowing),  and different light sensitivity of the sensor pixels. Amp glow is not influenced by either of them.

So I suppose you did your "two-step Image Calibration" different from your description:

Instead, by manually doing it, first using image calibration with the master flat only to get the result in the middile window and secondly using pixelmath to subtract a scaled difference (i.e., 1.67 = 5 min/ 3 min)of the master dark and the master bias removed the amp glow.

Bernd

Offline jliu

  • Newcomer
  • Posts: 20
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #5 on: 2019 February 25 21:09:09 »
Thanks for the comment.  I will try it.  It makes sense to apply the dark and bias correction first, but in this case I did flat field correct first followed by the latter. Probably, the ICX183 chip (9 mm x 13 mm) is small relative to the vignetting such that the discrepancy is unnoticeable.  The OTA used was the FSQ-106ED at f/5 which I think has a 45 mm advertised image circle.

Image Calibration works fine with my ICX814 based SXVR-H814 camera over the past 5 years, but this is a weird case with this camera. Oh well, I'll just have to do with what works for me.

Offline bulrichl

  • PixInsight Old Hand
  • ****
  • Posts: 438
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #6 on: 2019 February 26 01:46:52 »
As far as I know there is no Sony ICX183 but only a IMX183 (color and monochrome version). In the denomination, C stands for CCD, M for CMOS sensors. So you compare cameras with very different sensors:

The SXVR-H814 has a CCD sensor (ICX814) which does not show amp glow at all. Your other camera seems to have a CMOS sensor (IMX183M), to be specific the CMOS sensor with the strongest amp glow on the market. Nevertheless you can remove this strong amp glow with a correct image calibration. Please try the exact approach that I described above. I'm sure that it will work.

Bernd

Offline jliu

  • Newcomer
  • Posts: 20
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #7 on: 2019 February 27 17:26:22 »
Yes, it is an IMX183M.  As you pointed out, subtracting the dark and bias first before applying the flat works fine.  Thanks.

I purchased both the mono and color versions of the camera.  IC works fine with the OSC version.  My issue was with the mono camera.  Other than the amp glow, these chips exhibit low noise.

Offline bvalente

  • Member
  • *
  • Posts: 55
    • View Profile
Please try the exact approach that I described above. I'm sure that it will work.

Bernd

Hi Bernd - just wanted to say your process above worked brilliantly for me, i was searching the internet all today looking for the answer here. Thanks

Brian

Offline bulrichl

  • PixInsight Old Hand
  • ****
  • Posts: 438
    • View Profile
Hi Brian,

thank you for the feedback. I'm glad that it worked for you.

It's a pity when advice is given and no feedback comes back. Such threads then are not always helpful for users that have the same issue.

Bernd

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 6939
    • View Profile
    • http://pixinsight.com/
Quote
It's a pity when advice is given and no feedback comes back. Such threads then are not always helpful for users that have the same issue.

Absolutely! Should we write this sentence in stone and put it somewhere well visible?
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline rkymtnbiker

  • Newcomer
  • Posts: 1
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #11 on: 2019 November 13 08:27:35 »
Team,

I'm having a similar problem with my latest photos of the Western Veil with an ASI294 OSC camera.  I used the DarkFlat workflow described by Bernd.  I tried this workflow both with and without Dark Frame optimization.  This workflow has worked for other photos I've done and has completely removed the amp glow in my previous photos.  For these Lights, I used 240s exposures and a gain of 200.  I've attached a screenshot showing an original light frame, the MasterDark created from 30 shots of the identical exposures and temp for the lights, the same light frame after being calibrated by the MasterDark and MasterFlat, that calibrated light frame debayered to show the amp glow better, and finally the calibrated, CosmCorr, Debayered, Registered, integrated photo with ABE applied. 

Any Advice?

Thanks,

Ted

Update:  Well, I have good news and bad news:  Good news is that when I re-ran BPP without Dark Frame optimization, the amp glow is now completely removed!  Bad news is my integrated photo is WAY different than the one I processed 3 days ago.  Now, the integrated photo is completely green and ABE and Color Calibration will not automatically correct for the green tint.  I'm experimenting with custom settings in Color Calibration to see if I can fix that.

But, the amp glow is fixed!
« Last Edit: 2019 November 16 17:34:12 by rkymtnbiker »

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4531
    • View Profile
Re: Image Calibration Process Doesn't Seem to Properly Remove Amp Glow
« Reply #12 on: 2019 December 03 12:47:57 »
the green tint should be removable, if you want to post the XISF of the integrated image you're having trouble with someone can take a look.

rob