Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Flat fielding on a WO RedCat 51

  • Please log in to reply
10 replies to this topic

#1 enns

enns

    Lift Off

  • -----
  • topic starter
  • Posts: 5
  • Joined: 28 Jun 2017

Posted 03 December 2020 - 12:20 PM

Hi folks,

 

I recently dusted off my RedCat 51 (purchased from the kickstarter campaign but not used until covid days). I am having a major challenge getting my flats to remove the heavy vignetting with a FF camera. I am sure this is user error and a rookie mistake but I'm a bit stumped.

 

To check my work I used pixel math to divide a single light frame by a flat frame ( $T/flat * avg(flat) ) and the result is not good. Similarly when I stack and integrate the entire sequence I end up with a nasty looking vignetted master light frame.

 

The details are:

scope: WO RedCat 51

camera: Canon EOS R modified

 

Flats were taken immediately following light frames, using a light panel.

 

Thanks in advance for your insight!

 

flat frame.jpg

light frame.jpg

result.jpg

 



#2 sharkmelley

sharkmelley

    Skylab

  • *****
  • Posts: 4,181
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 03 December 2020 - 12:32 PM

Where's your bias?

 

Try ( ($T-bias)/(flat-bias) * avg(flat-bias) )

 

Mark


Edited by sharkmelley, 03 December 2020 - 12:32 PM.


#3 enns

enns

    Lift Off

  • -----
  • topic starter
  • Posts: 5
  • Joined: 28 Jun 2017

Posted 03 December 2020 - 08:14 PM

Thanks Mark, I did that with the bias also, with a similar result that shows heaving vignetting.

 

Rob



#4 jonnybravo0311

jonnybravo0311

    Viking 1

  • -----
  • Posts: 653
  • Joined: 05 Nov 2020
  • Loc: NJ, US

Posted 03 December 2020 - 09:52 PM

Where's your bias?

 

Try ( ($T-bias)/(flat-bias) * avg(flat-bias) )

 

Mark

Forgive me for the quick hijack, but I'm trying to learn/understand pixel math and your expression doesn't work when I enter it in PI's pixel math process. Specifically, it balks at the "avg(flat-bias)" part of the expression. I was able to get around this by first creating a separate "flat-bias" image and then using that in the expression you shared. In other words, I did this:

  1. Open a flat frame and a bias frame
  2. In pixel math, put in the expression: $T-bias
  3. Have it create a new image as output
  4. Drag the process triangle onto the flat frame
  5. Open a light frame and close the flat frame so that I now have a light frame, a bias frame and the result of my expression in step 2 (I renamed it to flatMinusBias)
  6. In pixel math, enter a new expression: (($T-bias)/flatMinusBias * avg(flatMinusBias))
  7. Drag process triangle onto the light frame
  8. Resulting new image frame should be what your original expression tried to obtain

Did I get that right?

 

Thanks!



#5 enns

enns

    Lift Off

  • -----
  • topic starter
  • Posts: 5
  • Joined: 28 Jun 2017

Posted 03 December 2020 - 10:52 PM

Yes... I found the same issue with the expression not working as written, and a similar workaround, to precompute (flat - bias).


  • jonnybravo0311 likes this

#6 jonnybravo0311

jonnybravo0311

    Viking 1

  • -----
  • Posts: 653
  • Joined: 05 Nov 2020
  • Loc: NJ, US

Posted 03 December 2020 - 11:11 PM

Yes... I found the same issue with the expression not working as written, and a similar workaround, to precompute (flat - bias).

Thanks for the confirmation :)



#7 sharkmelley

sharkmelley

    Skylab

  • *****
  • Posts: 4,181
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 04 December 2020 - 06:58 AM

Thanks Mark, I did that with the bias also, with a similar result that shows heaving vignetting.

Interesting.  I use a FF camera with the RedCat without problems but for some reason your flat isn't matching the light.  Try a dawn/dusk sky flat instead to see if that helps.

 

By the way, sorry for the mistake in the Pixelmath formula:  avg(flat-bias).  I had forgotten that avg works only for single images and not for expressions.

 

Mark



#8 enns

enns

    Lift Off

  • -----
  • topic starter
  • Posts: 5
  • Joined: 28 Jun 2017

Posted 04 December 2020 - 08:52 PM

Good news, yet confusing news. Through some trial and error, using the ImageCalibration process and then verifying with PixelMath, the following works:

 

(($T - bias) / flat) * avg(flat)

 

But if:

1. both the flats and lights are calibrated with bias

or

2. neither the flats nor lights are calibrated with bias

 

then the flat division results in a badly corrected frame.

 

The result I verified in PixelMath with RAW CR3 files for the light, flat, and bias frames.

 

I am confused about this result as it seems that either both, or neither the flat and light should be bias subtracted.



#9 sharkmelley

sharkmelley

    Skylab

  • *****
  • Posts: 4,181
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 05 December 2020 - 03:01 AM

Lights and flats should both be calibrated by the bias - it's a simple mathematical result.  If you fail to use the bias then dividing the light by the flat will usually under-correct or over-correct depending on the signal levels in each.

 

Mark



#10 enns

enns

    Lift Off

  • -----
  • topic starter
  • Posts: 5
  • Joined: 28 Jun 2017

Posted 05 December 2020 - 09:14 AM

Yes, thanks Mark. I think the root cause issue might be (just realized this):

 

Lights, Darks, Bias: ISO 800

Flats: ISO 100

 

Calibrating the flats with bias at a much higher ISO looks problematic. So will gather some new data at identical ISOs.



#11 sharkmelley

sharkmelley

    Skylab

  • *****
  • Posts: 4,181
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 05 December 2020 - 09:30 AM

Calibrating the flats with bias at a much higher ISO looks problematic. So will gather some new data at identical ISOs.

Yes, I think the EOS R is one of the cameras that has a different bias level at low ISO.

 

Mark




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics