Author Topic: Tutorial for raw_diag.mo?  (Read 2144 times)

Sapporo

  • Freshman
  • **
  • Posts: 55
Tutorial for raw_diag.mo?
« on: February 26, 2017, 11:07:20 AM »
Is there a tutorial for raw_diag.mo like the one for raw histogram http://www.magiclantern.fm/forum/index.php?topic=12096.0

a1ex

  • Administrator
  • Hero Member
  • *****
  • Posts: 11480
  • 5D Mark Free
Re: Tutorial for raw_diag.mo?
« Reply #1 on: February 26, 2017, 02:39:56 PM »
No, but this topic should help: http://www.magiclantern.fm/forum/index.php?topic=10111

Do you have a specific question about it?

Sapporo

  • Freshman
  • **
  • Posts: 55
Re: Tutorial for raw_diag.mo?
« Reply #2 on: February 27, 2017, 08:31:34 PM »
No, but this topic should help: http://www.magiclantern.fm/forum/index.php?topic=10111

Do you have a specific question about it?
No specific question. What I am after is something like The Children's Bible instead of the whole Bible. But your links in your link seems to be a good start http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/index.html
http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p2.html#SNR-DR

a1ex

  • Administrator
  • Hero Member
  • *****
  • Posts: 11480
  • 5D Mark Free
Re: Tutorial for raw_diag.mo?
« Reply #3 on: February 27, 2017, 08:43:50 PM »
Okay, so you are probably looking for general noise and DR theory, and how to measure DR at different settings.

There are two important methods in raw_diag for measuring DR:

- Optical Black + DR: it estimates DR from white level (from an overexposed subject) and noise from optical black area. This is a very crude approximation (though still useful to get a rough idea), mostly because we found later the optical black noise may not match the actual image noise, and the exact behavior can be model-specific.

- SNR Curve (2 shots): it estimates the noise profile from the difference between 2 pictures of the same static subject (also giving full well capacity and read noise). DR estimations with this method should be pretty good, as you can see the DR directly on the graph, but the other parameters (found by curve fitting) might not be very repeatable.

Relevant links for the second method:

http://www.magiclantern.fm/forum/index.php?topic=10111.msg117955#msg117955
http://www.magiclantern.fm/forum/index.php?topic=13787.0

Sapporo

  • Freshman
  • **
  • Posts: 55
Re: Tutorial for raw_diag.mo?
« Reply #4 on: March 05, 2017, 07:50:01 PM »
Thank you very much.