Wednesday, June 01, 2011

PTC and Color

Albert Theuwissen continues his PTC article series, this time covering color sensors' PTC. The article talks about color-specific artifacts, such as these nice curves:


  1. It would be useful to explore the effects of optical/electrical crosstalk between color channels. It's straightforward enough to separate a Bayer image into four planes and analyze independently.

  2. Why the blue channel has much less FPN than the red and green ones ?

  3. Albert writes in the article:

    "The very first part of the FPN curve is linear and is representing the PRNU of the various colours."

    I think the incoming light might have weak blue component, as often happens in real life. Then the small-signal blue FPN is lower.

    @ "It's straightforward enough to separate a Bayer image into four planes and analyze independently."

    The above graph's non-linear artifacts show that it's not that straightforward, at least not for Albert article's target audience.

  4. Vladimir, if you look at the ratio between RGB responses and the ration between RGB FPN, you will understand what I would like to ask.

  5. OK, I see. The responses differ by a factor of 3, while FPNs - by 5 or so. Then I'd guess it's just different PRNU numbers for RGB that Albert put in his simulator. May be Albert can explain why.

  6. You can see that the Red/Green FPN are almost the same, but their response curves differ a lot !!!!

  7. Albert TheuwissenJune 5, 2011 at 6:37 PM

    Dear all, please notice that this is indeed a SIMULATION, and you get out of it what you put into the simulator. This exercise is done just to verify the working methodology and capability of the PTC technique. If you want, I can also publish measured PTC results obtained from existing cameras and sensors. Albert.


All comments are moderated to avoid spam and personal attacks.