STF-8300C Overscan and Trim Regions?!

Discussion in 'STF Series CCD Cameras' started by Hamed-AM, May 3, 2020.

  1. Hamed-AM

    Hamed-AM Standard User

    Joined:
    May 3, 2020
    Messages:
    3
    Hi All,

    I would like to know about overscan and trim regions in STF-8300C camera if exists any.
    I have heard about physical and visual overscan strips in most professional CCDs, but I'm interested in to know if there is such a concept with STF-8300C too.

    One more thing is about the image size (as per fit(s) header file), active image area and effective image area (according to chip datasheet). These are 3352x2532, 3326x2504 and 3366x2544. This is a bit confusing, and I can not realize which one is the "real" pixel dimensions I need to use in my reduction process.
    Any clarification would be highly appreciated.

    Thanks
    Hamed
     
  2. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    4,293
    Were you in touch with @Bill about this?
    I sent the datasheet for the sensor to him.
    You can find the data sheet here:
    https://www.onsemi.com/pub/Collateral/KAF-8300-D.PDF
    It explains the regions on the sensor.
    The active pixels are the ones you normally will get from our cameras and software. The innermost dimensions.
     
  3. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    7,661
    The "Effective Image Area" is the recommended region of the chip to use. The "Active Image Area" adds a few more pixels on the edge that may actually work, but aren't guaranteed to work properly; for example, sometimes they aren't as sensitive as the rest of the pixels.

    You can enable overscan using the SBIG Driver Checker's Config Drv button. Normally it is set to 0. This adds extra pixels at the end of each line.

    Looking at the diagram on the sensor datasheet Page 3 it looks like you would need to enter 67, but only the last 39 pixels of that would be actual dark overscan pixels. The rest would need to be ignored.
     
  4. Hamed-AM

    Hamed-AM Standard User

    Joined:
    May 3, 2020
    Messages:
    3
    Colin and Doug, thank you very much for your responses to my issue.
    Doug told about the pixels (39 of them) at the end of a line to be supposed as dark overscan pixels. By the way, there are also so-called not-guaranteed pixels at the beginning of each line, as well as at the beginning and end of columns too. How should one deal with these extra pixels?! Is it yet wise to relay on quoted 332682504 area as the best region of the image, without trimming extra pixels off?!
     
  5. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    4,293
    The 3326 x 2504 active pixels are the useful ones for the image data.
    Toss out the not-guaranteed pixels - these are used during manufacturing for testing purposes, and you can't rely on their behaviour.
    If you had 2 different KAF-8300C sensors, made at different times, the values of these extra pixels could be different under identical conditions.
     
  6. Hamed-AM

    Hamed-AM Standard User

    Joined:
    May 3, 2020
    Messages:
    3
    Thank you, Colin
    So how is the layout of this useful pixels? Simply say 1-3326 and 1-2504?!
     
  7. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    4,293
    If you use the default settings in our drivers and software, they are 0-3325 and 0-2503.
    If you enable the overscan capabilities, then you need to count in by the rows and columns in the datasheet.
     

Share This Page