• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 97 subscribers
  • Views 2165 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TDA2SX: Some questions about Video Input Port

xuanbo
Expert 1550 points

Part Number: TDA2SX

Hi

   I am studying the VIP module of TDA2XX, and have some questions as follows:

Q1: the difference between 24/16/8-bit port
     

      From the manual we can see that, no matter what the input mode is, 24-bit YUV444 or 16-bit YUV422 or 8-bit YUV422, the input data are all Y0~7, Cb0~7, Cr0~7, which can be described as the 8 bit YUV format. For the 16-bit YUV422 and 8-bit YUV422, they have the same YUV format, the difference is the receive method:

    8-bit YUV422 mode receives 8 bit data per clock, but 16-bit YUV422 mode can receive 2*8 bit data per clock. Obviously, 16-bit YUV422 mode is faster.

    So the difference between  24/16/8-bit port is only the data transmission speed? The image "detail" degree they can describe are the same?

 Q2: If 8-bit port is slower, will it limit the frame rate of input signal? for example, can 8-bit port support 30ps, or 60ps?

 Q3: "Supports up to 4095 pixels wide input - without scaling and chroma up/down sampling" means the VIP support up to 4K resolution.

         Does this comment suitable for all  24/16/8-bit port mode? If i choose the 8 bit port mode, can it support 720p, 1080P or 4K input?

Q4: fld Pin

        For Progressive Source, FIELD ID is always ‘0.’  So if i use the progressive source, can i not use fld pin in the video input? Then i can configure this pin for other function.

       For example, vin1a's capture souce is OV10635 which is progressive mode.  Can i not connect the vin1a_fld0 pin, and use it for other module?

Thanks

xuanbo

  • 0
    TI__Guru**** 500535 points
    Hi xuanbo,

    Please find answers to your questions below.

    1, yes, that's correct, 8/16 are both supported for YUV422 data, where image details remain same, only interface changes. For YUV444, 24bit input, it is slightly different as the input format is different. You have full pixel with both the chroma components in YUV444.
    2, well it is not slower, but it takes two clock cycles to transmit one YUV422 pixels, whereas it takes only one clock cycle on 16bit interface. Yes, even 8bit can support 30 and 60 fps, but it depends on the resolution also. The max clock in 8bit is 165MHs, so resolution and fps is such that it is within this limit, VIP can support this resolution over 8bit interface.
    3, again it depends on the fps and pixel clock, as long as pixel clock is within 165MHz, VIP can 720p and 1080p resolution.
    4, fid input is not required for progressive input.

    Regards,
    Brijesh
  • 0 in reply to Brijesh Jadav
    Expert 1550 points

    Hi Brijesh

        Thank you for your explaination.

        About Q1, i have another question. In the "Video Driver Porting Guide for TDA2xx, TDA2Ex and TDA3xx":

        

          I am skeptical of "If the video sensor is providing 10, 12, 14 bit data, then this parameter should be set to FVID2_VIFW_16BIT ".

          Compared with 8 bit data video sensor, 10/12/14 bit data video sensor should have more image details. But according to the conclusion that VIP 8/16 bit port mode have the same image details. So why we need to set the param to FVID2_VIFW_16BIT here?

    Thanks

    xuanbo

  • 0 in reply to xuanbo
    TI__Guru**** 500535 points
    Hi xuanbo,

    Yes, that's correct, 12/10/14 image sensors have more details than 8bit sensors. But when you say, 16bit YUV422, bit depth does not change, it is still 8bit YUV422 data, but transmitted over 16bit interface.

    We require videoIfWidth because it is configuration in VIP. We need to tell whether it should expect data on 8bit or 16bit or 24bit interface.

    Rgds,
    Brijesh
  • 0 in reply to Brijesh Jadav
    Expert 1550 points
    Hi Brijesh

    So if we use 12/10/14 image sensors, "set videoIfWidth to FVID2_VIFW_16BIT " is not necessary?
    We can also use FVID2_VIFW_8BIT for 12/10/14 bit image sensors?

    Thanks
    xuanbo
  • 0 in reply to xuanbo
    TI__Guru**** 500535 points
    Hi xuanbo,

    Well it really depends on your input interface. How many input data lines are connected from your sensor to VIP? If there are more than 8, then you need to use 16bit. Otherwise, you can use 8bit.

    Typoically sensor sends 10/12/14 bit data over 10/12/14 data lines, it is not typical/possible to send 10/12/14 data over 8bit interface. So since it uses more than 8bit, you need to use 16bit interface for VIP.

    Regards,
    Brijesh
  • 0 in reply to Brijesh Jadav
    Expert 1550 points
    Hi Brijesh

    OK, i got it. Thank you very much.

    Regards
    xuanbo