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.

AM5728: Field detection logic for interlaced video

Part Number: AM5728

Hi,

On AM5728 video input port, how is the field ID detected in case of discrete sync capture?
Is it safe to rely on the  VPDMA write descriptor for detecting the field ID?

Regards,
Prasad.

  • Hi,

    The video experts have been notified. They will respond here.
  • Prasad said:
    On AM5728 video input port, how is the field ID detected in case of discrete sync capture?

    Please check TRM document section 9.4.5.6.9 Field ID Determination Using Dedicated Signal.

    Prasad said:
    Is it safe to rely on the  VPDMA write descriptor for detecting the field ID?

    Yes. The VIP_PARSER extracts the FIELD ID information from the sync (VS/HS or VS/FID) and sets the FID bit on the frame it sends to VPDMA.

    The VPDMA simply saves the FID bit value and includes it in the write descriptor.

  • Thanks Manisha for the quick response.

    We have configured the video port in H:V:DE mode. And we have set the following register fields for the same
    VIP_PORT_A[14] FID_DETECT_MODE = 1
    VIP_PORT_A[15] USE_ACTVID_HSYNC_N = 1
    VIP_PORT_A[22] DISCRETE_BASIC_MODE = 1

    I am not able to detect the field ID in this mode (using VSYNC skew). The field ID is always high in write descriptor.

    The TRM register description for VIP_PORT_A[22] DISCRETE_BASIC_MODE says
    In Basic discrete and ACTVID mode
    "capture works as follows: - VSYNC is expected to
    transition from inactive to active between ACTVID
    window. - This VSYNC transition allows the next line in an
    ACTVID envelope to be sent to a new VPDMA buffer. -
    FID value is determined by the FID pin value on the cycle
    where VSYNC transitions from inactive to active"

    Since we don't have field signal (H:V:DE mode), how do we detect field ID in this case?

    Regards,
    Prasad.
  • From TRM, below description on DISCRETE_BASIC_MODE field of VIP_PORT_A register. FID determined by VSYNC skew only in Normal Discrete Mode. 

    This register is valid for Discrete Sync mode only. 0 = RW 0x0
    Normal Discrete Mode. Hsync Style Capture operates as
    follows: - Captures line starting from HSYNC inactive to
    active condition. - VSYNC determined by sync window. -
    FID can be determined by VSYNC skew or captured from
    pin at first pixel in first line. ACTVID style capture works
    as follows: - Captures line during contiguous ACTVID
    envelope. - VSYNC is captured at the first pixel in each
    line. - FID is captured on first pixel of ACTVID window. 1
    = Basic Discrete Mode. When using hsync with Hsync
    Style Capture operates as follows: - The last line of active
    video ends on the pixel clock cycle where VSYNC
    transitions from inactive to active. - FID pin value is
    captured on this cycle and is used for the next field. - FID
    detection by VSYNC skew is not allowed. ACTVID style
    capture works as follows: - VSYNC is expected to
    transition from inactive to active between ACTVID
    window. - This VSYNC transition allows the next line in an
    ACTVID envelope to be sent to a new VPDMA buffer. -
    FID value is determined by the FID pin value on the cycle
    where VSYNC transitions from inactive to active. In basic
    discrete mode, there is no Vertical Ancillary Data.
    Therefore, VPDMA descriptors should not use Ancillary
    Data channels.

  • So, from your response what I understand is..

    In H:V:DE mode the VIP has to be configured in Normal Discrete Mode, Active Video style capture and VSYNC skew for Field ID determination.

    VIP_PORT_A[14] FID_DETECT_MODE = 1
    VIP_PORT_A[15] USE_ACTVID_HSYNC_N = 1
    VIP_PORT_A[22] DISCRETE_BASIC_MODE = 0

    I will try these settings and post the result. Thanks for the support.

  • Hi Prasad,

    Can you please update the results with the modified settings for us to be able to close this thread?

    Regards, 

    Manisha