• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 99 subscribers
  • Views 964 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.

OMAP5432, VIP, 10-bit, 12-bit Support

David Smalley
Expert 7210 points

First, can someone confirm there are 2 basic input modes for CPI data on OMAP5432

  1. “SYNC” mode described in TRM 8.2.2.3.1 – This mode uses discrete Hsync, Vsync signals and it copies 1 sample per pclk into DDR with no on-the-fly processing (eg CSC or scaling or Y, U/V sorting). The data can be any width up to 16-bits

  2. BT.656 mode described in TRM 8.2.2.3.2 – This mode uses embedded sync codes instead of Hsync, Vsync lines for synchronization. It assumes YUV data sequence as described in ITU.656 spec.  Sample width can only be 8 or 10-bits.

Second question – how to justify data bits when using BT.656 modes?

  1. Figure 8-11 indicates that for 8-bit mode, data should be left justified - cpi_d[11:4], correct?

  2. Then in 10-bit mode, should lines cpi_d[11:2] be used?

 Bonus question

  1. Customer ultimately wants to support 12-bit source. BT.656 spec only supports 8 or 10 bit . Then they would have to use 12-bit SYNC mode?  RAW data input acan be anywhere they want on the 16 data lines.

  • 0
    TI__Genius 14190 points

    David Smalley said:
    “SYNC” mode described in TRM 8.2.2.3.1 – This mode uses discrete Hsync, Vsync signals and it copies 1 sample per pclk into DDR with no on-the-fly processing (eg CSC or scaling or Y, U/V sorting). The data can be any width up to 16-bits

    Yes. However, on-the-fly processing can be done in ISP.

    David Smalley said:
    BT.656 mode described in TRM 8.2.2.3.2 – This mode uses embedded sync codes instead of Hsync, Vsync lines for synchronization. It assumes YUV data sequence as described in ITU.656 spec.  Sample width can only be 8 or 10-bits.

    Yes.

    David Smalley said:

    • Figure 8-11 indicates that for 8-bit mode, data should be left justified - cpi_d[11:4], correct?

    • Then in 10-bit mode, should lines cpi_d[11:2] be used?

    Data mapping can be configured from the ISIF.

    David Smalley said:
    Customer ultimately wants to support 12-bit source. BT.656 spec only supports 8 or 10 bit . Then they would have to use 12-bit SYNC mode?  RAW data input acan be anywhere they want on the 16 data lines.

    Correct, RAW12 should suffice for the use case.

    Thanks,

    Alex

  • 0 in reply to Alex Bashkov
    TI__Expert 7210 points
    Akex,

    Thanks for quick reply.

    OMAP5 TRM describes ISS ISP VP as interfacing to CCP2 or CSI2 in great detail but no mention of using it with CPI. Are you sure that it can accept input form CPI interface?
  • 0 in reply to David Smalley
    TI__Genius 14190 points
    David,

    The ISS supports five simultaneous pixel flows (CCP2, CSI2_A, CSI2_B, CSI2_C and parallel interface [CPI]), but only one of them at a time can use the video processing hardware (through the video port mux); the others can go directly to memory (CPI must always use ISP). CPI input to the ISP can be set from ISS_CTRL[3:2] INPUT_SEL.

    Hope it helps.

    Thanks,
    Alex
  • 0 in reply to David Smalley
    TI__Expert 7210 points
    Alex,

    Yes, this is very helpful!

    So to summarize, they can bring in 12-bit data in RAW mode ( Y and Cr, Cb all 12-bit samples) via CPI_d11:0 into ISP. ISP can pass full 12-bit samples unprocessed and the ISIF storage formatter can be set to right justify the 12-bit sample with 4 MSB's 00'd (per TRM 8.3.3.6.16)?

    And If they need to change camera mode to 10 or 8-bits they can just adjust the ISIF storage formatter to right shift the un-needed bits?

    Thanks,

    David
  • 0 in reply to Alex Bashkov
    TI__Prodigy 175 points
    David,

    On top of what Alex already replied, please refer also to TRM Table 8-424. ISS ISP VP Format Mapping.
    All connections are allowed when the CPI input is the source, though not listed in that table.

    Then, please check also TRM section 8.3.3.6.4 ISS ISP ISIF Interface, where it is explained how ISIF maps the incoming RAW data via the ISIF_CGAMMAWD[4:1] GWDI register bit field.

    Regards,
    Kaloyan
  • 0 in reply to David Smalley
    TI__Genius 14190 points

    David,

    Yes, your assumptions are correct. See also Kaloyan's additional inputs and let us know if this answers your questions.

    thanks,

    Alex