H_COUNT settings

[Kazuma_Sasaki]

Our system outputs video from LCDIF of iMX8DX.

Since the actual behavior is different from what is described in RM, let me confirm the setting of LCDIF_TRANSFER_COUNT[H_COUNT].

From RM(IMX8DQXPRM_Rev.0.pdf)

We set the YCbCr format, so I think the value of H_COUNT should be half the number of active pixels in a line.

But, from the results of our tests it seems that H_COUNT needs to be set to the number of active pixels in a line.

Is the description of RM wrong? or is our understanding wrong?

Test Result :

Probably only half of a frame is displayed.

Best Regards,

Kazuma Sasaki.

[Yuri]

Hello,

   When input data is in YCbCr 4:2:2 format (YCBCR422_INPUT is 1),
H_COUNT should be the number of pixels that should be fetched by the block and the
BYTE_PACKING_FORMAT should be 0xF.

(Note - YCBCR422_INPUT = 1 implies 2 pixels per 32 bits).

Regards,

Yuri.

[Kazuma_Sasaki]

Dear Yuri,

I appreciate your support.

Current RM description:

When input data is in YCbCr 4:2:2 format (YCBCR422_INPUT is 1),
H_COUNT should be the number of 32-bit words that should be fetched by the block and the
BYTE_PACKING_FORMAT should be 0xF.

Your comments:

When input data is in YCbCr 4:2:2 format (YCBCR422_INPUT is 1),
H_COUNT should be the number of pixels that should be fetched by the block and the
BYTE_PACKING_FORMAT should be 0xF.

Will RM be revised according to your comment in the near future?

Best Regards,

Kazuma Sasaki.