cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
boris.barak@vitec.com
Contributor
Contributor
481 Views
Registered: ‎05-29-2019

Use pixel repetition with hdmi_rx_ss for 480i/576i support

Jump to solution

We are using Zynq xczu4ev-fbvb900-1-e chip in our HDMI RX system solution.

Vivado version is v2018.3.

hdmi_rx_ss is used with conjunction of the software drivers from Xilinix.

The output of hdmi_rx_ss is native video one.

It seems hdmi_rx_ss issues wrong sample values in the case of 480i/576i input of YUV422 or YUV444 color scheme (see attached).

It's obviously could be seen on ILA snapshot view, sampling hdmi_rx_ss outputs. 

In the case of the rest of the formats everething seems good.

Attached Vivado ILA samples directly after mpsoc_wrapper (before any user logic).

The input video pattern is color bar 100%.

It definitly can be seen a good yellow color for the second color bar strip in the case of 480p.

On the other hand we can see a bad color for such case in 480i.

In addition repetited pixels are not the same !

Is any known issue exists here ?

 

Best regards.

Boris Barak.

 

Tags (2)
capture_hdmi_ch1_720x240i60_YCbCr422_12b.bmp
Vivado_ILA_480i_CB100_sample2.png
Vivado_ILA_480i_CB100.png
Vivado_ILA_480p_CB100.png
0 Kudos
1 Solution

Accepted Solutions
xud
Xilinx Employee
Xilinx Employee
331 Views
Registered: ‎08-02-2007

Hi boris.barak@vitec.com 

Please be aware of following notes from PG235 : 

Note: Using AXI4-Stream, pixel repetition only supports RGB and YUV444 color space, but not 12-bitYUV422. Using the native interface, you can create customized pixel replication logic that supports allcolor spaces including 12-bit YUV422.

 

View solution in original post

0 Kudos
2 Replies
boris.barak@vitec.com
Contributor
Contributor
412 Views
Registered: ‎05-29-2019

Hi,

I've realized that in the case of 480i/576i YUV422 pixels samples are repeated in the following way:

...PIX3=[V2Y3V2Y3],PIX2=[Y2U2Y2U2],PIX1=[V0Y1V0Y1],PIX0=[Y0U0Y0U0]

I'm wondering if this structure is menthioned somewhere in in the datasheet.

Finally my pixel subsampling mechamism (1440=>720) is considering this structure.

================================================================

There is another issue to verify.

In the test the hdmi source carries pattern of 100% color bar .

However the samples come from "hdmi_rx_ss" core, over native_video, are slightly different.

This can be seen on the attached ILA snapshots.

Indeed, for the second color (yellow)  strip we get the following samples: Y=0xD21, U=0x101, V=0x924.

Normalizing this to 8-bit gives: Y=0xD2, U=0x10, V=0x92.

Then converting to RGB : R=235, G=235, B=10.

B-level is supposed to be zero here, isn't it ?

Can you check ASAP, please ?

 

Best Regards.

Boris Barak.

 

Vivado_ILA_480i_CB100.png
Vivado_ILA_480i_CB100_sample2.png
0 Kudos
xud
Xilinx Employee
Xilinx Employee
332 Views
Registered: ‎08-02-2007

Hi boris.barak@vitec.com 

Please be aware of following notes from PG235 : 

Note: Using AXI4-Stream, pixel repetition only supports RGB and YUV444 color space, but not 12-bitYUV422. Using the native interface, you can create customized pixel replication logic that supports allcolor spaces including 12-bit YUV422.

 

View solution in original post

0 Kudos