Survey: YCbCr 4:2:2 availability / support with EDID 1.3 vs 1.4

So I have a Panasonic TV with HDMI. nvidia-settings allows me to pick RGB and YCbCr444 in the Color Space option. However according to the CEA extension block of its EDID, it shows that it supports YCbCr422 as well.

However, since its EDID is 1.3, there’s a chance that the nvidia driver would misinterpret similar info in the main block if it treats it with the same way it use on a EDID of version 1.4.

So I want to do a survey on whether anyone is ever capable to see/pick YCbCr422 in nvidia-setting on Linux. Would be nice if you can paste/upload your EDID in either ASCII or binary format as long as you can see either YCbCr444 or YCbCr422 on the option’s drop list (but not the help text :P).

I have actually already filed a bug report (Reference Number: 151126-000000). In case any moderator can confirm that the the technical team has already confirmed/identified the issue, it would be nice if you can let me know.

Content of my bug report in case anyone is interested in the details:

Does the nvidia linux driver supports YCbCr 4:2:2 output? It seems to me that the answer is positive according to the help text showed when pointing to the option; However, I can only get RGB and YCbCr 4:4:4 out of the list.

It seems to me this is because the driver interpret the EDID wrongly and ignore the true capability of my TV by ignoring the corresponding attribute in the CEA extension block.

Byte 3 of its CEA extension block is 72h (01110010b), which clearly shows that it supports both YCbCr 4:4:4 and YCbCr 4:2:2 since bit 5 and 4 is 1.

However, because the EDID is of version 1.3, while Byte 24 (18h) is only defined to be used to indicate capabilities on YCbCr color space in EDID 1.4. Therefore, the byte has a value of 0Ah (00001010b).

In EDID 1.3, when bit 4 is 0 and bit 3 is 1, it indicates “RGB color display”, even when bit 7 of Byte 20 (14h) is 1 (which indicates that it’s digital). However, in EDID 1.4, with the same values, it indicates that the display supports “RGB 4:4:4 & YCrCb 4:4:4”, which looks exactly the same as how the driver recogize my TV.

So apparently the driver is not aware of this difference in the two version of EDID and does not depend on the CEA extension block to determine YCbCr capabilities for EDID 1.3.

Attached is the dumped EDID in binary format.

panasonic.bin.gz (232 Bytes)