[lmyllari]

lmyllari, thank you. You're doing more scientific testing than I am.

I think it's best to wait for more details and a reproducible test case. If you'd like to file the bug report, please do. If you'd rather not, I'll take your information and try my best on the mailing list and bugzilla.

Cheers

Alchete, and any other who would like to help:

I think there is a driver issue with signaling the quantization range over HDMI. This is why the xrandr workaround is needed. When the driver notices it's connected to a TV (a CEA video mode and EDID indicating quantization range select support), it automatically sets 16-235 range. It should also tell the TV it's using this range (by sending AVI infoframes).

I think the infoframes are not being sent properly.

This is based on my observations from one system only, so I'd like to have a bit more data before reporting it.

If you have a DVD or bluray player (or some other video source) that lets you select between full and limited range RGB output, please test the following, and let me know with your source and TV make/model.

1) Try switching between limited and full range on your NUC with

Code:

xrandr --output HDMI1 --set "Broadcast RGB" "Full"

and

Code:

xrandr --output HDMI1 --set "Broadcast RGB" "Limited 16:235"

- does the picture change between black and grey black level?

2) Try the same switch with your other video source. Does the picture change between black and grey black level?

I have tested this with Oppo BDP-93 and KRP-500m, and the Oppo seems to control the monitor input range. There is no change in black level when switching RGB ranges. (I verified with manual settings from the monitor that the range had indeed changed).



(If you interested in technical details, see https://bugs.freedesktop.org/show_bug.cgi?id=46800#c20 for the driver code that implemented this stuff)