[noggin]

fritsch - not sure what you are saying? My TV reports 12 bit YCbCr as '12 bit', and my amp reports it as 36bit (Presumably the TV reports the single channel bit depth, the AVR the total number of bits used for each channel - though it may well not be taking into account subsampling. Things would get confusing if it started reporting 12-bit for 8-bit 4:2:0 for instance, or 20-bit for 10-bit 4:2:2)

Or are you saying that the AVR reporting a 12 bit input from an Intel GPU feed is just wrong?

Just had a look at my Sony BDP-S5500 Blu-ray player feeding through my Onkyo AVR into my Sony UHD TV.

Not playing any video - just in the player menu :

TV Reports :
1080p 12 bit

AVR Reports :
Input HDMI 1 : YCbCr 36 bit 1920x1080p 60Hz
Output HDMI Main : YCbCr 36 bit 1920x1080p 60Hz

I suspect my Blu-ray player is permanently outputting 12 bit depth YCbCr.

AIUI 12 bit YCbCr (and 10 and 16 bit) were added to relatively early HDMI specs as 'Deep Color' marketing... The Blu-ray spec is 8 bit 4:2:0 YCbCr, so I guess only file-based content you shoot yourself might benefit from this deep color (other than a small potential benefit of avoiding truncation on resampling maths errors)?

To return to the original question however, if the AVR is reporting that as an input source (and not an output source) then that's a bit odd (unless it is getting confused)? However it wouldn't be odd for an AVR to resample 8-bit to 10 or 12-bit if it is running in non-passthrough mode and scaling all content to a fixed output format (which could well be 12 bit rather than 8 bit)