02-21-2018, 01:01 PM
If the RK3328 TMDS clock is indeed limited to ~371MHz (hardware limitation, not just a bug in the driver), then they should not advertise the Chip specs as having HDMI 2.0a.
They would need at least 533MHz TMDS clock (for reduced blanking video modes), and more like 594MHz (standard 60Hz). If true, RK3328 really should be advertised as HDMI 1.4b.
(it's also possible that the Rock64 implementation introduced some clock limit that's not inherent to the chip, but rather some PCB capacitance, etc., that limits it)
also a clarification: I'm only referring to the 4K refresh rate capability of the graphics chip, not the CPU/GPU/VPU frame rate decoding capability with any codec. It's impressive that it can decode 4K content at or near 60fps, but pointless to do so at any frame rate greater than the refresh rate. There's still a lot of movie content that is 24fps, so it's useful to set display resolution of 3840x2160@24Hz in those cases, but that resolution spec was part of HDMI 1.4, back in 2009...
They would need at least 533MHz TMDS clock (for reduced blanking video modes), and more like 594MHz (standard 60Hz). If true, RK3328 really should be advertised as HDMI 1.4b.
(it's also possible that the Rock64 implementation introduced some clock limit that's not inherent to the chip, but rather some PCB capacitance, etc., that limits it)
also a clarification: I'm only referring to the 4K refresh rate capability of the graphics chip, not the CPU/GPU/VPU frame rate decoding capability with any codec. It's impressive that it can decode 4K content at or near 60fps, but pointless to do so at any frame rate greater than the refresh rate. There's still a lot of movie content that is 24fps, so it's useful to set display resolution of 3840x2160@24Hz in those cases, but that resolution spec was part of HDMI 1.4, back in 2009...