The big problem here is that the preferred timing being sent from the rift to windows is this:Īnd that modeline info doesn't exist in the detailed timing (Specifically Detailed Timing #3 is wrong. This is what proper timing characteristics for my LG IPS277L-BN looks like, and this is what windows enumerates in WDDM 2.0 to figure out what to send to the screen (at least on AMD). rather then cached registry data (like you would get using a DVI-I adapter). Using EnTech Moninfo, I insured that both of my devices were enumerating DDC/EDID 1.3 information in real-time mode. (or if you want, hdmi to your monitor, DVI-D to HDMI adapter on the port BELOW, this will set a more expected display ordering.

The correct way to get direct mode is to use the dvi BELOW the dvi NEXT to your hdmi port, and use the hdmi port for the rift. Most reference cards (ie, sapphire r9 290) share the dvi port with the hdmi port next to it. (this works, because one display is registry cached, not direct.) Most people I assume are plugging the rift into HDMI, while additionally using the dvi-d port RIGHT NEXT to it for their standard display. (this is why it'll work if you change to the DVI adapter.) This, additionally, is also the reason for the black screen crash on AMD r9 290 devices, when both devices are in direct mode. This will require an update whenever you guys are ready. The rift's EIA/CEA-861 Information is incorrect in the firmware. This is the root cause of why the dk2 will not go into extended mode on a WDDM 2.0 driver. Now that I'm slept, I've gotten to the bottom of it. I have another thread where I was throwing a lot of speculation and ideas at the wall.

Wddm 2.0 graphics card