Later...
I looked at the video several times, and it appears that the tail rotor visibly "spun down" from time index 14:15:12 to 14:15:13 (when the tail rotor goes out of view to the right). There's only a few frames to see this, but that's how it appears to me. Perhaps a broken TR trans or TR shaft coupling maybe. Since hover has a high power requirement, the TR would also have a high torque requirement at hover, thus an understandable time for failure.
Very tragic looking stuff.
One thought (and I hope a useful question). In the video, the rotation rate of the fuselage continues to accelerate to the right, even after it has first contacted the flight deck. This suggests that power (and torque) is still being applied to the MR after the yaw to the right has started. If you were in this situation, wouldn't you want to lower the collective fully immediately, and then take your chances touching down on the flight deck with a decreasing yaw rate?
I'm not saying anything against the pilots, who were in an extremely difficult situation. But the video makes it clear that the right yaw was in a constant state of accelerating rate for about 5 seconds after the event started (from 14:15:11 until 14:15:16), until just before the second impact with the flight deck, when the yaw rate begins to slow again.