VR Expansion Plugin

[mordentral]

@3agle

Hi all,
may very well have been covered before in thread but it is very long and I haven’t been able to search for it effectively, so I apologize if it has been covered.

I’ve encountered plugin in a project I’m working on (I didn’t implement), and I am experiencing a problem with HMD position replicating across the network ( functions fine), causing the local position to jitter at high ping. I am under the impression from some of the code paths in UReplicatedVRCameraComponent that the sending client should not use replicated values, yet that does appear to be happening (the higher the ping the more jitter when laterally moving the headset). Turning off replication on the component results in no lateral movement of the HMD in the world. Is expected behaviour for the plugin? I would have expected the local player would always update using the transform given by the HMD. Could it be that there is some sort of authoritative HMD position flag somewhere?

Also the code is based on plugin version 4.22 and there have been modifications to it within the project, so fully willing to accept that is not a plugin issue, I just want to establish what the plugin *should *be doing, if possible.

Turning off replication will not effect the local movement of the HMD normally no, it sounds like your client somehow set it up so that the player is not fully possessing the pawn and is receiving camera replication on the camera they are using for their local view. The clients send their local camera transform out to the server via RPC and it is returned back to all (non owning) clients via a replicated variable.

Not entirely sure how that could happen, it would take some effort to break it as much as that. Its possible that they removed the Cond_SkipOwner part of the replicated variable?
I still have users in multiplayer in 4.22 so it is highly unlikely that a feature back port broke any of that (still possible, just very unlikely).

The camera updates its local position on GetCameraView, which should set later in the frame than replication, so the mostly likely outcome is some sort of setup where they aren’t fully controlling their pawn somehow.