Thread Rating:
  • 1 Vote(s) - 5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Help  Obi Rope Photon Unity Networking2
#11
(17-08-2021, 08:23 PM)JungleEd Wrote: Is there anyone working on adding a network layer to obi rope or obi fluids?

I don't know for sure, but TheMunk seems to be doing so.

I don't think a network layer will ever be officially added to Obi since it's far beyond the scope of the asset: it's a physics engine, not a networking system.

(17-08-2021, 08:23 PM)JungleEd Wrote: Would moving a particle(synching) during the simulation break the simulation?

Nope. Obi is unconditionally stable so no matter how much you abuse particles during simulation, it will never break unlike Unity's physics (by break I mean entering an unstable state from which it cannot recover). However I don't think it makes any sense to do this in the context of networking, unless you sync all particle's positions and velocities in a way that all constraints are satisfied as far as the client simulation goes. Otherwise the client simulation would take over and move your particles back to where they "belong" right after you make any change.
Reply
#12
(18-08-2021, 07:47 AM)josemendez Wrote: I don't know for sure, but TheMunk seems to be doing so.

Using Photon Fusion (beta) seems to be working fine. 200 particle positions synched with a tick rate of 60 and lag compensation looks alright and latency is very low with local peer-to-peer. I'm targeting the Quest 2 so may have to up the tick rate to 72 to smooth out the motion or implement client prediction, although I think that will require the clients to run the simulation too as well as syncing velocities which basically doubles the network traffic.

The cool thing is that skipping the simulation on clients means Obi can run many more substeps/iterations for better simulation on a PC server. Basically, only the amount of particles that can be synced over the network is the limiting factor now.
Reply