we have integrated our solution with Omniverse. At the end we want to use Omniverse for background rendering of images and for real time rendering. Currently we use websocket streaming…
The streamed video feels a little bit sluggish. Even on the same machine (my workstation in the office is a Xenon with 10 cores and a Quadro RTX 5000 and my workstation at home is an AMD Threadripper with 32 cores and a RTX3080, so hardware shouldn´t be the problem :-) ). Is this related to the websocket connection? Should we switch to WebRTC now? What can we do to optimize this?
We have setup a Linux machine with everything we need (nucleus, create and our own kit agent). Everything works as expected. Now we would like to setup a machine with new hardware, that could be moved to out datacenter later, so what is the recommendation regarding the graphics card to use (when money is not the problem, and we want to achieve the maximum possible concurrent user count) for our scenario (real time server sided rendering)
btw. It “feels” like the live stream is in 21.3. more sluggish than in 21.1