These past two months since our last alpha test had one primary focus and that was VOIP.
We had plans for first party VOIP support for a long time, pretty much since the inception of this project, and we are very happy with what we have achieved here.

tl;dr: This is first party, directly integrated into the client and server, no configuration required other than opening a second port on the server. Grid based spatial routing works out of the box

We spent a lot of time researching the audio engine and integrate with it fully.
We have our own audioAudioEventMetadata entries, game parameters linked to custom RTPCs, our custom sound banks and we stream VOIP PCM directly into wwise.

We separate between spatial and direct routing, where direct is comparable to a discord call without any distance attenuation or environment occlusion effects. Ideal for implementing phone calls, tactical radio communication or team chat.

We also have scripting capability to control direct voice routing. We allow connecting clients together in exclusive or shared mode

• Exclusive:
  Voice data is only emitted via the direct connection. Whether you are directly next to the emitter entity or on the other side of the map, you will not hear a difference
• Shared:
  Voice data is emitted from the entity too. This can create realistic effects like having a phone conversation with someone and then walking closer to the room where that entity resides and also hearing the muffled entity talk

It is controlled via scripting API on the server side. Example of how to connect all clients on the server to each other:

On the client side we have robust handling of input device selection (with hot plug support), a ton of settings for individual volumes, gain boosting, VAD (Voice Activity Detection) with configurable threshold, PTT (Push to talk), Noise Suppression via RNNoise and/or SpeexDSP.

Some extensions for scripting client sided DSP settings on direct audio and control over spatial falloff are planned for a future update

Media