according to the latest release notes (CloudXR 3), it is possible to have bidirectional audio as in audio from the HMD to the server. Sending audio from the server to the HMD works flawlessly, the other way around not so much.
It is unclear to us, how to access the audio input from the HMD on our app running on the server; we were hoping CloudXR would expose a microphone to the system we could then grab as an input, but this is not the case neither on the AWS AMI provided by Nvidia nor our local servers. We’ve added the -ra/-sa flags, but so far have no clue how to process this data serverside (or where).
Is there anything we are missing?