Rebroadcasting of output audio

[partymehard]

I'm running my own mumble server on a small DigitalOcean droplet and recently I've found that I've been broadcasting any output audio through mumble.

This includes anyone who speaks, any music, or any sound from game, being broadcast as if my microphone were picking it up through my speakers. Thing is that I'm using a headset/mic combo, so there isn't any sound coming through the speakers.

I've switched to Push-To-Talk for now, but would prefer to be using voice recognition.

[kissaki]

You’re talking about the audio output of your *client*, right?

That has nothing to do with the server then.

Generally, Mumble listens to a device provided by the system.

What you want to look at is your systems recording settings. E.g. on windows, in the recoding devices dialog verify that only one recording device is active and set as default device, and that it is your microphone. Then in Mumble check that this microphone is selected as the input device (using default system device *should* work, but to check that it’s not a wrong device that is being used, select one specifically).

Some audio drivers allow multiple recording devices to be active in parallel, others not.

Some headsets and computer cases have bad separation of the microphone cable from the audio output cable, resulting in feedback from one to the other.

Furthermore, virtual devices can be set up that combine several devices into one recording device - which is not what you want. Just make sure you don’t use one as the default input device.