Stereo Sound Not Working!

[rabcor]

Since version 1.2.4, Mumble supports the Opus codec which is a stereo codec. (unlike celt and speex which are mono)

Or as it says here: https://wiki.mumble.info/wiki/1.2.4#Next_generation_audio_codec

 

"The Opus codec is designed to handle a wide range of interactive audio applications, including Voice over IP, videoconferencing, in-game chat, and even remote live music performances. It can scale from low bit-rate narrowband speech to very high quality stereo music."

 

(heh, stereo so next gen, that totally hasn't existed in like... every audio format since mp3... and this was supposed to be "next gen" in 2013? but i digress...)

I set my server to force the Opus codec (actually this option shouldn't even be around anymore, considering that everyone should have a much higher version than 1.2.4 by now) just to be sure, but my sound is always delivered in mono format.

How do I fix this? How do I get stereo sound? It is very important to me.

On Teamspeak 3 I had no problems with the stereo sound with the opus voice/music codecs (but the latency on LAN sucked ass which was why I wanted to change to mumble, the latency surely didn't let me down, but the mono sound made me kinda sad) I am forwarding sound from speakers to mumble, I need it to be output as stereo at least (hell i'd actually prefer 6 channel)

[mkrautz]

Mumble records speech and encodes to mono, but upmixes to stereo (or even more channels) on the receiving end when necessary (positional audio, etc.)

Why do you need stereo? Are you in a setup where you have users with more than one mic or a stereo mic setup? Or something else? Regular headset mics will only record to a single channel anyway, so I suppose your use case is special in some way?

Or are you using a music bot where the mono sound is a problem?

[mkrautz]

Ah, I actually *read* the whole post and saw your use case now. Sorry about that!

We don't currently support stereo or > 2 channel encoding. But the protocol can support it just fine. (Backwards compatibility is not guaranteed, though!)

There might be problems with multichannel recordings vs. UDP packet size, where try to keep under

the standard ~1500 MTU to avoid fragmenting packets.

However, for "special" use cases like yours, TCP would probably works just as well...

Is this something you're interested in seeing in Mumble?

[rabcor]

Yes it would be very helpful to me if Mumble supported this, Mumble is actually the only intuitive and low latency cross platform sound sharing application that fits the bill for what I need to do (which is yes streaming music from one computer to another, as well as other audio (e.g. I might be watching videos or playing games) I am essentially sharing one set of speakers with two (or more) computers) I know pulseaudio supposedly can do this, but pulseaudio and I do not agree with one another which is how I came to be here.

In this use case since I am hosting a server for this specific purpose, backwards compatibility would be a non-issue, and I have no qualms about using TCP over UDP either. The most important thing is that the latency and quality remains relatively the same and having true 2ch support. Because of the lack of stereo sound support I am right now using a hardware solution (e.g. line out to line in) but truth be told I was having better results with Mumble (less noise on a virtual sound card, noise is pretty bad with my hardware solution) and I would go back to it if I could have two channels.

My needs are met with just 2 channels since that provides me with acceptable sound quality even if it is upmixed to 6 channels, but mono upmixed to 2 (or 6) channels does not.

Packet size is also not a big issue since I'm doing this over LAN anyways, but even if I weren't, it'd be a fair price to pay for the added audio channel.