Set different output devices per user

[lukasbestle]

I use Mumble for podcasting (as an awesome open, powerful and high quality Skype alternative). In the end, I'd like to have the recordings of the other speakers (Mumble users) in different channels in my recording software.

Right now, I use multiple Mumble clients on the same local machine (my recording machine) and locally mute all speakers – except the one I want to record with the client instance – in the clients. Every client instance then only outputs a specific speaker's voice and I can send the outputs to different local audio devices for processing.

That definitely works, but it is very hacky and complicated and lengthy to set up each time (the instances share their configuration, so every time I open a new one I have to change the output device again).

What I propose is a setting in the speaker information window that lets you define a custom output device just for that speaker. By default, this would always be the one set in the global preferences, but for use-cases like mine you could easily configure it to be whatever you need. Maybe this could also be hidden if the "Advanced" checkbox in the preferences is not checked.

[ubuntuaddicted]

In linux, mumble already has multichannel recording within the client. No need to open more than 1 instance of mumble.

[ubuntuaddicted]

Sorry about a second post (I can't edit I'm on a mobile device) but maybe I misunderstood your goal. You want to have each persons audio be processed differently LIVE? On Linux Tech and Gaming podcast we use mumble as well. We just each have our audio fine tuned for our unique environments and everything just works. I'm the hub, so they call me and I encode and stream that.

[lukasbestle]

Thanks for your quick reply.

Recording inside Mumble to multiple tracks works, yes. We used that in a situation in which my internal software audio routing using the OS X kernel extension "Soundflower" didn't work. It's an awesome feature if you simply want to record without setting up other software.

What I want is streaming to multiple audio output devices. After the podcast recording is over, I want to have my audio tracks separate from each other to be able to remove coughs and other background noise (sometimes, someone forgets to mute/not unmute himself). I could use one track for live streaming and then import my Mumble recordings afterwards, but that just doubles the work, so it would be great to have multiple tracks in the DAW in the first place.

So what I need in Mumble is a setting which lets me define multiple live audio outputs (one per speaker) in a single instance of my Mumble client.

[ubuntuaddicted]

Now I understand your needs & situation better. That would be a desirable feature in mumble. If I were 1 of the devs I would make it happen but sadly I'm not. I'm just a user like you. I'll keep my eye on this as its something Linux Tech and Gaming show may utilize as well.

[lukasbestle]

Great that there's another one who might need this feature. Can anyone of the devs comment on this?

[mkrautz]

@ubuntuaddicted I'm guessing that Linux Tech and Gaming Show would need this to work with PulseAudio for it to be useful?

@lukasbestle What would you need this to work with? I gather that you're on OS X. Can you explain what your setup with Soundflower is?

[lukasbestle]

I am on OS X, exactly. But I guess that this can be applied to other operating systems and setups as well.

What we want to achieve is the following routing setup (using a custom Soundflower installation with additional audio devices):

• I use an audio interface that is connected via USB. That audio goes into my DAW, together with another track for jingles and other sound effects.
• The DAW then applies my effects (compressor, EQ) and sends the audio out to a Soundflower device (for people on other operating systems, that could be done by using an interface that supports loopback or a similar software that does the same as Soundflower; maybe something like JACK could work, but that requires a totally different audio API as far as I can tell).
• I set the microphone input in Mumble to that Soundflower device. Now, the other speakers can hear my voice and the jingles.
• So far, so good. Up to now, this all works with the current Mumble release. What's interesting is how to get the other tracks into my DAW. For that I'd like to use additional Soundflower devices (one per speaker) and set them as outputs in Mumble. I can then access those devices from my DAW, record them into separate tracks and send everything out to a live streaming server if I want (using another Soundflower device).

The only requirement for this setup is the possibility to define multiple audio outputs in Mumble. It doesn't even have to be super-special, as Soundflower devices appear just like any other pair of headphones or internal speakers in applications like Mumble.

I heard on Twitter that Mumble currently creates a mix before doing any output stuff, so this might require a rewrite so that you can grab the track audio before it gets mixed.

[lukasbestle]

@mkrautz: Are there any news on this?

[Rapturoso]

I have been following this for a long time. Surely there's some movement and news on this desperately needed functionality? I know of at least 60 or so producers of live streams and videos that would absolutely find this functionality a massive time saver. Output to different Windows / Pulse audio driver devices or virtual audio inputs on DAWs would be great.

[Krzmbrzl]

To give a (very late) answer to your question: AFAIK there's still noone working on this feature. And I don't think that it'll receive much attention from the core-devs. There are simply way too many other issues in the queue that affect more users.

We'd be very happy to accept PullRequests though :)

[hagmgoe]

Hello

Another attempt to see how far now is the right time to ask the feature. I would be very pleased. :-)

[Krzmbrzl]

Afaik my comment above is still valid

[Heljick]

Still no move from the devs about this feature ? Will be a big add in 2021 with all the streams we are doing those days

[Krzmbrzl]

No nothing new yet. Nobody has found to the time to work on this yet.