Splitting audio

In this document we will be setting up Virtual Audio Inputs and splitting audio outputs in OBS. Remember! This method is highly untested and it works very well on my PC at the moment. But if you have any issues, post it here.

qpwgraph

This is the app that will let us split audio easily with a GUI. You can download it by executing this in your terminal

sudo dnf install qpwgraph

We now need to setup qpwgraph such that it launches each time with the correct arguments on startup.
Copy the following contents into a file named qpwgraph.desktop and drop it in ~/.config/autostart/

[Desktop Entry]
Name=qpwgraph
Version=1.0
GenericName=PipeWire Graph/Patchbay
Comment=qpwgraph is a PipeWire graph Qt GUI interface
Exec=qpwgraph %f -axm
Icon=org.rncbc.qpwgraph
Categories=AudioVideo;Audio;Video;Midi;X-Alsa;X-PipeWire;Qt;
MimeType=application/x-qpwgraph-patchbay;
Keywords=PipeWire;MIDI;ALSA;JACK;Qt;
Terminal=false
Type=Application

We will be using this application later on to setup output routing.

NOTE: If you are using a window manager, make sure to add qpwgraph <path_to_config> -axm to your autostart script for your window manager by replacing <path_to_config> with the path of the config you save later on during the setup.
Eg: In i3, say a config is in ~/Speedrunning.qpwgraph, then we put exec qpwgraph ~/Speedrunning.qpwgraph -axm in ~/.config/i3/config.

Setting up Virtual Audio Devices

virtual-input-1 and virtual-input-2 can be renamed to make things more easily identifiable for yourself. But for the purpose of this guide, I'll be using those two names.

We're going to use .conf files to configure virtual devices. These files should be placed in ~/.config/pipewire/pipewire.conf.d (for user configuration) or /etc/pipewire/pipewire.conf.d (for system-wide configuration).

First, add a file virtual-input-1.conf with the following contents.

context.objects = [
	{	factory = adapter
		args = {
			factory.name = support.null-audio-sink
			node.name = "virtual-input-1"
			media.class = Audio/Sink
		}
	}
]

Next, add another file virtual-input-2.conf with the following contents.

context.objects = [
	{	factory = adapter
		args = {
			factory.name = support.null-audio-sink
			node.name = "virtual-input-2"
			media.class = Audio/Sink
		}
	}
]

This will create two virtual audio devices whenever pipewire service starts. If you want to have more virtual devices, you may add more to your liking. To take immediate effect, you can simply restart pipewire service with the command

systemctl --user daemon-reload && systemctl --user restart pipewire

qpwgraph should now show two nodes with the names virtual-input-1 and virtual-input-2.
Again, you can change out any instance of virtual-input-1 and virtual-input-2 to something else as they are just names. I just use them as it is to keep things consistent across the guide.

Setting up OBS

Now head to OBS and make a scene called Audio for the sake of simplicity. Inside the scene, add two JACK Input Client sources called virtual-input-1 and virtual-input-2. Just leave the properties as default. After adding the sources within the scene, add the scene to the scenes that you want audio in.

Setting up Audio Splits

Open up qpwgraph and you should see the nodes OBS Studio: virtual-input-1 and OBS Studio: virtual-input-2. Connect them to the virtual devices by dragging the monitor_FL and monitor_FR of virtual-input-1 to virtual-input-1:in_1 and virtual-input-1:in_2 of OBS Studio: virtual-input-1 respectively. Do the same for OBS Studio: virtual-input-2.

If you wanna not have your Spotify music appear on VODs, you can route Spotify into one of the virtual devices, like virtual-input-2, by dragging output_FL to playback_FL and output_FR to playback_FR.

Now on OBS, go into Settings, then to the Output tab. Under Streaming, tick the Twitch VOD Track box. If you're on simple output mode, it should default to 2. If you're in advanced output mode, pick 2.

Now go back to the your OBS scene and click on the gear icon under Audio Mixer. There, you'll deselect track number 2 of the audio that contains your Spotify music. In this case, that is virtual-input-2. So deselect track 2 of virtual-input-2 and now your Spotify music will not be inside your Twitch VODs.

Similarly, route any other audio that you want into your virtual devices, i.e. your Minecraft instances or lecture videos.

Now hit Ctrl + S to save changes in qpwgraph.

If you have any issues regarding Spotify desktop app connections flickering from one virtual device to the other, you may post an issue here.

All done!

We are done with setting up audio splitting on Linux.
Again, this is very dependent on your hardware and setup and this is highly untested. If you get any issues while setting this up, post an issue here or even better, join the resetti Discord server and we'll try and help you.