OBS is not designed for independent streams coming from one device. I get the strong impression that it was made originally for the stereotypical bedroom streamer, who only has the one mic and possibly headphones, and even that was botched in terms of the headphones being written by someone who understood software but not audio. (if it hiccups, expand the buffer...even after it's become uselessly out of sync) And despite many complaints, it still hasn't been fixed.
For the specific situation of a 2-channel device that is actually 2 independent sources, there's a workaround:
- Set OBS to stereo
- Create 2 copies of that source. Now you have 4 channels total, for those that count stereo as 2.
- In the Advanced Audio Properties, set both to Mono, and Pan/Balance them all the way to either side.
Now you have each of the 2 inputs controllable separately, panned center. That works (and ends up the way it does) because the Pan/Balance control is actually *before* the Mono switch, which is backwards from pretty much every professional thing. In this particular case, one non-professional thing (wacky processing order) partially cancels another non-professional thing (not choosing individual channels of a device), to get something that is at least partially usable.
If you want anything more than that specifically, you need to move your audio processing to an external DAW (Digital Audio Workstation: essentially a complete sound studio in one app), do everything in there, and bring the finished soundtrack into OBS as its only input, to pass through unchanged. A physical console with a USB line-in to the computer, works too, if you can get all of the signals in there and do what you need with them.
It helps a lot if you can control the DAW (or a physical digital console) from OBS, and for that, there's the
Advanced Scene Switcher plugin, which can send OSC (Open Sound Control) messages. Figure out how your DAW (or digital board) wants to receive OSC, set up Adv. SS for that, and now you don't have to switch apps all the time.
For a nice fade in one of my rigs, I ended up using the OSC messages to turn on and off a control signal, that is generated in the DAW, and goes to the sidechain input of several noise gates. When the control signal is on, the gate is open; when the control signal is off, the gate is closed (muted). The timing controls of each gate, produce the fade.