Separate graphs with different Input and Output devices do not retain settings when switching.

[ThatGuyCND]

Describe the bug When switching between graphs — one Session with two separate graphs, each with their own input and output devices — the input and output devices are not kept separate.

Is this a PEBKAC issue? Am I misunderstanding the active sessions? Is only the active graph meant to be processing rather than all graphs in the active session?

To Reproduce Steps to reproduce the behavior:

1. In the first graph:
   • set input device to XLR capture device (Zoom AMS-44)
   • set output to Virtual Audio Driver (Blackhole 2ch)
   • Ensure that the output device is set as the System Input Device
2. In the second graph:
   • set input device to Virtual Audio Driver (VB Cable)
   • set output device to DAC (FiiO K9)
   • Ensure that the input device is set as the System Output Device
3. With audio playing while the second graph is selected, note that the first graph is not processing the mic input correctly (this can be validated via a virtual meeting). Click on the first graph in the active session, note that the nodes for input and output device labels have not changed, but expanding the Node section will show you that they are in fact set incorrectly and are showing the devices for the second graph. You can also hear that the audio processing updated to use the interstitial nodes defined for that graph.

Expected behavior Discreet inputs and outputs do not change when switching between graphs. NOTE: After fixing the input and output nodes in graph 1, switching back to graph 2 displays the same issue. Fixing the nodes in graph 2 now processes graph 1's input device through the plugins in graph 2, BUT the mic is still output to the correct virtual device.

Screenshots Graph 1:

Graph 2:

Noted Error:

Desktop (please complete the following information):

• OS: MacOS 12.7.3 — Apple M1
• Element Version: 0.46.6

Standalone or Plugin uh ... both? IDK how to answer this.

Additional context I have not setup any aggregate devices, and trying to keep my mic processing separate from my audio processing, I believe that what I've setup is correct, BUT coming from AULab (which doesn't work in modern MacOS 😭 😞 ) where I didn't have to previously setup any aggregate devices, I'm willing to adapt, change, update my Element configuration to use the software in a better way. Maybe I just hit an edge case?

Audio Devices:

[mfisher31]

The audio device applies to the entire session, not on a per-graph basis