OK good to know. I’ll see if I can learn anything more from your Unity project. I don’t have much hope of fixing the issue if I can’t reproduce it, though. Something I noticed is that the test scene is using a different volume camera window configuration from the default. This is something we need to make more clear, but the default volume camera output configuration determines the “startup” configuration for your app. So when the default configuration in settings doesn’t match the one in the startup scene, we end up opening and closing a volume with the default configuration in quick succession when the app is opened. You don’t ever see this because the volumes fade in and out, but I wonder if that first volume could be “stealing” the audio and causing issues.
Can you try setting the default volume configuration in Project Settings > PolySpatial > Default Volume Camera Window config to the same asset that you use in the volume camera for the startup scene?
Also, unrelated to this audio issue, I noticed that your scene has an ARSession and you’ve set a hand tracking usage description. ARKit is not actually required for the gaze/pinch gesture to work, and furthermore it won’t work when you don’t have an ImmersiveSpace open (enabled in Unity by using an Unbounded volume camera). I don’t think this would affect the audio issue, but you should be able to remove the AR Session from your scene, disable Initialize Hand Tracking On Startup, and remove the usage description. In fact, you can even de-activate the Apple visionOS AR plugin/loader if you don’t intend to use ARKit features.