[Snap] Give add-ons access to audio

[benfrancis]

Some add-ons need to record and/or play back audio.

This may require the pulseaudio interface.

[dilyn-corner]

I'd strongly recommend trying to keep the audio stack as simple as possible -- audio-{playback,record} interfaces may be sufficient for you, especially in this context. For instance, are you using pipewire at all? If not, I'd recommend avoiding the pulseaudio interface. Using alsa directly keeps things simple. Of course, if you truly need the features pulseaudio provides that's another story.

[benfrancis]

The gateway application itself doesn't use audio*, but various add-ons do. E.g. the voice-addon depends on the the mic, sound-player and deepspeech npm packages and the voco add-on is another popular example.

I will need to work with the community to figure out all of the dependencies of audio add-ons and what work might be needed to get them working on the Ubuntu Core based distribution. These add-ons are very popular with some community members, but I would ideally like to encourage human-computer interaction to move to the separate WebThings Shell package in the 2.0 timeframe, so for me personally this isn't a big priority.

* There used to be a built-in smart assistant experiment but it was split out as a separate add-on.