Maybe the audio pausing 'feature' was to reduce disk IO and CPU usage whilst recording video. I imagine trying to process a FLAC file or similar high bitrate stream would have a somewhat adverse reaction on video performance. If you try to add 720p it may be worth enabling for specific modes.

I noted a few pages back a post relating to the "recorded from device" tag. They mentioned how it is generated from concat device-id, random string and hashing (somewhat more detailed than what I put here). Haven't heard anything from you Nicolai on this. Does it solve the issue?