Remove silences from source audio

[thestumonkey]

We are storing the raw audio for playing, but there's lots of silence in it. Using VAD, we can remove a lot of the silences and reduce load on the main transcrption and recongition services

[linear]

FRI-25 Remove silences from source audio

[AnkushMalaker]

This was the default behaviour earlier, but now I don't see that much use. Feel free to open with some more info
Default storage is opus, and models handle silence innately now a days