For a few years now, discord doesn't need you to send an update when you stop speaking. This is now handled entirely client-side. The client automatically detects when the stream of datagrams stops, and sets the speaking flag off locally.
This solves the problem of tracks ending early because of buffering, which is caused by early speaking stop updates. Instead, only send these updates when playback starts or the mask changes (priority/soundshare/voice), but never when the playback stops.
For a few years now, discord doesn't need you to send an update when you stop speaking. This is now handled entirely client-side. The client automatically detects when the stream of datagrams stops, and sets the speaking flag off locally.
This solves the problem of tracks ending early because of buffering, which is caused by early speaking stop updates. Instead, only send these updates when playback starts or the mask changes (priority/soundshare/voice), but never when the playback stops.