Audio generates its own synchronization data — the closed loop between speech synthesis and self-transcription.
audio.currentTime reaches each word's start. No pre-existing alignment. No human timing. The audio creates its own sync data. Family 1 claim scope.