Any audio or video, any length. Silence and music are skipped before transcription; doubtful passages are highlighted.