Mark main segment and split it into vocal chunks
Choose an audio track, set song start/end, and extract the main segment.