Text this: Extending automatic transcripts in a unified data representation towards a prosodic-based metadata annotation and evaluation