MAJOR UPDATE
Each speaker is separated as the transcript builds, so you don't replay a section to work out who was talking.