Add the start and end time associated with media fragments to the RDF output

[yunjiali]

We can use Ninsuan ontology
The start and end time must be there in order to calculate the whole length of the media fragments