It is unlikely that multiple formats for short timecodes would be used in one document. Therefor we will parse all the short timecodes to see if they are all less then the duration as HH:MM. If they are we will use that, otherwise we will parse them as MM:SS.