In a video editing apparatus (100) a first video data set describes a first scene in a first video. Among second video data sets describing second scenes contained in a plurality of second videos, third video data sets are identified that represent third scenes having the highest degree of similarity with the first scenes. Audio data sets associated with the third scenes are evaluated. Among a plurality of second audio data sets such audio data sets describing soundtracks having the highest degree of similarity with the first soundtracks may be identified. One of the third audio data sets may be combined with the first video data set to generate a media output data set where an audio track is added to the video data set in accordance with the preferences of the user.