A method and apparatus for splitting the audio of media content into separate content files without introducing boundary artifacts is described. A computing system for streaming media content including audio and video comprises a database to store a plurality of content files and a server coupled to the database to deliver the plurality of content files over one or more network connections. Each of the plurality of content files comprises an encoded portion of the video encoded according to a frame rate and having a fixed-time duration and an encoded portion of the audio encoded according to a codec-enforced frame size and having a plurality of full audio frames having the codec-enforced frame size. A duration of the encoded portion of the audio of one or more of the plurality of content files is greater than or less than the fixed-time duration.