Audio stream management techniques are described. In an implementation, a unique representation is generated by a client for each of a plurality of audio streams associated with television content received at the client. Each of the unique representations is displayed in a user interface at the client to be selectable by a user to cause output of a respective audio stream with the television content.