Whereas the polysemous term media is unambiguous within a single domain, multiple ambiguous interpretations can now be found in discussions of electronic convergent mass and communication media. The different interpretations of media in multimedia can be broadly classified as convergent when considering the end product, such as a news web site, or divergent when considering discrete elements, such as text, audio and video. The latter interpretation is commonly used in higher education multimedia applications. This paper revisits theories of multimedia in learning with reference to current standards for learning object metadata to produce a simple model of divergent or discrete media in multimedia. The model clearly distinguishes media form from content, technology and domain with the aim of reducing ambiguity.