Providing definition and temporal structure for multimedia data

Currently, multimedia (MM) applications are prepared in an ad hoc fashion. Images or video are usually produced by the MM author. The MM application itself is composed with a specific combination of a hardware platform and authoring software, and it typically contains its own specific data. Today, MM data are usually associated with a particular platform or application, but future MM developers will use repositories of reusable data shared among various applications and across various platforms. The layered multimedia data model (LMDM) emphasisizes the sharing of data and components by dividing the process of MM application development into smaller, more manageable pieces. In particular, the LMDM emphasizes the conceptual separation, manipulation, and presentation of data. When combined with a collection of MM data, the LMDM enables description of a wide variety of MM compositions. This paper describes the first two layers of the LMDM, those concerned with data definition and with the specification of temporal structures that can be reused by various MM presentations. Several examples demonstrate the advantages of the layered paradigm: modular, reusable components and isolation of platform dependencies.