Building temporal structures in a layered multimedia data model

Abstract

The Lagered Multimedia Data Model (LdfDhf) aib in the qwcification of multimedia compositions by dividing the problem into smaller, more manageable pieces. In this paper we describe the lower two la~ert of the LMDM, the Data Definition Layer, which allows the qwcification of multimedia objects in a databaae, and the Data Manipulation Layer, which allows the~pecification of temporal~tructurw built from thoue objects. Several ezampleg demondrute the advantage of the layered paradigm: limple~pecificationa, and modular, rvwabie component. Background and MotivationMultimedia (MM) is not only presentation-driven, it is also data-driven. Although one typically associates multimedia with graphics and sound, MM data may have relevance independent of its prmrttation, such as high volume photo data fkom satellites, much of which is never viewed. Similarly, MM data is eomtimes manipulated without any presentation occurring, as with digital audio mixing. In addition, a eingle data object may have several presentations, such as a video with various CO1OM or resolutions. These issues lead to the conclusion that a MM model must describe both MM data and its presentation, while recognizing the conceptual diference between them.In a reeent paper we introduced a new model for specification of MM data and MM compositions, which follows the layered architecture paradigm [10]. The strength of the Layered Mdtimedia Data Model (LMDM) is its emphasis on sharing, portability, and reueabfity of MM data and abstractions. Figure 1, follows the layered model paradigm. Such an approach, in addition to making a clear separation between data management, presentation management, and user-interaction, has several other advantagea. The LMDM, shown inFirst, hardwar-and OSspecific concernscan be isolated in particular layers, remdting in a more general model. Second, components defined in one layer can be used in conjunction with diferent sets of components ikom other layers, improving their reusability y. Finally, the layered paradigm has long been accepted in the cornmunie tiona community, and multimedia may be viewed as a novel method of communication.A layered MM model was previously proposed in [7], although the choice of layers was somewhat difFerent. ScriptX, the devic~independent MM scripting language currently under development [6], was also designed around a layered model the general structure of which is sirn.ilar to the LMDM. Data Definition LayerThe Data Definition Layer (DDL) allows data specification of a MM data object clae~, and it provides the abstraction of data objects for the higher layers. Objects can include persistent data, instructions to generate data, or a reference to a run-time data generator. Database management tools for accessing and querying persistent data are assumed to be in place. Objects contain metadata describing data format and semantics. Operators are provided for data transfor- mations. A detailed description of the DDL is pro- 271

Hypervideos and Interactive Multimedia Presentations

“…NCL 2.0 contained 21 modules from 11 functional areas [Silva et al 2004 -Further Multimedia Presentation Models and Languages: Further multimedia presentation/document models and languages are described by Adali et al [1999Adali et al [ , 2000, Adiba and Zechinelli-Martini [1999], Assimakopoulos [1999], Deng et al [2002a], and Scherp and Boll [2005]. Further models are ZYX [Boll et al 1999[Boll et al , 2000Boll and Klas 2001], the Layered Multimedia Data Model (LMDM) [Schloss and Wynblatt 1994]; Madeus [Layaida and Sabry-Ismail 1996], and MPGS [Bertino et al 2000]. Interchange formats are the CWI Multimedia Interchange Format (CMIF) [Bulterman et al 1991] and the Procedural Markup Language (PML) [Ram et al 1999].…”

Modeling video objects in 4DIS temporal database system

“…in a way similar to the querying of conventional database systems on literal data [2, 5, 12. 14], Second, we would like to be able to specie" the temporal chaxac-teristics of video objects so that temporal requirements of various presentation operators could be enforced and supported [7,9]. The focus of this paper is on the latter issue which has been receiving less attention.…”

“…In [9], an event calculus is proposed to describe the relationships and synchronization among several multimedia objects. Similarly, in [7], seven temporal relationships among multiple multimedia objects are defined: before, meets, overlaps, during, starts, finishes, and equals.…”

Extending databases to support image editing

“…Ideas similar to our work can be found in [3] and [8]. In [3], multimedia objects are treated as binary large objects (BLOBs).…”

“…For example, there is no way to describe the dithering needed to double the size of an image. In [8], a notion similar to transforms is presented. There, scripts are passed to image editing tools in order to create new images.…”