Building Digital Video Editing System Based on Dynamic Synchronous Transfer Mode Network

Abstract

For the compilation of the edited video stream, normally at least two others are manipulated (A/B roll). With conventional “linear editing”, video/audio sequences are stored on the magnetic media with compilation via replication in real time. This causes some limitations in flexibility, since final compilation must be in the correct chronological sequence and is constrained by the serial linear tape access time. “Non-linear editing” systems require all clip/sequence data to be stored on the random access medium, which affords play-out and manipulation/compilation abilities. Digital formats afford the opportunity for non-linear editing but certain constraints must be considered. Volume of data associated with audio and video streams is very large and requires high capacity storage. While adequate storage is often available, the communications between the workstations and access to the storage becomes a bottleneck. Paper is devoted to investigation of the existing experience of building Distributed Digital Video Editing Systems and use in them widely available computing platforms such as PCs in order to apply this for the project DIVIDEND (Dlgital VIDeo Editing based on the Network testbeD). Project DIVIDEND addresses the demand of the media industry and video producers in particular, for non-linear editing facilities which are currently unavailable. It intends to build an experimental testbed based on high speed network technology called Dynamic synchronous Transfer Mode (DTM). System architecture and specifications characteristics are described.