Real-Time Adaptive Content-Based Synchronization of Multimedia Streams

Abstract

Traditional synchronization schemes of multimedia applications are based on temporal relationships between inter- and intrastreams. These schemes do not provide good synchronization in the presence of random delay. As a solution, this paper proposes an adaptive content-based synchronization scheme that synchronizes multimedia streams by accounting for content in addition to time. This approach to synchronization is based on the fact that having two streams sampled close in time does not always imply that these streams are close in content. The proposed scheme primary contribution is the synchronization of audio and video streams based on content. The secondary contribution is adapting the frame rate based on content decisions. Testing adaptive content-based and adaptive time-based synchronization algorithms remotely between the American University of Beirut and Michigan State University showed that the proposed method outperforms the traditional synchronization method. Objective and subjective assessment of the received video and audio quality demonstrated that the content-based scheme provides better synchronization and overall quality of multimedia streams. Although demonstrated using a video conference application, the method can be applied to any multimedia streams including nontraditional ones referred to as supermedia like control signals, haptic, and other sensory measurements. In addition, the method can be applied to synchronize more than two streams simultaneously.