Prediction Methods for MPEG-4 and H.264 Video Transmission

Video services became a large part of internet network traffic. Therefore understanding of video coding standards and video traffic sources, such as video trace files is highly important. In this paper we concentrate on the basic characteristics of mpeg-4 and h.264 video coding standards. We describe the concept of the i, p and b frames in these standards since they are the main feature of every video trace file. Then we describe the content of the video trace files since the trace files are important for researchers to investigate network performance and understanding of network features. These are the important issues in terms of assuring the quality of service (QoS) in multimedia applications spread across the internet. Traffic prediction and bandwidth allocation are the crucial parts in terms of QoS. In this kind of applications, the artificial neural networks are vastly used. Therefore we illustrate the results of neural networks for video traffic prediction using both mpeg-4 and h.264 trace files.