Modeling 3D video traffic using a Markov modulated gamma process

Abstract

Three-dimensional (3D) television and video streaming has become very popular over the last few years. Multiview encoded video is used to support 3D video applications. The statistical characteristics of multiview video are significantly different from the traditional single-view video and therefore existing video traffic models are no more applicable for this type of video. In this paper, we propose and evaluate a new model for multiview video that is based on a Markov process. To the best of our knowledge, there is only one other model for multiview video that has been proposed in the literature by Rossi et al. We compared the two models using Q-Q plots and the autocorrelation function of the frame sizes along with QoS metrics of the resulting packet traces estimated by simulation. The comparison results show that our model has less complexity and better accuracy.