Critical bandwidth allocation techniques for stored video delivery across best-effort networks

Abstract

We propose two new techniques for the delivery of compressed prerecorded video streams across best-effort networks like the Internet. Current approaches for the delivery of stored video across best-effort networks typically alter the quality of the video frames, the frame rate delivered to the user, or a combination of both. By using network feedback, these algorithms continually adjust the video quality to fit within the available network resources. These approaches, however do not take advantage of the a priori information available from stored video streams, namely the frame sizes that the movie consists of. We show how monitoring the a priori information and actively monitoring a client-side buffer can help smooth the video frame rate delivered to the user, providing a more consistent quality of video.