Adaptive Video Quality Throttling Based on Network Bandwidth for Virtual Classroom Systems

Abstract

Current e-Learning solutions enable viewing and interaction with participants who are geographically distant. However, these systems often are prone to delay, jitter, and packet loss owing to network fluctuation. The dynamic nature of bandwidth congestion requires us to dynamically adapt the quality of video streaming. In this paper, we attempt to alter the quality of streaming on the fly, by monitoring packet loss due to network congestion. To achieve this, we setup a simplistic classroom architecture consisting of one local and two remote classrooms with the camera feeds of the participants displayed in each other’s location. The camera feeds are flagged in accordance to their predetermined priority level, and the flagging is altered in accordance to the dynamic interactions in the classroom. In the advent of network congestion, the packet loss of the recipient is monitored, and the capture resolutions of all the other feeds are altered dynamically, to make allowance for the prominent streams to remain unaffected. The effectiveness of the system is measured using participant feedback. Results indicate that the participants do not feel any perceivable change in the quality of the multimedia content presented to them.