Coded-domain multipoint video bridging

Multipoint video conferencing is a natural extension of point-to-point video conferencing. Multipoint video conferencing usually includes a multipoint video bridge which combines the video signals from the multiple conference participants to provide a single output video signal which can be displayed to and shared by all the participants. The video signals are typically received and transmitted over a network in a compressed digital format. Video bridging can be done in the pixel-domain or in the coded-domain. The coded domain video bridging has the advantages of being able to achieve lower end-to-end delay and lower system cost. However, to make the coded domain video bridging practical, several technical issues need to be addressed. These technical issues include frame-synchronization of the multiple input video signals, multiplexing and combining of the video signals, and possible video quality degradations due to the combining operations. In this paper, we discuss these technical issues for a coded-domain multipoint video bridge based on Group-Of-Block (GOB) video units in the H.261 video standard. We demonstrate the feasibility of a software-based coded-domain video bridge with a fully functional prototype built around a low-cost personal computer. We also show simulation results concerning the quality degradation issue