Client-Side Echo Cancellation for Multi-Party Audio Conferencing

August 16, 2006
Junlin Li | Intern, Communication and Collaboration Systems Group, MSR

We introduce a new structure for multi-party audio conferencing, where the MCU (Multipoint Conferencing Unit) mixes all the received signals and rebroadcast the same stream to all clients. Each client then removes its own signal before playback. This architecture reduces bandwidth and computation requirements at the MCU. The major challenge, however, is that the client’s own speech is non-linearly distorted at the MCU. This makes it hard to remove from the mixed signal, producing an echo-like distortion. In this work, we solve that problem by proposing a client-side echo cancellation algorithm which completely removes the echo, with minimal artifacts. The Mean Opinion Score (MOS) results imply that the proposed algorithm works well, which makes the proposed multi-party audio conferencing structure practical and promising.

Speaker Details

Junlin Li is a third year PhD student of the School of Electrical and Computer Engineering in Georgia Institute of Technology. He completed his undergraduate and master in Electric Engineering from Tsinghua University, China. He has worked on speech coding, 3D graphic coding and processing. Junlin currently research focus is on distributed and collaborative signal processing in wireless sensor networks. This summer, he is working with Li-wei He and Dinei Florencio of Communication and Collaboration Systems group at MSR, on client-side echo cancellation for multi-party audio conferencing systems.