This chapter presents the design and optimization of multiparty audiovisual conferencing systems, which are multi-metric, multi-control applications studied in Chap. 5 . These systems operate on the principle that only one participant speaks at a time to avoid the “double talk” phenomenon. We survey six representative proprietary systems by examining their architectures, capabilities, and features. These were built on the speed and reliability of today’s Internet, enabling many concurrent users to participate in a shared session. Most rely on a global network of servers to manage and forward packets in a centralized or distributed fashion. Next, we study the optimization of perceptual quality in a v-party audiovisual conferencing application. Instead of solving it directly as a multi-metric, multi-control problem, we consider a conversation in each conversational turn to consist of \(v-1\) speaker-listener pairs. In each pair, the listener optimizes the \(\mathrm {MED}\) to achieve the best ASQ-INT tradeoff, independent of other listeners. We also found that fluctuations in the silence periods that each listener waits for from the speaker affect the perceptual quality for the \(v-1\) listeners in a conversational turn. To reduce these fluctuations, we utilize playout scheduling algorithms to equalize the silence periods while waiting for the speech segment from the speaker.