A conferencing system includes an input configured to receive N encoded speech signals from N terminals, and a signal processing arrangement configured to determine L encoded signals, of the N encoded speech signals, each indicative of an amount of sound that is louder than amounts of sound indicated by signals of the N encoded signals other than the L signals, the signal processing arrangement being further configured to produce at least N minus L sets of signals similar to the L signals and to transmit at least a set of the similar signals toward each of the terminals other than the terminals from which the L signals were received.