Abstract: Methods, systems and computer program products for performing voice conferencing over a data network, such as an internet protocol (IP) network are provided. The conferencing is for use in environments including N incoming channels and N outgoing channels. Each of the N incoming channels is associated with a corresponding one of the N outgoing channels, where N≧3. A different audio packet is received over each of the N incoming channels. The energy level of each of the different audio packets is determined so that a first highest energy packet and second highest energy packet can be identified. Also identified are the incoming channels over which the first highest and second highest energy packets are received. Next, the highest energy packet is sent to each of the N outgoing channels except an outgoing channel associated with incoming channel over which the highest energy packet was received.