To provide a multi-point conference system for acquiring a delay time in data transmission/reception and performing echo canceling by taking it into account so as to enhance the quality of voice data that are multicast.

A terminal 1 transmits a synchronous packet part of which is rewritten. A data mixing apparatus 20M puts together only identification information denoting a sender in the synchronous packet with a mixing packet. The terminal 1 for receiving the mixing packet uses a mixing packet inspection means 9 to discriminate establishment of synchronization at a point of time when discriminating the continuous acquisition of packets including its own sender information. When the synchronization is established, since a delay time for data transmission/reception is obtained, the voice quality of a conference can be enhanced by adding the delay time to the echo cancel processing.