It is made possible for a DTMF tone signal associated with a button depressed on an IP telephone terminal to arrive at the opposite party even if an IP telephone terminal does not have a function of sending a DTMF tone signal. If a DTMF tone signal transmission request signal detection unit in a proxy detects a DTMF tone signal transmission request signal inserted in an RTP packet or a SIP packet received from an IP telephone terminal or detects a DTMF digit signal contained in the DTMF tone signal transmission request signal, a DTMF tone signal generation unit generates a DTMF tone signal associated with the DTMF digit signal. A voice encoder encodes the DTMF tone signal in the same way as a voice signal. An RTP payload replacement unit replaces a voice signal in a payload of the RTP packet with the encoded DTMF tone signal.