摘要 |
PROBLEM TO BE SOLVED: To provide a method for automatic high degree voice recognition without a remarkable change to a secondary channel and a present standard by extracting plural voice characteristic signals from a received coding voice signal separately converging the received coding voice signal to an audio voice signal and applying them to a voice recognition system. SOLUTION: A hand set 201 generates a voice coder/parameter to transmit it to a radio base station 220 in response to a user input by utterance. In the radio base station 220, the received coding voice signal is provided on a path to be supplied to a public switch 230. Further, the coding voice signal received by the radio base station 220 is supplied to an automatic voice/speaker recognition (ASR) characteristic system 235 and an ASR system 240 arranged in the radio base station 220. In these systems, the decoded same coding display generating the voice signal is used as a substitute, and an ASR characteristic parameter is extracted with the ASR characteristic system 235 that the ASR system 240 uses. |