VOICE TRANSLATION SYSTEM, METHOD OF VOICE TRANSLATION AND PROGRAM

摘要

PROBLEM TO BE SOLVED: To provide a voice translation system capable of preventing collision of speeches between a user A and a user B and which allows the user A to check whether the information contained in a speech made by the user A is properly translated to the user B.SOLUTION: The voice translation system includes: a first voice unit; a second voice unit; and a server that performs voice translation between the first voice unit and the second voice unit. The server includes: voice recognition means that performs voice recognition on a voice signal of a first language as an input; first machine translation means that translates from the recognized first language into a second language; second machine translation means that inversely translates the translated second language into the first language; first voice synthesis means that generates a first voice synthesis signal using the inverted translation result as an input; and second voice synthesis means that generates a second voice synthesis signal using the translation result as an input. Voice time length measurement means measures the second voice synthesis signal or a time length of the first voice synthesis signal, and outputs a voice while synchronizing the start time and termination time between the first voice and the second voice according to the time length.