| 摘要 |
PROBLEM TO BE SOLVED: To provide a segment boundary matching system for making segment boundaries of voices between two systems be matched with each other. SOLUTION: The voice signal of a system A is guided to a featured value extraction part 12 and a feature point detection part 13 via a delay adjusting part 11, the feature value (a) of each segment is extracted by the featured value extraction part 12, and a feature point (b) is detected by the feature point detection part 13. Also, the voice signal of a system B is guided to a featured value extraction part 14 and a feature point detection part 15, a featured value (c) of each segment is extracted by the featured value extraction part 13, and a feature point (d) is detected by the feature point detection part 15. The featured values (a), (c) and the feature points (b), (d) are sent to a monitor terminal 6. The monitor terminal 6 calculates the deviation of the segment boundaries between the two systems on the basis of relative positions from the segment boundary of the featured values (b), (d) of both systems, and makes the segment boundaries be matched with each other by correcting the deviation. COPYRIGHT: (C)2007,JPO&INPIT
|
