| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide an apparatus and method for highly precisely executing a detecting process of an utterance section from a sound signal. <P>SOLUTION: A sound processing apparatus includes: a direction estimation unit that detects direction points indicating a sound source direction of a sound signal for each of blocks divided by a predetermined time unit; and a direction tracking unit that connects the direction points to each other between the plurality of blocks and detects a section. The direction estimation unit generates: a dead angle pattern, or a correspondence data between a direction and a gain corresponding to a dead angle filter, which is a spatial filter with low sensitivity toward the sound source direction; and a directivity pattern, or a correspondence data between a direction and a gain corresponding to a spatial filter with high sensitivity toward the sound source direction; and detects a direction having a combination of a minimum part of the dead angle pattern and a maximum part of the directivity pattern as the direction points which indicate the sound source direction. <P>COPYRIGHT: (C)2012,JPO&INPIT |
