| 摘要 |
PROBLEM TO BE SOLVED: To provide a speech transmission apparatus for creating a transmission speech signal without cut of a head part of a speaker's speech, which gives no sense of incongruity to a listener at a transmission destination, by recovering difference between sound collection timing and transmission timing. SOLUTION: A sound collection signal control section 11 buffers a sound collection signal and outputs it to a transmission speech signal creation section 130 according to output directions of the transmission control section 13. A sound collection signal analysis section 12 acquires pitch information and utterance period information by detecting pitch/utterance of the sound collection signal for each predetermined unit, and forwards them to the transmission control section 13. When it is detected that the same pitch continues according to the pitch information, the transmission control section 13 performs thinning processing of the continuous pitch, and creates and outputs a transmission speech signal which is shorter in time length than the original sound collection signal. The thinning processing is continuously performed so that the transmission speech signal may not be cut in the middle, and it is stopped when the sound collection timing matches the transmission timing, and the sound collection signal is output as it is, as the transmission speech signal. COPYRIGHT: (C)2009,JPO&INPIT
|
