摘要 |
PROBLEM TO BE SOLVED: To provide a speech processing device capable of estimating a speech section of a received speech without depending on ambient noise.SOLUTION: A speech processing device comprises: an acquisition unit for acquiring a transmitted speech; a detection unit for detecting a first speech section of the transmitted speech; and a calculation unit for calculating a first feature amount including section length of the first speech section, or the number of vowel sound in the first speech section. The speech processing device further comprises a determination unit for determining frequency of appearance of a second feature amount in which the first feature amount is less than a first threshold, in the transmitted speech, and an estimation unit for estimating a speech time of the received speech on the basis of the frequency. |