| 摘要 |
PROBLEM TO BE SOLVED: To suitably improve input environment by a user oneself, according to presented information, by presenting not only a magnitude of an input speech signal, but also quality information, when the speech signal is input to various devices or the like. SOLUTION: A gain of the speech signal inputted from a speech signal input section 101 is adjusted to an appropriate level in an input gain adjusting section 102, and the signal is converted into a digital signal by an A/D converter section 103. A first signal level is calculated, for each frame of the digital signal, which is cut out for every fixed period, in a first signal level calculating section 104; and a speech likelihood value is calculated from a feature of the digital signal calculated by a spectrum feature calculating section 105, and a probability model (Gaussian Mixture Model) parameter of a clean speech signal. Then, the first signal level and the speech likelihood value of the same frame are displayed on a display section 108. COPYRIGHT: (C)2009,JPO&INPIT
|
