| 摘要 |
During a speech recognition process, it is determined if an error has occurred, and a speech parameter is analysed if this is the case. If the speech parameter is below or above a threshold value (S), a message (OUT) is output to the user, individualised to maintain a range of values for the speech parameter, requesting that he provides a new input with a correction made to the speech parameter. The speech parameter may be the signal-to-noise ratio of the input speech signal, the keeping within a range of frequencies, the speaking speed, or loudness. A telecommunications system is also claimed.
|
