| 摘要 |
The present invention relates to a method and device at speech-to-speech translation. A given speech in a first language is recognized in a speech recognition equipment (A). The speech recognition equipment produces a text which is transferred to a translator (B) for translation into a second language. Parallel to these procedures, fundamental tone information is assembled for the first speech. The fundamental tone information influences the prosody generation (G), which influences a text-to-speech converter (C). From the text-to-speech converter a speech in a second language is obtained, the synthesis of which is essentially in accordance with the synthesis of the first speech.
|
