| 摘要 |
A bandwidth efficient digital voice communication system ( 10 ) can include a speech-to-text converter ( 22 ) for converting a voice signal to a text representation, a speech parameter extractor ( 28 ) for extracting user identifiable parameters from a voice signal, and a text-to-speech converter ( 44 ) for converting the text representation and the user identifiable parameters from the voice signal in real time. The user identifiable parameters can be among pitch rate, gain, fundamental frequency, and formants. The speech-to-text converter can further include a text to phoneme segmentation module ( 24 ) providing flags to indicate a plurality of phoneme boundaries, an automatic phonetic segmentation module ( 26 ) coupled to the text to phoneme segmentation module and the speech parameter extractor, and a voice parameter manager ( 29 ) for receiving the text representation, the flags indicative of the plurality of phoneme boundaries, and the user identifiable parameters from the voice signal.
|
