| 摘要 |
PURPOSE: An apparatus for converting voices in a mobile communication system is provided to convert a source(subscriber) signal into a target signal to transmit the converted signal, so as to supply a communication service to a source(subscriber) having an abnormal pronunciation and to generate a synchronized sound using a self voice. CONSTITUTION: If a source(subscriber) requests a signal(voice) conversion to an exchange station, to convert a self signal into a target signal, the exchange station receives voice data of the source, and makes a rule to map the voice data with the target signal. A voice converting device of the exchange station transmits the target signal and a source signal to voice analyzers(201,202), respectively. The voice analyzers(201,202) analyze the source signal and the target signal to extract acoustic featuring parameters relating to the extracted signals. A voice converter(203) generates a vector space of a finite degree relating to the signals, according to degrees of LPC(Linear Predictive Coding) cepstral coefficients extracted through the voice analyzers(201,202). The voice converter(203) displays the LPC cepstral coefficients on the vector space, and finds an LPC cepstral coefficient converted when a mapping is performed between an LPC cepstral coefficient relating to the source signal and an LPC cepstral coefficient relating to the target signal. A voice synthesizer(204) synthesizes the source signal with the target signal to perform a voice conversion, by using the LPC cepstral coefficients extracted from the voice analyzers(201,202) and the LPC cepstral coefficient from the voice converter(203).
|
