| 摘要 |
PURPOSE:To simplify the constitution of hardware, to avoid the reduction in transmission speed, and to suppress the unnecessary degradation in S/N by using a pitch-forecasted pulse train converting means. CONSTITUTION:A decoded multipulse train and an encoded maximum amplitude are inputted to a pitch extracting and V (voiced)/UV (unvoiced) discriminating device 2 which discriminates V/UV and outputs the result as pitch and V/UV data, and this device 2 discriminates a V section and an UV section in each analysis frame area. The decoded multipulse train is used instead of a waveform to extract the pitch, and pitch and V/UV data is outputted, and this data is supplied to a pulse train converter 3 and a pitch data assembler 5. The pulse train converter 3 reduces multipulses in the whole of V and UV sections in accordance with a desired encoding conversion speed. For example, a specific number of pulses which are determined in accordance with the desired encoding conversion speed are selected in the order of amplitude absolute value and the other pulses are removed with respect to the UV section. Thus, transmission is less delayed and the degradation in S/N is suppressed. |
