| 摘要 |
PURPOSE:To improve the sound quality in a low-encoding-speed area by replacing information of multipulses with information of a transversal filter by using pitch information and determining an impulse sequence inputted to the transversal filter so as to generate the multipulses. CONSTITUTION:When a voice signal Si is a voiced signal, a transversal filter coefficient determining device 15 uses the pitch information inputted from a pitch extractor 2 to determine a coefficient (of the transversal filter) for substituting the information of multipulses outputted by a multipulse retrieval device 14 and the information of a tap position. the determined coefficient and tap position are quantized by a coefficient quantizer 16 and outputted to a multiplexing circuit 23. The output signal of the coefficient quantizer 16 is decoded and inputted to the constituted transversal filter to determine the impulse sequence for generating the multipulses as the sound source information of the voice signal Si. Consequently, the sound quality in an area where the encoding speed is 4.8kb/s is improved. |
