| 摘要 |
PURPOSE: To provide a linear predictive parameter coding device applied to CELP coding and capable of coding to high performance even in a low bit rate. CONSTITUTION: A linear predictive parameter of a voice in sub-frame dividing the frame to plural parts is calculated for an input voice sectioned in frame by a sub-frame unit linear prediction parameter calculation means 11, and a quantization target is calculated so that a difference between an interpolation value between the quantization target and the former frame quantization parameter and input linear prediction parameters of respective sub-frames become minimum, and the obtained quantization target is quantized by a parameter quantization means 13, and a quantization code and a quantization parameter are outputted. |
