| 摘要 |
PURPOSE:To eliminate the need for auxiliary information and to compensate sufficiently a high frequency component by providing plural adaptive differential PCM coders comprising quantizers or the like and processing a voice signal one by one sample. CONSTITUTION:Only a low frequency component is extracted from a sampled input voice signal by an LPF 15, a half the sampling signal is interleaved by a 1/n interleaving device 16 and each one sample is inputted to the 1st adaptive difference PCM coder ADPCM 17. The low frequency component of the input signal is quantized by a quantizer 23 and a residual signal is sent to a decoder. The low frequency component decoded in the ADPCM 17 is fed to the 2nd ADPCM 20, the forecast value of the input signal including the low frequency component and the high frequency component is fed to an adder 29, where the difference between the forecast value and the low frequency component is calculated and the result is quantized by a quantizer 30. Then the result is sent as a signal to compensate the high frequency component lost by the interleaving. Thus, since the signal is processed at each sample, the processing speed is quickened, no compensation information is required and the high frequency component is compensated sufficiently.
|
