| 摘要 |
PURPOSE:To decrease greatly and visually the effect of the quantizing noise to the reproduced sound as well as miniaturize the scale of the device, by using the transversal filter for the predictor. CONSTITUTION:Aural signal Xj to be coded is supplied through terminal 1, and the predictive value including the quantizing noise is subtracted by subtractor 10 to be quantized through quantizer 20. The quantized signal is supplied to the differential pulse modulation DPCM decoder from terminal 2 as well as to subtractor 50 and filter 130. Subtractor 50 carries out the subtraction between the input and the output of quantizer 20, and the delivers output -nj equal to the noise component generated at quantizer 20 to then supply it to filter 70. Both filters 70 and 130 are formed with the transversal filters, and the sum of the outputs of filters 70 and 130 is applied to the subtractor in the form of the predictive value. As a result, the stability of the reception side is assured, and at the same time the high-range noise and the low-frequency noise can be eliminated to obtain an extremely high-quality sound. |
