| 摘要 |
<p>PURPOSE:To easily adapt the above system to the ATM network and to improve the efficiency by using a prediction coefficient adapted through the use of a full differential code when a voice signal for a special period is subject to difference coding. CONSTITUTION:A voice signal inputted at a prescribed sampling period is converted into a digital signal and stored in a memory 1 at each prescribed frame period. The, the signal is read sequentially by control signals 13, 14 and subject to differential coding at a subtractor 15 and a quantizer 2. The subtractor 15 takes a difference between a prediction signal outputted from a prediction device 6 and an input voice signal to calculate a difference signal. Moreover, a differential code bit number used for the adaptation of the prediction coefficient is made variable in response to the property of the voice signal inputted for a prescribed period to calculate the said difference signal. Thus, the full difference code is used to code the voice signal, then the system is easily adapted to the ATM network and the efficient voice packet transmission system is obtained.</p> |
