| 摘要 |
PURPOSE:To encode voice waveforms which have high tone qualities by providing a linear prediction analyzing part at a framework encoder, and transmitting the frequency characteristics of the voice waveforms to a decoder in the shape of a linear prediction parameter. CONSTITUTION:A linear prediction coefficient obtained by a linear prediction analyzing part 6 is encoded by a parameter encoding part 5 by using a parameter code book 4 in order to search parameter information, towards an input voice signal 3. Then, an average waveform of one pitch is found based on pitch information obtained by a pitch analyzing part 7 towards the input voice signal 3, and the basic waveform is operated for filtering by the linear prediction coefficient in order to obtain a basic residual waveform. Moreover, the time sequences (framework) of the plural kinds of pulses indicating the shape are retrieved by a framework retrieving part 8 in order to obtain framework information, and the information of the waveforms strung among the frameworks is encoded by a framework waveform encoding part 9 by using a framework waveform code book 10 in order to find the waveform information among frameworks. Then, the information obtained by an encoder 1 is transmitted to a decod er 2. Thus, the voice waveforms which have the high tone qualities can be encoded by the simple data processing of a low pit rate. |
