| 摘要 |
PURPOSE:To quantize part of information determining the features of the whole zero-phase shifted fundamental waveform and to generate the whole zero-phase shifted fundamental waveform on a composition side from the information by reproducing a part which is not quantized with representative data which has necessary length, and superposing it in each repetition pitch cycle and synthesizing a voice. CONSTITUTION:A transmission part 14 multiplexes a pitch cycle 8, a maximum amplitude 11, and a vector index 13 and sends them to a reception part 16 through a transmission line 15. On a reception side, the vector index 13 received by a reception part 16 is reproduced by a reverse quantization part 17 and converted into normalized waveform data 18, and a reproduction part 19 finds reproduced waveform data 20 with a maximum amplitude 11' received by the reception part 16. A superposition part 21 superposes the reproduced waveform data 20 in each pitch cycle 8' received by the reception part 16, copies it to a voice waveform, and sets the result as frame-length synthesized voice data 22 in a two-plane buffer memory 23 for output. Consequently, all data can be reproduced on the reception side although only part of information is used on the transmission side. |
