| 摘要 |
PURPOSE: To reduce the processing time, and to improve the natural sound of the synthesizing signal by leaving a part of the original signal, which includes the basic information of each block, as it is, and changing the residual part. CONSTITUTION: In this sound signal synthesizing method by connection of waveform, which present the unit of element of the sound signal, with the lapse of time, at least the waveform related to the sound to be output as voice is divided into plural blocks, which are synchronized with the basic frequency of the signal and which correspond to the response of sound pipe in relation to a continuous stimulus impact of a vocal cord, and weighting is performed in each block, and the signal generated by the weighting is substituted with a copy thereof when the time is shifted by the quantity, which depends on the rhythm element information, and the sifted signal are overlapped and added to each other for sound synthesis. In each block of the original signal to be regenerated by the synthesis, a non-changed part of the original signal, which includes the basic information and which is not changed in the synthesized signal so as to be regenerated, is discriminated, and weighting, overlap and addition are performed to the residual block. |
