Регулювання швидкості подання голосової інформації

Abstract

Speed-rate control of voice messages playback is of important practical weight for oral communication and voice signals processing technologies. The time-scaling method based on the differentiated expansion/compression of the speech unit duration reproducing natural regularities of the speech units temporal modifications is presented. The algorithm of speech segmentation into phone sections determined by six basic temporal groups has been elaborated. Procedures of adaptive transformations of different nature sound structures have been developed. High intelligibility and naturalness of the reproduced oral messages confirm the effectiveness of the adaptive time-domain voice modifications.