摘要
By taking advantage of the simple tone structure of distinct pitch contours in Mandarin speech, pitch information is orthogonally trans- formed and vector quantized. An average bit rate of 0.78 bits/frame (34.67 bits/s) for voiced sounds was achieved.
原文 | English |
---|---|
文章編號 | 61370 |
頁(從 - 到) | 1317-1320 |
頁數 | 4 |
期刊 | IEEE Transactions on Communications |
卷 | 38 |
發行號 | 9 |
DOIs | |
出版狀態 | Published - 9月 1990 |