发明名称 ENCODING METHOD AND DECODING METHOD FOR SPEECH SIGNAL
摘要 PURPOSE:To improve the reproduced sound quality on a synthesis side by deciding whether or not an intermediate peak is large to a certain extent or above and finding and sending position information. CONSTITUTION:The output signal of a filter 12 is sent to a peak search part 26, which detects the maximum peak in a block. A pitch peak detection part 27 detects a peak which repeats in pitch cycles (p) on the basis of the position of the maximum peak and basic pitch information from a high-precision pitch search part 16. Then a local maximum detection part 28 detects peaks which become maximum nearby intermediate positions between respective pitch peak positions and nearby intermediate positions between the respective pitch peak positions and intermediate positions (what is called a maximum peak). Then a classification part 29 perform classification into types I and II by using statistics (total value) e1-e3, and output 1-bit phase code data which becomes 0 at the time of the type I and 1 at the time of the type II to an encoding part 21.
申请公布号 JPH0792998(A) 申请公布日期 1995.04.07
申请号 JP19930245758 申请日期 1993.09.30
申请人 SONY CORP 发明人 CHIYAN JIYOSEFU
分类号 G10L19/02;G10L13/00;H03M7/30;(IPC1-7):G10L7/04;G10L9/00 主分类号 G10L19/02
代理机构 代理人
主权项
地址