发明名称 VOICE DECODER
摘要 <p>PURPOSE:To decrease unnaturalness in a background noise in a silent mode by decreasing the Q value of a synthetic filter by a correction means corresponding to a measured Q value. CONSTITUTION:A synthetic filter coefficient generation circuit 3 generates a synthetic filter coefficient from a code string inputted from an input terminal 1, and outputs the synthetic coefficient to the synthetic filter 4 in a sounding state, and to a Q value measuring circuit 6 in a silent state. The Q value of the synthetic filter calculated by the synthetic filter coefficient generation circuit 3 is measured by the Q value measuring circuit 6. The synthetic filter coefficient found by the synthetic filter coefficient generation circuit 3 is changed by a synthetic filter coefficient correction circuit 7 based on the measured Q value. In the sounding state, voice output can be obtained by inputting an excitation signal outputted from an excitation signal generation circuit 2 and the synthetic filter coefficient outputted from the synthetic filter coefficient generation circuit 3 to the synthetic filter 4.</p>
申请公布号 JPH06125281(A) 申请公布日期 1994.05.06
申请号 JP19920272173 申请日期 1992.10.12
申请人 NEC CORP 发明人 HAYATA TOSHIHIRO
分类号 G10L19/06;G10L11/02;G10L19/00;G10L21/02;H04B1/46;H04B14/04;(IPC1-7):H04B1/46;G10L9/14 主分类号 G10L19/06
代理机构 代理人
主权项
地址