发明名称 VOICE INPUTTING DEVICE
摘要 PROBLEM TO BE SOLVED: To obtain a signal into which noise is hardly mixed and whose S/N is satisfactory by converting a voice signal into plural electric signals with plural kinds of directivity and plural frequency characteristics, and adding these plural electric signals. SOLUTION: A micro-unit constituted of microphones 41, 42, and 43 is fixed to a front pillar. The characteristics of this micro-unit are added by post-filters 44-49 and directivity adding parts 50 and 51. The directivity characteristics have a dip with the generating direction of an engine noise and the generating direction of a road noise generated due to the contact of a tire with a road for example, and the frequency characteristics have a dip with the frequencies of an engine noise and the frequencies of the road noise. Therefore, the noise can be effectively removed from a voice recognition processed voice signal according to the directivity characteristics and the frequency characteristics. On the other hand, the generating direction and frequencies of voice are made different from those of each noise so that the voice signal can be efficiently taken out.
申请公布号 JP2000069583(A) 申请公布日期 2000.03.03
申请号 JP19980239160 申请日期 1998.08.25
申请人 FUJITSU TEN LTD 发明人 YAMATO TOSHITAKA;SAKIYAMA KAZUHIRO;KAMIYA MASAHIRO
分类号 G10L15/20;G10L21/02;H04R1/28;H04R3/00;(IPC1-7):H04R1/28 主分类号 G10L15/20
代理机构 代理人
主权项
地址