发明名称 SPEECH ENHANCEMENT WITH MULTIPLE SENSORS USING PRECEDING CLEAR SPEECH
摘要 FIELD: information technology. ^ SUBSTANCE: method and device determine channel characteristic for an alternative sensor using an alternative sensor signal and an air-conduction microphone signal. Channel characteristic and prior probability distribution for clear speech values is then used to estimate the value of clear speech. ^ EFFECT: detection and correction of a speech signal when distorted by speech of another speaker. ^ 17 cl, 6 dwg
申请公布号 RU2407074(C2) 申请公布日期 2010.12.20
申请号 RU20070147463 申请日期 2006.06.06
申请人 MAJKROSOFT KORPOREJSHN 发明人 LJU TSZYCHEHN;ASERO ALEKHANDRO;CHZHAN CHZHEN'JU
分类号 G10L21/02 主分类号 G10L21/02
代理机构 代理人
主权项
地址