首页> 外文会议>Signal Processing and Communications Applications Conference >Gender identification of a speaker from voice source
【24h】

Gender identification of a speaker from voice source

机译:语音来源中说话人的性别识别

获取原文

摘要

There are close relationship between physiological properties of glottis and vocal tract and the age and gender of a speaker. Since physiological changes in glottis and vocal tract affects the voice of a speaker, the features obtained from speech signal can be used to determine the age and gender. In this study, the relation between the gender of a speaker and voice source occurred in glottis is investigated. For this purpose, the TIMIT database is used consisting of 438 male and 192 female speakers. The voice source of each speaker is obtained from the /a/ phoneme of each speech by inverse filtering. CIQ,SQ and dH1H2 parameters calculated from the voice source are used with respect to a threshold value to determine the gender of the speaker. In this study, 99% success for CIQ, 97% success for SQ and 94% success for dH1H2 are obtained.
机译:声门和声道的生理特性与说话者的年龄和性别之间有着密切的关系。由于声门和声道的生理变化会影响说话者的声音,因此从语音信号中获取的特征可用于确定年龄和性别。在这项研究中,研究了说话人的性别与声门中发生的声源之间的关系。为此,使用TIMIT数据库,其中包括438位男性和192位女性使用者。通过反向滤波从每个语音的/ a /音素获得每个扬声器的语音源。根据阈值使用从语音源计算的CIQ,SQ和dH1H2参数来确定说话者的性别。在这项研究中,CIQ成功率达99%,SQ成功率达97%,dH1H2成功率达94%。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号