您好,欢迎来到帮我找美食网。
搜索
您的当前位置:首页Experiments in Gaussian-mixture-model speaker recognition

Experiments in Gaussian-mixture-model speaker recognition

来源:帮我找美食网
FEMALE SPEECH40 )%20 ni( ytiliba10 borp ssiM 5 mel−cepstrum from synthesized speechrec−cepstra from LSFs w pitch 2 rec−cepstrum from LSFs w/o pitch mel−cepstra from all−pole spectrum 1 0.50.5 1 2 5 10 20 40 False Alarm probability (in %)MALE SPEECH40 )%20 ni( ytiliba10 borp ssiM 5 mel−cepstrum from synthesized speechrec 2 −cepstra from LSFs w pitch rec−cepstrum from LSFs w/o pitch mel 1 −cepstra from all−pole spectrum 0.50.5 1 2 5 10 20 40 False Alarm probability (in %)LOG−SPECTRUM1008010820LOG−MEL ENERGIES60a) 4020001000200030004000500005101520251008060402010820b) 00100020003000400050000510152025Frequency (Hz)Mel−Filter #FEMALE SPEECH40 )%20 ni( ytiliba10 borp ssiM 5 mel−cepstrum from: synthesized speech 2 all−pole spectrum w pitch striations all−pole spectrum w/o pitch striations 1 0.50.5 1 2 5 10 20 40 False Alarm probability (in %)MALE SPEECH40 )%20 ni( ytiliba10 borp ssiM 5 melsynthesized speech 2 −cepstrum from: all−pole spectrum w pitch striations all−pole spectrum w/o pitch striations 1 0.50.5 1 2 5 10 20 40 False Alarm probability (in %)

因篇幅问题不能全部显示,请点此查看更多更全内容

Copyright © 2019- banwoyixia.com 版权所有 湘ICP备2023022004号-1

违法及侵权请联系:TEL:199 1889 7713 E-MAIL:2724546146@qq.com

本站由北京市万商天勤律师事务所王兴未律师提供法律服务