{"title":"Automatic epoch extraction for closed-phase analysis of speech","authors":"A. Kounoudes, P. Naylor, M. Brookes","doi":"10.1109/ICDSP.2002.1028254","DOIUrl":null,"url":null,"abstract":"This paper presents an automatic method to determine the instants of glottal closure (GCIs), or epochs, from the speech signal alone without the need of a laryngograph signal. The proposed algorithm incorporate a new technique for estimating GCI candidates and dynamic programming to select the best candidates according to predefined cost functions. Results show accuracy in estimation to within /spl plusmn/0.25ms on 87% of the test database and less that 1% false alarms and misses. Preliminary experiments using the telephone -degraded NTIMIT database have shown that the algorithm continues to perform well even in the presence of noise.","PeriodicalId":351073,"journal":{"name":"2002 14th International Conference on Digital Signal Processing Proceedings. DSP 2002 (Cat. No.02TH8628)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2002-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"4","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2002 14th International Conference on Digital Signal Processing Proceedings. DSP 2002 (Cat. No.02TH8628)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICDSP.2002.1028254","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 4
Abstract
This paper presents an automatic method to determine the instants of glottal closure (GCIs), or epochs, from the speech signal alone without the need of a laryngograph signal. The proposed algorithm incorporate a new technique for estimating GCI candidates and dynamic programming to select the best candidates according to predefined cost functions. Results show accuracy in estimation to within /spl plusmn/0.25ms on 87% of the test database and less that 1% false alarms and misses. Preliminary experiments using the telephone -degraded NTIMIT database have shown that the algorithm continues to perform well even in the presence of noise.