Tuning the performance of automatic speaker recognition in different conditions: effects of language and simulated voice disguise

IF 0.5 4区社会学 Q4 CRIMINOLOGY & PENOLOGY

International Journal of Speech Language and the Law Pub Date : 2020-01-10 DOI:10.1558/ijsll.39778

Radek Skarnitzl, Maral Asiaee, Mandana Nourbakhsh

{"title":"Tuning the performance of automatic speaker recognition in different conditions: effects of language and simulated voice disguise","authors":"Radek Skarnitzl, Maral Asiaee, Mandana Nourbakhsh","doi":"10.1558/ijsll.39778","DOIUrl":null,"url":null,"abstract":"Automatic speaker recognition applications have often been described as a ‘black box’. This study explores the benefit of tuning procedures (condition adaptation and reference normalisation) implemented in an i-vector PLDA framework ASR system, VOCALISE. These procedures enable users to open the black box to a certain degree. Subsets of two 100-speaker databases, one of Czech and the other of Persian male speakers, are used for the baseline condition and for the tuning procedures. The effect of tuning with cross-language material, as well as the effect of simulated voice disguise, achieved by raising the fundamental frequency by four semitones and resonance characteristics by 8%, are also examined. The results show superior recognition performance (EER) for Persian than Czech in the baseline condition, but an opposite result in the simulated disguise condition; possible reasons for this are discussed. Overall, the study suggests that both condition adaptation and reference normalisation are beneficial to recognition performance.","PeriodicalId":43843,"journal":{"name":"International Journal of Speech Language and the Law","volume":"26 1","pages":"209–229-209–229"},"PeriodicalIF":0.5000,"publicationDate":"2020-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Speech Language and the Law","FirstCategoryId":"98","ListUrlMain":"https://doi.org/10.1558/ijsll.39778","RegionNum":4,"RegionCategory":"社会学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"CRIMINOLOGY & PENOLOGY","Score":null,"Total":0}

引用次数: 2

Abstract

Automatic speaker recognition applications have often been described as a ‘black box’. This study explores the benefit of tuning procedures (condition adaptation and reference normalisation) implemented in an i-vector PLDA framework ASR system, VOCALISE. These procedures enable users to open the black box to a certain degree. Subsets of two 100-speaker databases, one of Czech and the other of Persian male speakers, are used for the baseline condition and for the tuning procedures. The effect of tuning with cross-language material, as well as the effect of simulated voice disguise, achieved by raising the fundamental frequency by four semitones and resonance characteristics by 8%, are also examined. The results show superior recognition performance (EER) for Persian than Czech in the baseline condition, but an opposite result in the simulated disguise condition; possible reasons for this are discussed. Overall, the study suggests that both condition adaptation and reference normalisation are beneficial to recognition performance.

查看原文本刊更多论文

调整不同条件下的自动说话人识别性能:语言和模拟声音伪装的影响

自动语音识别应用通常被描述为一个“黑匣子”。本研究探讨了在i矢量PLDA框架ASR系统VOCALISE中实施的调谐程序(条件适应和参考归一化)的好处。这些程序在一定程度上使用户能够打开黑匣子。两个100人说话数据库的子集，一个是捷克语数据库，另一个是波斯语数据库，用于基线条件和调优程序。本文还研究了跨语言材料调音的效果，以及通过将基频提高四个半音和共振特性提高8%来实现的模拟语音伪装的效果。结果表明，在基线条件下，波斯语的识别性能优于捷克语，而在模拟伪装条件下则相反;讨论了可能的原因。总的来说，研究表明条件适应和参考归一化都有利于识别性能。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

International Journal of Speech Language and the Law Multiple-

CiteScore

1.00

自引率

25.00%

发文量

期刊介绍： The International Journal of Speech, Language and the Law is a peer-reviewed journal that publishes articles on any aspect of forensic language, speech and audio analysis. Founded in 1994 as Forensic Linguistics, the journal changed to its present title in 2003 to reflect a broadening of academic coverage and readership. Subscription to the journal is included in membership of the International Association of Forensic Linguists and the International Association for Forensic Phonetics and Acoustics.