SEAC: Serbian Emotional Amateur Cellphone Speech Corpus

Author:

Suzić Siniša1,Nosek Tijana1,Sečujski Milan1,Popović Branislav1,Krstanović Lidija1,Vujović Mia1,Simić Nikola1,Janev Marko2,Jakovljević Nikša1,Delić Vlado1

Affiliation:

1. University of Novi Sad, Faculty of Technical Sciences

2. Serbian Academy of Sciences and Arts, Institute of Mathematics

Abstract

Abstract Emotional speech recognition and synthesis of expressive speech are highly dependable on the availability of emotional speech corpora. In this paper, we present the creation and verification of the Serbian Emotional Amateur Cellphone Speech Corpus (SEAC), which was released by the University of Novi Sad, Faculty of Technical Sciences in 2022, as the first amateur emotional speech corpus in Serbian language, recorded over cellphones. The corpus contains emotional speech elicited from 53 different speakers (24 male and 29 female) in 5 different emotional states (neutral, happiness, sadness, fear and anger), and its total duration amounts to approximately 8 hours of speech data. Initial objective evaluation of the corpus has confirmed high correlation between the behaviour of acoustic parameters corresponding to different emotional states in the newly recorded corpus and the existing Serbian emotional speech corpus recorded by 6 professional actors, which was used as a source for reference recordings. The corpus was further evaluated through listening tests concerned with human emotion recognition. Finally, we present the results of experiments concerning emotion recognition and speaker recognition by several conventional machine learning systems carried out on the corpus, as well as the results of a cross-lingual emotion recognition experiment involving a state-of-the-art machine learning system based on deep convolutional neural networks.

Publisher

Research Square Platform LLC

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3