arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2203.16032cs.SDeess.AS
收藏

ConferencingSpeech 2022挑战赛:在线会议应用的非侵入式客观语音质量评估(NISQA)挑战赛

ConferencingSpeech 2022 Challenge: Non-intrusive Objective Speech Quality Assessment (NISQA) Challenge for Online Conferencing Applications

Tencent Ethereal Audio Lab.(腾讯天籁音频实验室);Technical University of Berlin(柏林工业大学);Microsoft Corp.(微软公司);Indiana University Bloomington(印第安纳大学伯明顿分校);Southern University of Science and Technology(南方科技大学);XiDian University(西安电子科技大学)
Gaoxiong Yi, Wei Xiao, Yiming Xiao, Babak Naderi, Sebastian Möller, Wafaa Wardah, Gabriel Mittag, Ross Cutler, Zhuohuang Zhang, Donald S. Williamson, Fei Chen, … 展开作者

Gaoxiong Yi, Wei Xiao, Yiming Xiao, Babak Naderi, Sebastian Möller, Wafaa Wardah, Gabriel Mittag, Ross Cutler, Zhuohuang Zhang, Donald S. Williamson, Fei Chen, Fuzheng Yang, Shidong Shang

摘要:

随着语音通信系统(如在线会议应用程序)的进步,我们可以与人们无缝合作,无论他们身在何处。然而,在在线会议期间,语音质量可能会受到背景噪声、混响、数据包丢失、网络抖动等的显著影响。由于其性质,语音质量传统上在实验室进行主观测试,最近也在遵循ITU-T Rec.P.800系列国际标准的众包中进行评估。然而,这些方法成本高昂,无法应用于客户数据。因此,需要一种有效的客观评估方法来评估或监测正在进行的会话的语音质量。ConferencingSpeech 2022挑战赛针对语音质量评估任务的非侵入式深度神经网络模型。我们开源了一个训练语料库,其中包含超过86 K不同语言的语音片段,通过众包提供了各种合成和实时降级及其相应的主观质量分数。18支队伍提交了他们的模型进行评估。盲法测试集包括约4300个来自广泛降解范围的夹子。本文描述了挑战,数据集和评估方法,并报告了最终结果。

英文摘要:

With the advances in speech communication systems such as online conferencing applications, we can seamlessly work with people regardless of where they are. However, during online meetings, speech quality can be significantly affected by background noise, reverberation, packet loss, network jitter, etc. Because of its nature, speech quality is traditionally assessed in subjective tests in laboratories and lately also in crowdsourcing following the international standards from ITU-T Rec. P.800 series. However, those approaches are costly and cannot be applied to customer data. Therefore, an effective objective assessment approach is needed to evaluate or monitor the speech quality of the ongoing conversation. The ConferencingSpeech 2022 challenge targets the non-intrusive deep neural network models for the speech quality assessment task. We open-sourced a training corpus with more than 86K speech clips in different languages, with a wide range of synthesized and live degradations and their corresponding subjective quality scores through crowdsourcing. 18 teams submitted their models for evaluation in this challenge. The blind test sets included about 4300 clips from wide ranges of degradations. This paper describes the challenge, the datasets, and the evaluation methods and reports the final results.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑