语音/音频处理学术速递[5.19]

今日论文合集：cs.SD语音25篇，eess.AS音频处理35篇。

cs.SD语音

【1】 ONE-PEACE: Exploring One General Representation Model Toward Unlimited Modalities标题：一种和平：探索一种通向无限模式的表征模式链接：https://arxiv.org/abs/2305.11172

作者：Peng Wang,Shijie Wang,Junyang Lin,Shuai Bai,Xiaohuan Zhou,Jingren Zhou,Xinggang Wang,Chang Zhou机构：DAMO Academy, Alibaba Group, Huazhong University of Science and Technology备注：30 pages, 9 figures, 18 tables

【2】 Unsupervised Multi-channel Separation and Adaptation标题：无监督多通道分离与自适应链接：https://arxiv.org/abs/2305.11151

作者：Cong Han,Kevin Wilson,Scott Wisdom,John R. Hershey机构：Columbia University,Google

【3】 QPGesture: Quantization-Based and Phase-Guided Motion Matching for Natural Speech-Driven Gesture Generation标题：QPGesture：基于量化和相位引导的自然语音驱动手势生成链接：https://arxiv.org/abs/2305.11094

作者：Sicheng Yang,Zhiyong Wu,Minglei Li,Zhensong Zhang,Lei Hao,Weihong Bao,Haolin Zhuang机构： Tsinghua Shenzhen International Graduate School, Tsinghua University, Huawei Cloud Computing Technologies Co., Ltd , Huawei Noah’s Ark Lab, The Chinese University of Hong Kong备注：15 pages, 12 figures, CVPR 2023 Highlight

【4】 A Comparative Study on E-Branchformer vs Conformer in Speech Recognition, Translation, and Understanding Tasks标题：E-Branchformer与Conform在语音识别、翻译和理解任务中的对比研究链接：https://arxiv.org/abs/2305.11073

作者：Yifan Peng,Kwangyoun Kim,Felix Wu,Brian Yan,Siddhant Arora,William Chen,Jiyang Tang,Suwon Shon,Prashant Sridhar,Shinji Watanabe机构：Carnegie Mellon University, Pittsburgh, PA, USA, ASAPP Inc., Mountain View, CA, USA备注：Accepted at INTERSPEECH 2023. Code: this https URL

【5】 FunASR: A Fundamental End-to-End Speech Recognition Toolkit标题：FunASR：一个基本的端到端语音识别工具包链接：https://arxiv.org/abs/2305.11013

作者：Zhifu Gao,Zerui Li,Jiaming Wang,Haoneng Luo,Xian Shi,Mengzhe Chen,Yabin Li,Lingyun Zuo,Zhihao Du,Zhangyu Xiao,Shiliang Zhang机构：Speech Lab of DAMO Academy, Alibaba Group, China备注：5 pages, 3 figures, accepted by INTERSPEECH 2023

【6】 GETMusic: Generating Any Music Tracks with a Unified Representation and Diffusion Framework标题：GETMusic：使用统一的表示和扩散框架生成任何音乐曲目链接：https://arxiv.org/abs/2305.10841

作者：Ang Lv,Xu Tan,Peiling Lu,Wei Ye,Shikun Zhang,Jiang Bian,Rui Yan机构：†Microsoft Research Asia, ‡Gaoling School of Artifical Intelligence, Renmin University of China, §National Engineering Research Center for Software Engineering, Peking University备注：15 pages, 4 figures

【7】 A Lexical-aware Non-autoregressive Transformer-based ASR Model标题：一种基于词汇感知的非自回归变换ASR模型链接：https://arxiv.org/abs/2305.10839

作者：Chong-En Lin,Kuan-Yu Chen机构：National Taiwan University of Science and Technology, Taiwan备注：Accepted by Interspeech 2023

【8】 Validation of an ECAPA-TDNN system for Forensic Automatic Speaker Recognition under case work conditions标题：案例工作条件下ECAPA-TDNN自动说话人识别系统的验证链接：https://arxiv.org/abs/2305.10805

作者：Francesco Sigona,Mirko Grimaldi机构：Centro di ricerca interdisciplinare sul linguaggio (CRIL), University of Salento, Lecce, Italy

【9】 Whisper-KDQ: A Lightweight Whisper via Guided Knowledge Distillation and Quantization for Efficient ASR标题：Whisper-KDQ：一种基于引导知识提取和量化的高效ASR轻量级Whisper链接：https://arxiv.org/abs/2305.10788

作者：Hang Shao,Wei Wang,Bei Liu,Xun Gong,Haoyu Wang,Yanmin Qian机构：MoE Key Lab of Artificial Intelligence, AI Institute, X-LANCE Lab, Department of Computer Science and Engineering, Shanghai Jiao Tong University, Shanghai, China

【10】 CLAPSpeech: Learning Prosody from Text Context with Contrastive Language-Audio Pre-training标题：CLAPSpeech：通过对比语言从文本上下文中学习韵律-音频预训练链接：https://arxiv.org/abs/2305.10763

作者：Zhenhui Ye,Rongjie Huang,Yi Ren,Ziyue Jiang,Jinglin Liu,Jinzheng He,Xiang Yin,Zhou Zhao机构：Zhejiang University, Bytedance, ByteDance备注：Accepted by ACL 2023 (Main Conference)

【11】 Noise-aware Speech Separation with Contrastive Learning标题：基于对比学习的噪声感知语音分离链接：https://arxiv.org/abs/2305.10761

作者：Zizheng Zhang,Chen Chen,Xiang Liu,Yuchen Hu,Eng Siong Chng机构：School of Software and Microelectronics, Peking University, China, School of Computer Science and Engineering, Nanyang Technological University, Singapore

【12】 Diffusion-Based Speech Enhancement with Joint Generative and Predictive Decoders标题：基于扩散的产生式和预测式联合解码器语音增强链接：https://arxiv.org/abs/2305.10734

作者：Hao Shi,Kazuki Shimada,Masato Hirano,Takashi Shibuya,Yuichiro Koyama,Zhi Zhong,Shusuke Takahashi,Tatsuya Kawahara,Yuki Mitsufuji机构：Graduate School of Informatics, Kyoto University, Kyoto, Japan, Sony Group Corporation, Tokyo, Japan

【13】 Attention-based Encoder-Decoder Network for End-to-End Neural Speaker Diarization with Target Speaker Attractor标题：基于注意力的带目标说话人吸引子的端到端神经说话人二值化编解码器网络链接：https://arxiv.org/abs/2305.10704

作者：Zhengyang Chen,Bing Han,Shuai Wang,Yanmin Qian机构：MoE Key Lab of Artificial Intelligence, AI Institute, X-LANCE Lab, Department of Computer Science and Engineering, Shanghai Jiao Tong University备注：Accepted by InterSpeech 2023

【14】 RMSSinger: Realistic-Music-Score based Singing Voice Synthesis标题：RMSSinger：基于真实感乐谱的歌声合成链接：https://arxiv.org/abs/2305.10686

作者：Jinzheng He,Jinglin Liu,Zhenhui Ye,Rongjie Huang,Chenye Cui,Huadai Liu,Zhou Zhao机构：Zhejiang University, ByteDance备注：Accepted by Finding of ACL2023

【15】 Accurate and Reliable Confidence Estimation Based on Non-Autoregressive End-to-End Speech Recognition System标题：基于非自回归端到端语音识别系统的准确可靠置信度估计链接：https://arxiv.org/abs/2305.10680

作者：Xian Shi,Haoneng Luo,Zhifu Gao,Shiliang Zhang,Zhijie Yan机构：Speech Lab of DAMO Academy, Alibaba Group, China备注：5 pages, 4 figures

【16】 a unified front-end framework for english text-to-speech synthesis标题：一种统一的英语文语合成前端框架链接：https://arxiv.org/abs/2305.10666

作者：Zelin Ying,Chen Li,Yu Dong,Qiuqiang Kong,YuanYuan Huo,Yuping Wang,Yuxuan Wang机构：Speech, Audio & Music Intelligence (SAMI), ByteDance备注：5 pages, 3 figures

【17】 Speech Separation based on Contrastive Learning and Deep Modularization标题：基于对比学习和深度模块化的语音分离链接：https://arxiv.org/abs/2305.10652

作者：Peter Ochieng机构：Department of Computer Science., University of Cambridge.备注：arXiv admin note: substantial text overlap with arXiv:2212.00369

【18】 ZeroPrompt: Streaming Acoustic Encoders are Zero-Shot Masked LMs标题：ZeroPrompt：流声音编码器是Zero-Shot掩码LM链接：https://arxiv.org/abs/2305.10649

作者：Xingchen Song,Di Wu,Binbin Zhang,Zhendong Peng,Bo Dang,Fuping Pan,Zhiyong Wu机构：Tsinghua Univ., Beijing, China ,Horizon Inc., Beijing, China ,WeNet Open Source Community备注：accepted by interspeech 2023

【19】 ML-SUPERB: Multilingual Speech Universal PERformance Benchmark标题：ML-Superb：多语言语音通用性能基准链接：https://arxiv.org/abs/2305.10615

作者：Jiatong Shi,Dan Berrebbi,William Chen,Ho-Lam Chung,En-Pei Hu,Wei Ping Huang,Xuankai Chang,Shang-Wen Li,Abdelrahman Mohamed,Hung-yi Lee,Shinji Watanabe机构：Carnegie Mellon University, National Taiwan University, Meta AI, Rembrand备注：Accepted by Interspeech

【20】 Prompting the Hidden Talent of Web-Scale Speech Models for Zero-Shot Task Generalization标题：激发网络规模语音模型中的隐藏人才进行零命中任务概括链接：https://arxiv.org/abs/2305.11095

作者：Puyuan Peng,Brian Yan,Shinji Watanabe,David Harwath机构：Department of Computer Science, The University of Texas at Austin, USA, Language Technology Institute, Carnegie Mellon University, USA备注：Interspeech 2023

【21】 FastFit: Towards Real-Time Iterative Neural Vocoder by Replacing U-Net Encoder With Multiple STFTs标题：FastFit：用多个短时傅立叶变换代替U-Net编码器实现实时迭代神经网络声码器链接：https://arxiv.org/abs/2305.10823

作者：Won Jang,Dan Lim,Heayoung Park机构：Kakao Enterprise Corporation, Republic of Korea备注：Accepted to INTERSPEECH 2023

【22】 TempAdaCos: Learning Temporally Structured Embeddings for Few-Shot Keyword Spotting with Dynamic Time Warping标题：TempAdaCos：利用动态时间规整学习时间结构嵌入的Few-Shot关键词检测链接：https://arxiv.org/abs/2305.10816

作者：Kevin Wilkinghoff,Alessia Cornaggia-Urrigshardt机构：Fraunhofer FKIE, Fraunhoferstr. , Wachtberg, Germany

【23】 Listen, Think, and Understand标题：倾听、思考和理解链接：https://arxiv.org/abs/2305.10790

作者：Yuan Gong,Hongyin Luo,Alexander H. Liu,Leonid Karlinsky,James Glass机构：MIT CSAIL, MIT-IBM Watson AI Lab备注：Preprint, work in progress

【24】 Data Augmentation for Diverse Voice Conversion in Noisy Environments标题：噪声环境下不同语音转换的数据增强链接：https://arxiv.org/abs/2305.10684

作者：Avani Tanna,Michael Saxon,Amr El Abbadi,William Yang Wang机构：University of California, Santa Barbara备注：Interspeech 2023 Show and Tell, 2 pp

【25】 Use of Speech Impairment Severity for Dysarthric Speech Recognition标题：语音损伤严重程度在节律障碍语音识别中的应用链接：https://arxiv.org/abs/2305.10659

作者：Mengzhe Geng,Zengrui Jin,Tianzi Wang,Shujie Hu,Jiajun Deng,Mingyu Cui,Guinan Li,Jianwei Yu,Xurong Xie,Xunying Liu备注：Accepted to INTERSPEECH2023

eess.AS音频处理

【1】 mdctGAN: Taming transformer-based GAN for speech super-resolution with Modified DCT spectra标题：MdctGAN：利用改进的DCT谱实现语音超分辨率的驯服Transformer基GaN链接：https://arxiv.org/abs/2305.11104

作者：Chenhao Shuai,Chaohua Shi,Lu Gan,Hongqing Liu机构：Nanyang Technological University, Singapore, Xidian University, China, Brunel University London, UK, Chongqing University of Posts and Telecommunications, China备注：5 pages, 4 figures, INTERSPEECH 2023

【2】 Prompting the Hidden Talent of Web-Scale Speech Models for Zero-Shot Task Generalization标题：激发网络规模语音模型中的隐藏人才进行零命中任务概括链接：https://arxiv.org/abs/2305.11095

【3】 Improving Generalization Ability of Countermeasures for New Mismatch Scenario by Combining Multiple Advanced Regularization Terms标题：组合多个高级正则化项提高新失配场景对策的泛化能力链接：https://arxiv.org/abs/2305.10940

作者：Chang Zeng,Xin Wang,Xiaoxiao Miao,Erica Cooper,Junichi Yamagishi机构：National Institute of Informatics, Japan ,SOKENDAI, Japan备注：Accepted by interspeech2023

【4】 Diffusion-Based Mel-Spectrogram Enhancement for Personalized Speech Synthesis with Found Data标题：基于扩散的Mel谱图增强用于个性化语音合成链接：https://arxiv.org/abs/2305.10891

作者：Yusheng Tian,Wei Liu,Tan Lee机构：Department of Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR

【5】 FastFit: Towards Real-Time Iterative Neural Vocoder by Replacing U-Net Encoder With Multiple STFTs标题：FastFit：用多个短时傅立叶变换代替U-Net编码器实现实时迭代神经网络声码器链接：https://arxiv.org/abs/2305.10823

作者：Won Jang,Dan Lim,Heayoung Park机构：Kakao Enterprise Corporation, Republic of Korea备注：Accepted to INTERSPEECH 2023

【6】 Locate and Beamform: Two-dimensional Locating All-neural Beamformer for Multi-channel Speech Separation标题：定位与波束形成：用于多通道语音分离的二维定位全神经波束形成器链接：https://arxiv.org/abs/2305.10821

作者：Yanjie Fu,Meng Ge,Honglong Wang,Nan Li,Haoran Yin,Longbiao Wang,Gaoyan Zhang,Jianwu Dang,Chengyun Deng,Fei Wang机构：College of Intelligence and Computing, Tianjin University, Tianjin, China, Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Beijing Xiaoju Technology Co., Ltd., Beijing, China备注：Accepted by Interspeech 2023. arXiv admin note: substantial text overlap with arXiv:2212.03401

【7】 TempAdaCos: Learning Temporally Structured Embeddings for Few-Shot Keyword Spotting with Dynamic Time Warping标题：TempAdaCos：利用动态时间规整学习时间结构嵌入的Few-Shot关键词检测链接：https://arxiv.org/abs/2305.10816

作者：Kevin Wilkinghoff,Alessia Cornaggia-Urrigshardt机构：Fraunhofer FKIE, Fraunhoferstr. , Wachtberg, Germany

【8】 Listen, Think, and Understand标题：倾听、思考和理解链接：https://arxiv.org/abs/2305.10790

作者：Yuan Gong,Hongyin Luo,Alexander H. Liu,Leonid Karlinsky,James Glass机构：MIT CSAIL, MIT-IBM Watson AI Lab备注：Preprint, work in progress

【9】 Enhancing Speech Articulation Analysis using a Geometric Transformation of the X-ray Microbeam Dataset标题：利用X射线微束数据集的几何变换增强语音清晰度分析链接：https://arxiv.org/abs/2305.10775

作者：Ahmed Adel Attia,Mark Tiede,Carol Y. Espy-Wilson机构：University of Maryland College Park, MD, Haskins Laboratories, New Haven, CT

【10】 A Multi-Task Learning Framework for Sound Event Detection using High-level Acoustic Characteristics of Sounds标题：利用声音高层声学特征检测声音事件的多任务学习框架链接：https://arxiv.org/abs/2305.10729

作者：Tanmay Khandelwal,Rohan Kumar Das机构：Fortemedia Singapore, Singapore备注：Accepted for Publication at INTERSPEECH 2023

【11】 Data Augmentation for Diverse Voice Conversion in Noisy Environments标题：噪声环境下不同语音转换的数据增强链接：https://arxiv.org/abs/2305.10684

作者：Avani Tanna,Michael Saxon,Amr El Abbadi,William Yang Wang机构：University of California, Santa Barbara备注：Interspeech 2023 Show and Tell, 2 pp

【12】 Use of Speech Impairment Severity for Dysarthric Speech Recognition标题：语音损伤严重程度在节律障碍语音识别中的应用链接：https://arxiv.org/abs/2305.10659

作者：Mengzhe Geng,Zengrui Jin,Tianzi Wang,Shujie Hu,Jiajun Deng,Mingyu Cui,Guinan Li,Jianwei Yu,Xurong Xie,Xunying Liu备注：Accepted to INTERSPEECH2023

【13】 Investigating Range-Equalizing Bias in Mean Opinion Score Ratings of Synthesized Speech标题：合成语音平均评分中的幅度均衡偏差研究链接：https://arxiv.org/abs/2305.10608

作者：Erica Cooper,Junichi Yamagishi机构：National Institute of Informatics, Japan备注：Accepted to Interspeech 2023

【14】 Improving Speaker Verification with Self-Pretrained Transformer Models标题：利用自训练Transformer模型改进说话人确认链接：https://arxiv.org/abs/2305.10517

作者：Junyi Peng,Oldřich Plchot,Themos Stafylakis,Ladislav Mošner,Lukáš Burget,Jan Černocký机构：Omilia - Conversational Intelligence, Athens, Greece备注：Accepted to Interspeech 2023

【15】 ONE-PEACE: Exploring One General Representation Model Toward Unlimited Modalities标题：一种和平：探索一种通向无限模式的表征模式链接：https://arxiv.org/abs/2305.11172

【16】 Unsupervised Multi-channel Separation and Adaptation标题：无监督多通道分离与自适应链接：https://arxiv.org/abs/2305.11151

作者：Cong Han,Kevin Wilson,Scott Wisdom,John R. Hershey机构：Columbia University,Google

【17】 QPGesture: Quantization-Based and Phase-Guided Motion Matching for Natural Speech-Driven Gesture Generation标题：QPGesture：基于量化和相位引导的自然语音驱动手势生成链接：https://arxiv.org/abs/2305.11094

【18】 A Comparative Study on E-Branchformer vs Conformer in Speech Recognition, Translation, and Understanding Tasks标题：E-Branchformer与Conform在语音识别、翻译和理解任务中的对比研究链接：https://arxiv.org/abs/2305.11073

【19】 Self-supervised Fine-tuning for Improved Content Representations by Speaker-invariant Clustering标题：基于说话人不变聚类改进内容表示的自监督精调链接：https://arxiv.org/abs/2305.11072

作者：Heng-Jui Chang,Alexander H. Liu,James Glass机构：MIT CSAIL, USA备注：Accepted to Interspeech 2023

【20】 FunASR: A Fundamental End-to-End Speech Recognition Toolkit标题：FunASR：一个基本的端到端语音识别工具包链接：https://arxiv.org/abs/2305.11013

【21】 Making More of Little Data: Improving Low-Resource Automatic Speech Recognition Using Data Augmentation标题：充分利用小数据：使用数据增强改进低资源自动语音识别链接：https://arxiv.org/abs/2305.10951

作者：Martijn Bartelds,Nay San,Bradley McDonnell,Dan Jurafsky,Martijn Wieling机构：University of Groningen, Stanford University, University of Hawai’i at M¯anoa备注：Accepted at ACL 2023

【22】 GETMusic: Generating Any Music Tracks with a Unified Representation and Diffusion Framework标题：GETMusic：使用统一的表示和扩散框架生成任何音乐曲目链接：https://arxiv.org/abs/2305.10841

【23】 A Lexical-aware Non-autoregressive Transformer-based ASR Model标题：一种基于词汇感知的非自回归变换ASR模型链接：https://arxiv.org/abs/2305.10839

作者：Chong-En Lin,Kuan-Yu Chen机构：National Taiwan University of Science and Technology, Taiwan备注：Accepted by Interspeech 2023

【24】 Validation of an ECAPA-TDNN system for Forensic Automatic Speaker Recognition under case work conditions标题：案例工作条件下ECAPA-TDNN自动说话人识别系统的验证链接：https://arxiv.org/abs/2305.10805

作者：Francesco Sigona,Mirko Grimaldi机构：Centro di ricerca interdisciplinare sul linguaggio (CRIL), University of Salento, Lecce, Italy

【25】 Whisper-KDQ: A Lightweight Whisper via Guided Knowledge Distillation and Quantization for Efficient ASR标题：Whisper-KDQ：一种基于引导知识提取和量化的高效ASR轻量级Whisper链接：https://arxiv.org/abs/2305.10788

【26】 CLAPSpeech: Learning Prosody from Text Context with Contrastive Language-Audio Pre-training标题：CLAPSpeech：通过对比语言从文本上下文中学习韵律-音频预训练链接：https://arxiv.org/abs/2305.10763

作者：Zhenhui Ye,Rongjie Huang,Yi Ren,Ziyue Jiang,Jinglin Liu,Jinzheng He,Xiang Yin,Zhou Zhao机构：Zhejiang University, Bytedance, ByteDance备注：Accepted by ACL 2023 (Main Conference)

【27】 Noise-aware Speech Separation with Contrastive Learning标题：基于对比学习的噪声感知语音分离链接：https://arxiv.org/abs/2305.10761

【28】 Diffusion-Based Speech Enhancement with Joint Generative and Predictive Decoders标题：基于扩散的产生式和预测式联合解码器语音增强链接：https://arxiv.org/abs/2305.10734

【29】 Attention-based Encoder-Decoder Network for End-to-End Neural Speaker Diarization with Target Speaker Attractor标题：基于注意力的带目标说话人吸引子的端到端神经说话人二值化编解码器网络链接：https://arxiv.org/abs/2305.10704

【30】 RMSSinger: Realistic-Music-Score based Singing Voice Synthesis标题：RMSSinger：基于真实感乐谱的歌声合成链接：https://arxiv.org/abs/2305.10686

作者：Jinzheng He,Jinglin Liu,Zhenhui Ye,Rongjie Huang,Chenye Cui,Huadai Liu,Zhou Zhao机构：Zhejiang University, ByteDance备注：Accepted by Finding of ACL2023

【31】 Accurate and Reliable Confidence Estimation Based on Non-Autoregressive End-to-End Speech Recognition System标题：基于非自回归端到端语音识别系统的准确可靠置信度估计链接：https://arxiv.org/abs/2305.10680

作者：Xian Shi,Haoneng Luo,Zhifu Gao,Shiliang Zhang,Zhijie Yan机构：Speech Lab of DAMO Academy, Alibaba Group, China备注：5 pages, 4 figures

【32】 a unified front-end framework for english text-to-speech synthesis标题：一种统一的英语文语合成前端框架链接：https://arxiv.org/abs/2305.10666

作者：Zelin Ying,Chen Li,Yu Dong,Qiuqiang Kong,YuanYuan Huo,Yuping Wang,Yuxuan Wang机构：Speech, Audio & Music Intelligence (SAMI), ByteDance备注：5 pages, 3 figures

【33】 Speech Separation based on Contrastive Learning and Deep Modularization标题：基于对比学习和深度模块化的语音分离链接：https://arxiv.org/abs/2305.10652

作者：Peter Ochieng机构：Department of Computer Science., University of Cambridge.备注：arXiv admin note: substantial text overlap with arXiv:2212.00369

【34】 ZeroPrompt: Streaming Acoustic Encoders are Zero-Shot Masked LMs标题：ZeroPrompt：流声音编码器是Zero-Shot掩码LM链接：https://arxiv.org/abs/2305.10649

【35】 ML-SUPERB: Multilingual Speech Universal PERformance Benchmark标题：ML-Superb：多语言语音通用性能基准链接：https://arxiv.org/abs/2305.10615

机器翻译，仅供参考

语音/音频处理学术速递[5.19]

评论 0