语音/音频处理学术速递[6.2]

今日论文合集：cs.SD语音27篇，eess.AS音频处理34篇。

cs.SD语音

【1】 Differentiable Allpass Filters for Phase Response Estimation and Automatic Signal Alignment标题：用于相位响应估计和信号自动对准的可微全通滤波器链接：https://arxiv.org/abs/2306.00860

作者：Anders R. Bargum,Stefania Serafin,Cumhur Erkut,Julian D. Parker机构：Multi Sensory Experience Lab, Aalborg University, Copenhagen, Denmark备注：Collaboration done while interning/employed at Native Instruments. Accepted for publication in Proc. DAFX'23, Copenhagen, Denmark, September 2023. Sound examples at this https URL

【2】 Adapting a ConvNeXt model to audio classification on AudioSet标题：将ConvNeXt模型应用于AudioSet的音频分类链接：https://arxiv.org/abs/2306.00830

作者：Thomas Pellegrini,Ismail Khalfaoui-Hassani,Etienne Labbé,Timothée Masquelier机构：Artificial and Natural Intelligence Toulouse Institute (ANITI), France, Institut de Recherche en Informatique de Toulouse (IRIT), France, Universit´e de Toulouse, CNRS, Toulouse INP, UT, France, CerCo UMR , CNRS, Universit´e de Toulouse, France备注：Accepted at INTERSPEECH 2023

【3】 Vocos: Closing the gap between time-domain and Fourier-based neural vocoders for high-quality audio synthesis标题：Vocos：缩小时域和基于傅立叶的神经声码器之间的差距，以实现高质量的音频合成链接：https://arxiv.org/abs/2306.00814

作者：Hubert Siuzdak

【4】 Adaptive Contextual Biasing for Transducer Based Streaming Speech Recognition标题：基于换能器的流传输语音识别的自适应上下文偏置链接：https://arxiv.org/abs/2306.00804

作者：Tianyi Xu,Zhanheng Yang,Kaixun Huang,Pengcheng Guo,Ao Zhang,Biao Li,Changru Chen,Chao Li,Lei Xie机构：Northwestern Polytechnical University, Xi’an, China

【5】 SlothSpeech: Denial-of-service Attack Against Speech Recognition Models标题：SlothSpeech：针对语音识别模型的拒绝服务攻击链接：https://arxiv.org/abs/2306.00794

作者：Mirazul Haque,Rutvij Shah,Simin Chen,Berrak Şişman,Cong Liu,Wei Yang机构：University of Texas at Dallas, USA, University of California, Riverside, USA

【6】 Enhancing the Unified Streaming and Non-streaming Model with Contrastive Learning标题：用对比学习增强流与非流的统一模型链接：https://arxiv.org/abs/2306.00755

作者：Yuting Yang,Yuke Li,Binbin Du机构：NetEase Yidun AI Lab, Hangzhou, China备注：Accepted by INTERSPEECH 2023

【7】 UnDiff: Unsupervised Voice Restoration with Unconditional Diffusion Model标题：UnDiff：基于无条件扩散模型的无监督语音恢复链接：https://arxiv.org/abs/2306.00721

作者：Anastasiia Iashchenko,Pavel Andreev,Ivan Shchekotov,Nicholas Babaev,Dmitry Vetrov机构：Vetrov, Samsung AI Center, Moscow, HSE University, Moscow, Skolkovo Institute of Science and Technology, Moscow, Artificial Intelligence Research Institute, Moscow, ∗equal contribution备注：Accepted to Interspeech 2023

【8】 Stuttering Detection Using Speaker Representations and Self-supervised Contextual Embeddings标题：基于说话人表示和自监督上下文嵌入的口吃检测链接：https://arxiv.org/abs/2306.00689

作者：Shakeel A. Sheikh,Md Sahidullah,Fabrice Hirsch,Slim Ouni机构：the date of receipt and acceptance should be inserted later备注：Accepted in International Journal of Speech Technology, Springer 2023 substantial overlap with arXiv:2204.01564

【9】 Encoder-decoder multimodal speaker change detection标题：编解码器多模式说话人变化检测链接：https://arxiv.org/abs/2306.00680

作者：Jee-weon Jung,Soonshin Seo,Hee-Soo Heo,Geonmin Kim,You Jin Kim,Young-ki Kwon,Minjae Lee,Bong-Jin Lee机构：NAVER Corporation, Republic of Korea ,NAVER Cloud Corporation, Republic of Korea备注：5 pages, accepted for presentation at INTERSPEECH 2023

【10】 EmoMix: Emotion Mixing via Diffusion Models for Emotional Speech Synthesis标题：EmoMix：基于扩散模型的情感语音合成情感混合链接：https://arxiv.org/abs/2306.00648

作者：Haobin Tang,Xulong Zhang,Jianzong Wang,Ning Cheng,Jing Xiao机构：Ping An Technology (Shenzhen) Co., Ltd., Shenzhen, China, University of Science and Technology of China, Hefei, China备注：Accepted by 24th Annual Conference of the International Speech Communication Association (INTERSPEECH 2023)

【11】 Adaptation and Optimization of Automatic Speech Recognition (ASR) for the Maritime Domain in the Field of VHF Communication标题：自动语音识别(ASR)在VHF通信领域中的海事领域适配和优化链接：https://arxiv.org/abs/2306.00614

作者：Emin Cagatay Nakilcioglu,Maximilian Reimann,Ole John机构：fer.de备注：None

【12】 Masked Autoencoders with Multi-Window Attention Are Better Audio Learners标题：具有多窗口注意的蒙版自动编码器是更好的音频学习工具链接：https://arxiv.org/abs/2306.00561

作者：Sarthak Yadav,Sergios Theodoridis,Lars Kai Hansen,Zheng-Hua Tan机构： Department of Electronic Systems, Aalborg University, Denmark, Pioneer Centre for Artificial Intelligence, Denmark, Department of Applied Mathematics and Computer Science, DTU, Denmark, Department of Informatics and Telecommunications

【13】 Speech inpainting: Context-based speech synthesis guided by video标题：语音修复：视频引导的基于上下文的语音合成链接：https://arxiv.org/abs/2306.00489

作者：Juan F. Montesinos,Daniel Michelsanti,Gloria Haro,Zheng-Hua Tan,Jesper Jensen机构：Universitat Pompeu Fabra, Department of Information and Communication Technologies, Spain, Aalborg University, Department of Electronic Systems, Denmark, Oticon AS, Denmark备注：Accepted in Interspeech23

【14】 Inspecting Spoken Language Understanding from Kids for Basic Math Learning at Home标题：考察幼儿对家庭基础数学的口语理解链接：https://arxiv.org/abs/2306.00482

作者：Eda Okur,Roddy Fuentes Alba,Saurav Sahay,Lama Nachman机构：Intel Labs, USA, Intel Labs, Mexico备注：Proceedings of the 18th Workshop on Innovative Use of NLP for Building Educational Applications (BEA) at ACL 2023

【15】 Towards hate speech detection in low-resource languages: Comparing ASR to acoustic word embeddings on Wolof and Swahili标题：在低资源语言中实现仇恨语音检测：将ASR与Wolof和斯瓦希里语上的声学单词嵌入进行比较链接：https://arxiv.org/abs/2306.00410

作者：Christiaan Jacobs,Nathanaël Carraz Rakotonirina,Everlyn Asiko Chimoto,Bruce A. Bassett,Herman Kamper机构：Department of Electrical & Electronic Engineering, Stellenbosch University, South Africa, Universitat Pompeu Fabra, Spain, African Institute for Mathematical Sciences, Cape Town, South Africa备注：Accepted to Interspeech 2023

【16】 Transfer Learning for Underrepresented Music Generation标题：未被充分代表的音乐世代的迁移学习链接：https://arxiv.org/abs/2306.00281

作者：Anahita Doosti,Matthew Guzdial机构：Department of Computing Science, Alberta Machine Intelligence Institute (Amii), University of Alberta, Edmonton, AB, Canada备注：None

【17】 Strategies for improving low resource speech to text translation relying on pre-trained ASR models标题：基于预训练ASR模型的低资源语音到文本翻译改进策略链接：https://arxiv.org/abs/2306.00208

作者：Santosh Kesiraju,Marek Sarvas,Tomas Pavlicek,Cecile Macaire,Alejandro Ciuba机构：Univ. Grenoble Alpes, France. ,University of Pittsburgh, USA.

【18】 MuseCoco: Generating Symbolic Music from Text标题：MuseCoco：从文本生成符号音乐链接：https://arxiv.org/abs/2306.00110

作者：Peiling Lu,Xin Xu,Chenfei Kang,Botao Yu,Chengyi Xing,Xu Tan,Jiang Bian机构：‡Microsoft Research Asia, ZZhejiang University, Shanghai Jiao Tong University, ^Nanjing University, ]Dalian University of Technology

【19】 MERT: Acoustic Music Understanding Model with Large-Scale Self-supervised Training标题：MERT：具有大规模自我监督训练的声乐理解模型链接：https://arxiv.org/abs/2306.00107

作者：Yizhi Li,Ruibin Yuan,Ge Zhang,Yinghao Ma,Xingran Chen,Hanzhi Yin,Chenghua Lin,Anton Ragni,Emmanouil Benetos,Norbert Gyenge,Roger Dannenberg,Ruibo Liu,Wenhu Chen,Gus Xia,Yemin Shi,Wenhao Huang,Yike Guo,Jie Fu机构：University of Sheffield, Beijing Academy of Artificial Intelligence, Carnegie Mellon University, University of Waterloo, Queen Mary University of London, University of Michigan Ann Arbor, Dartmouth College, New York University

【20】 How to Construct Perfect and Worse-than-Coin-Flip Spoofing Countermeasures: A Word of Warning on Shortcut Learning标题：如何构建完美和劣于硬币的欺骗对策：对捷径学习的一句警示链接：https://arxiv.org/abs/2306.00044

作者：Hye-jin Shim,Rosa González Hautamäki,Md Sahidullah,Tomi Kinnunen机构：University of Eastern Finland, Finland, University of Oulu, Finland, TCG CREST, India备注：Interspeech 2023

【21】 Speaker-specific Thresholding for Robust Imposter Identification in Unseen Speaker Recognition标题：基于特定说话人阈值的隐身说话人识别方法链接：https://arxiv.org/abs/2306.00952

作者：Ashutosh Chaubey,Sparsh Sinha,Susmita Ghose机构：LG Ad Solutions, India, LG Ad Solutions, USA

【22】 Harmonic enhancement using learnable comb filter for light-weight full-band speech enhancement model标题：基于可学习梳状滤波器的轻量级全频带语音增强模型的谐波增强链接：https://arxiv.org/abs/2306.00812

作者：Xiaohuai Le,Tong Lei,Li Chen,Yiqing Guo,Chao He,Cheng Chen,Xianjun Xia,Hua Gao,Yijian Xiao,Piao Ding,Shenyi Song,Jing Lu机构：Key Laboratory of Modern Acoustics, Nanjing University, Nanjing , China, RTC Lab, ByteDance, China, NJU-Horizon Intelligent Audio Lab, Horizon Robotics, Beijing , China备注：accepted by Interspeech 2023

【23】 Spoken Language Identification System for English-Mandarin Code-Switching Child-Directed Speech标题：面向英汉码转换儿童语音的口语识别系统链接：https://arxiv.org/abs/2306.00736

作者：Shashi Kant Gupta,Sushant Hiray,Prashant Kukde机构：RingCentral Innovation, India, RingCentral Inc., USA备注：Accepted by Interspeech 2023, 5 pages, 1 figure, 4 tables

【24】 A Teacher-Student approach for extracting informative speaker embeddings from speech mixtures标题：一种从混合语音中提取信息性说话人嵌入的师生方法链接：https://arxiv.org/abs/2306.00634

作者：Tobias Cord-Landwehr,Christoph Boeddeker,Cătălin Zorilă,Rama Doddipatla,Reinhold Haeb-Umbach机构：Haeb-Umbach, Paderborn University, Department of Communications Engineering, Paderborn, Germany, Toshiba Cambridge Research Laboratories, Cambridge, United Kingdom备注：Accepted for Interspeech 2023

【25】 Speaker verification using attentive multi-scale convolutional recurrent network标题：基于注意力多尺度卷积递归网络的说话人确认链接：https://arxiv.org/abs/2306.00426

作者：Yanxiong Li,Zhongjie Jiang,Wenchang Cao,Qisheng Huang机构：School of Electronic and Information Engineering, South China University of Technology, Guangzhou, China备注：21 pages, 6 figures, 8 tables. Accepted for publication in Applied Soft Computing

【26】 A Multi-dimensional Deep Structured State Space Approach to Speech Enhancement Using Small-footprint Models标题：一种基于小足迹模型的多维深层结构状态空间语音增强方法链接：https://arxiv.org/abs/2306.00331

作者：Pin-Jui Ku,Chao-Han Huck Yang,Sabato Marco Siniscalchi,Chin-Hui Lee机构：Georgia Institute of Technology, U.S.A, Kore University of Enna, Italy, Norwegian University of Science and Technology, Norway备注：Accepted to Interspeech 2023. Code will be released at this https URL

【27】 Audio-Visual Speech Separation in Noisy Environments with a Lightweight Iterative Model标题：基于轻量级迭代模型的噪声环境下的视听语音分离链接：https://arxiv.org/abs/2306.00160

作者：Héctor Martel,Julius Richter,Kai Li,Xiaolin Hu,Timo Gerkmann机构： Department of Computer Science and Technology, Tsinghua University, Beijing, China, Signal Processing (SP), Department of Informatics, Universit¨at Hamburg, Hamburg, Germany, Chinese Institute for Brain Research (CIBR), Beijing, China备注：Accepted by Interspeech 2023

eess.AS音频处理

【1】 Speaker-specific Thresholding for Robust Imposter Identification in Unseen Speaker Recognition标题：基于特定说话人阈值的隐身说话人识别方法链接：https://arxiv.org/abs/2306.00952

作者：Ashutosh Chaubey,Sparsh Sinha,Susmita Ghose机构：LG Ad Solutions, India, LG Ad Solutions, USA

【2】 Harmonic enhancement using learnable comb filter for light-weight full-band speech enhancement model标题：基于可学习梳状滤波器的轻量级全频带语音增强模型的谐波增强链接：https://arxiv.org/abs/2306.00812

【3】 Spoken Language Identification System for English-Mandarin Code-Switching Child-Directed Speech标题：面向英汉码转换儿童语音的口语识别系统链接：https://arxiv.org/abs/2306.00736

作者：Shashi Kant Gupta,Sushant Hiray,Prashant Kukde机构：RingCentral Innovation, India, RingCentral Inc., USA备注：Accepted by Interspeech 2023, 5 pages, 1 figure, 4 tables

【4】 A Teacher-Student approach for extracting informative speaker embeddings from speech mixtures标题：一种从混合语音中提取信息性说话人嵌入的师生方法链接：https://arxiv.org/abs/2306.00634

【5】 Frame-wise and overlap-robust speaker embeddings for meeting diarization标题：用于满足二值化的逐帧和抗重叠的说话人嵌入链接：https://arxiv.org/abs/2306.00625

作者：Tobias Cord-Landwehr,Christoph Boeddeker,Cătălin Zorilă,Rama Doddipatla,Reinhold Haeb-Umbach机构：⋆ Paderborn University, Department of Communications Engineering, Paderborn, Germany, † Toshiba Cambridge Research Laboratories, Cambridge, United Kingdom备注：ICASSP 2023

【6】 Automatic Data Augmentation for Domain Adapted Fine-Tuning of Self-Supervised Speech Representations标题：自监督语音表示域自适应微调的自动数据增强链接：https://arxiv.org/abs/2306.00481

作者：Salah Zaiem,Titouan Parcollet,Slim Essid机构：LTCI, T´el´ecom Paris, Institut Polytechnique de Paris, France, Samsung AI Center, Cambridge, United-Kingdom, University of Cambridge, United-Kingdom备注：6 pages,INTERSPEECH 2023

【7】 Speech Self-Supervised Representation Benchmarking: Are We Doing it Right?标题：演讲自我监督表征标杆：我们做得对吗？链接：https://arxiv.org/abs/2306.00452

作者：Salah Zaiem,Youcef Kemiche,Titouan Parcollet,Slim Essid,Mirco Ravanelli机构：LTCI, T´el´ecom Paris, Institut Polytechnique de Paris, France, Hi! PARIS Engineering Team, France, Capgemini, France, Samsung AI Center, Cambridge, United-Kingdom, University of Cambridge, United-Kingdom备注：None

【8】 Speaker verification using attentive multi-scale convolutional recurrent network标题：基于注意力多尺度卷积递归网络的说话人确认链接：https://arxiv.org/abs/2306.00426

【9】 A Multi-dimensional Deep Structured State Space Approach to Speech Enhancement Using Small-footprint Models标题：一种基于小足迹模型的多维深层结构状态空间语音增强方法链接：https://arxiv.org/abs/2306.00331

【10】 Speaker-independent Speech Inversion for Estimation of Nasalance标题：非特定人语音倒置的噪声估计链接：https://arxiv.org/abs/2306.00203

作者：Yashish M. Siriwardena,Carol Espy-Wilson,Suzanne Boyce,Mark K. Tiede,Liran Oren机构：University of Maryland College Park, Maryland, USA, University of Cincinnati, Ohio, USA, Haskins Laboratories, Connecticut, USA备注：Interspeech 2023

【11】 Audio-Visual Speech Separation in Noisy Environments with a Lightweight Iterative Model标题：基于轻量级迭代模型的噪声环境下的视听语音分离链接：https://arxiv.org/abs/2306.00160

【12】 Differentiable Allpass Filters for Phase Response Estimation and Automatic Signal Alignment标题：用于相位响应估计和信号自动对准的可微全通滤波器链接：https://arxiv.org/abs/2306.00860

【13】 Adapting a ConvNeXt model to audio classification on AudioSet标题：将ConvNeXt模型应用于AudioSet的音频分类链接：https://arxiv.org/abs/2306.00830

【14】 Vocos: Closing the gap between time-domain and Fourier-based neural vocoders for high-quality audio synthesis标题：Vocos：缩小时域和基于傅立叶的神经声码器之间的差距，以实现高质量的音频合成链接：https://arxiv.org/abs/2306.00814

作者：Hubert Siuzdak

【15】 Adaptive Contextual Biasing for Transducer Based Streaming Speech Recognition标题：基于换能器的流传输语音识别的自适应上下文偏置链接：https://arxiv.org/abs/2306.00804

作者：Tianyi Xu,Zhanheng Yang,Kaixun Huang,Pengcheng Guo,Ao Zhang,Biao Li,Changru Chen,Chao Li,Lei Xie机构：Northwestern Polytechnical University, Xi’an, China

【16】 SlothSpeech: Denial-of-service Attack Against Speech Recognition Models标题：SlothSpeech：针对语音识别模型的拒绝服务攻击链接：https://arxiv.org/abs/2306.00794

作者：Mirazul Haque,Rutvij Shah,Simin Chen,Berrak Şişman,Cong Liu,Wei Yang机构：University of Texas at Dallas, USA, University of California, Riverside, USA

【17】 Improved Cross-Lingual Transfer Learning For Automatic Speech Translation标题：用于语音自动翻译的改进的跨语言迁移学习链接：https://arxiv.org/abs/2306.00789

作者：Sameer Khurana,Nauman Dawalatabad,Antoine Laurent,Luis Vicente,Pablo Gimeno,Victoria Mingote,James Glass机构： and Victoria Mingote arewith the University of Zaragoza, Antoine Laurent is with LIUMUniversity

【18】 Enhancing the Unified Streaming and Non-streaming Model with Contrastive Learning标题：用对比学习增强流与非流的统一模型链接：https://arxiv.org/abs/2306.00755

作者：Yuting Yang,Yuke Li,Binbin Du机构：NetEase Yidun AI Lab, Hangzhou, China备注：Accepted by INTERSPEECH 2023

【19】 UnDiff: Unsupervised Voice Restoration with Unconditional Diffusion Model标题：UnDiff：基于无条件扩散模型的无监督语音恢复链接：https://arxiv.org/abs/2306.00721

【20】 How Generative Spoken Language Modeling Encodes Noisy Speech: Investigation from Phonetics to Syntactics标题：生成性口语模型如何编码有噪音的语音：从语音到句法的研究链接：https://arxiv.org/abs/2306.00697

作者：Joonyong Park,Shinnosuke Takamichi,Tomohiko Nakamura,Kentaro Seki,Detai Xin,Hiroshi Saruwatari机构：The University of Tokyo, Japan备注：Accepted to INTERSPEECH 2023

【21】 Stuttering Detection Using Speaker Representations and Self-supervised Contextual Embeddings标题：基于说话人表示和自监督上下文嵌入的口吃检测链接：https://arxiv.org/abs/2306.00689

【22】 Encoder-decoder multimodal speaker change detection标题：编解码器多模式说话人变化检测链接：https://arxiv.org/abs/2306.00680

【23】 EmoMix: Emotion Mixing via Diffusion Models for Emotional Speech Synthesis标题：EmoMix：基于扩散模型的情感语音合成情感混合链接：https://arxiv.org/abs/2306.00648

【24】 Adaptation and Optimization of Automatic Speech Recognition (ASR) for the Maritime Domain in the Field of VHF Communication标题：自动语音识别(ASR)在VHF通信领域中的海事领域适配和优化链接：https://arxiv.org/abs/2306.00614

作者：Emin Cagatay Nakilcioglu,Maximilian Reimann,Ole John机构：fer.de备注：None

【25】 Masked Autoencoders with Multi-Window Attention Are Better Audio Learners标题：具有多窗口注意的蒙版自动编码器是更好的音频学习工具链接：https://arxiv.org/abs/2306.00561

【26】 The Effects of Input Type and Pronunciation Dictionary Usage in Transfer Learning for Low-Resource Text-to-Speech标题：输入类型和发音词典使用对低资源文语转换迁移学习的影响链接：https://arxiv.org/abs/2306.00535

作者：Phat Do,Matt Coler,Jelske Dijkstra,Esther Klabbers机构：Campus Fryslˆan, University of Groningen, the Netherlands, Fryske AkademyMercator Research Centre, the Netherlands, ReadSpeaker, the Netherlands备注：Accepted at INTERSPEECH 2023

【27】 Speech inpainting: Context-based speech synthesis guided by video标题：语音修复：视频引导的基于上下文的语音合成链接：https://arxiv.org/abs/2306.00489

【28】 Inspecting Spoken Language Understanding from Kids for Basic Math Learning at Home标题：考察幼儿对家庭基础数学的口语理解链接：https://arxiv.org/abs/2306.00482

【29】 Towards hate speech detection in low-resource languages: Comparing ASR to acoustic word embeddings on Wolof and Swahili标题：在低资源语言中实现仇恨语音检测：将ASR与Wolof和斯瓦希里语上的声学单词嵌入进行比较链接：https://arxiv.org/abs/2306.00410

【30】 Transfer Learning for Underrepresented Music Generation标题：未被充分代表的音乐世代的迁移学习链接：https://arxiv.org/abs/2306.00281

作者：Anahita Doosti,Matthew Guzdial机构：Department of Computing Science, Alberta Machine Intelligence Institute (Amii), University of Alberta, Edmonton, AB, Canada备注：None

【31】 Strategies for improving low resource speech to text translation relying on pre-trained ASR models标题：基于预训练ASR模型的低资源语音到文本翻译改进策略链接：https://arxiv.org/abs/2306.00208

作者：Santosh Kesiraju,Marek Sarvas,Tomas Pavlicek,Cecile Macaire,Alejandro Ciuba机构：Univ. Grenoble Alpes, France. ,University of Pittsburgh, USA.

【32】 MuseCoco: Generating Symbolic Music from Text标题：MuseCoco：从文本生成符号音乐链接：https://arxiv.org/abs/2306.00110

【33】 MERT: Acoustic Music Understanding Model with Large-Scale Self-supervised Training标题：MERT：具有大规模自我监督训练的声乐理解模型链接：https://arxiv.org/abs/2306.00107

【34】 How to Construct Perfect and Worse-than-Coin-Flip Spoofing Countermeasures: A Word of Warning on Shortcut Learning标题：如何构建完美和劣于硬币的欺骗对策：对捷径学习的一句警示链接：https://arxiv.org/abs/2306.00044

作者：Hye-jin Shim,Rosa González Hautamäki,Md Sahidullah,Tomi Kinnunen机构：University of Eastern Finland, Finland, University of Oulu, Finland, TCG CREST, India备注：Interspeech 2023

机器翻译，仅供参考

语音/音频处理学术速递[6.2]

评论 0