今日论文合集：cs.SD语音31篇，eess.AS音频处理33篇。cs.SD语音【1】 Generating Realistic Images from In-the-wild Sounds标题：从野外声音生成逼真的图像链接：https://arxiv.org/abs/2309.02405作者：Taegyeong Lee,Jeonghun Kang,Hyeonyu Kim,Taehwan Kim备注：Accepted to ICCV 2023【2】 Voice Morphing: Two Identities in One Voice标题：声音变形：一个声音中的两个身份链接：https://arxiv.org/abs/2309.02404作者：Sushanta K. Pani,Anurag Chowdhury,Morgan Sandler,Arun Ross备注：Accepted oral paper at BIOSIG 2023【3】 The Batik-plays-Mozart Corpus: Linking Performance to Score to Musicological Annotations标题：蜡染-戏剧-莫扎特语料库：将演奏和配乐与音乐学注释联系起来链接：https://arxiv.org/abs/2309.02399作者：Patricia Hu,Gerhard Widmer备注：To be published in the Proceedings of the International Society for Music Information Retrieval Conference (ISMIR)【4】 Self-Similarity-Based and Novelty-based loss for music structure analysis标题：基于自相似和新奇损失的音乐结构分析链接：https://arxiv.org/abs/2309.02243作者：Geoffroy Peeters【5】 FSD: An Initial Chinese Dataset for Fake Song Detection标题：FSD：一个用于假歌检测的初始中文数据集链接：https://arxiv.org/abs/2309.02232作者：Yuankun Xie,Jingjing Zhou,Xiaolin Lu,Zhenghao Jiang,Yuxin Yang,Haonan Cheng,Long Ye备注：Submitted to ICASSP 2024【6】 Bring the Noise: Introducing Noise Robustness to Pretrained Automatic Speech Recognition标题：带来噪声：将噪声稳健性引入预先训练的自动语音识别链接：https://arxiv.org/abs/2309.02145作者：Patrick Eickhoff,Matthias Möller,Theresa Pekarek Rosin,Johannes Twiefel,Stefan Wermter备注：Submitted and accepted for ICANN 2023 (32nd International Conference on Artificial Neural Networks)【7】 Evaluating Methods for Ground-Truth-Free Foreign Accent Conversion标题：无事实基础的外国口音转换的评价方法链接：https://arxiv.org/abs/2309.02133作者：Wen-Chin Huang,Tomoki Toda备注：Accepted to the 2023 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC). Demo page: this https URL Code: this https URL【8】 RADIO: Reference-Agnostic Dubbing Video Synthesis标题：广播：参考不可知的配音视频合成链接：https://arxiv.org/abs/2309.01950作者：Dongyeun Lee,Chaewon Kim,Sangjoon Yu,Jaejun Yoo,Gyeong-Moon Park备注：Under review【9】 TODM: Train Once Deploy Many Efficient Supernet-Based RNN-T Compression For On-device ASR Models标题：TODM：一次训练为设备上ASR模型部署多个高效的基于超网的RNN-T压缩链接：https://arxiv.org/abs/2309.01947作者：Yuan Shangguan,Haichuan Yang,Danni Li,Chunyang Wu,Yassir Fathullah,Dilin Wang,Ayushi Dalmia,Raghuraman Krishnamoorthi,Ozlem Kalinli,Junteng Jia,Jay Mahadeokar,Xin Lei,Mike Seltzer,Vikas Chandra备注：Meta AI; Submitted to ICASSP 2024【10】 A Comparative Analysis of Pretrained Language Models for Text-to-Speech标题：文语转换中几种预训练语言模型的比较分析链接：https://arxiv.org/abs/2309.01576作者：Marcel Granero-Moya,Penny Karanasou,Sri Karlapati,Bastian Schnell,Nicole Peinelt,Alexis Moinet,Thomas Drugman备注：Accepted for presentation at the 12th ISCA Speech Synthesis Workshop (SSW) in Grenoble, France, from 26th to 28th August 2023【11】 BadSQA: Stealthy Backdoor Attacks Using Presence Events as Triggers in Non-Intrusive Speech Quality Assessment标题：BadSQA：在非侵入式语音质量评估中使用在线事件作为触发器的秘密后门攻击链接：https://arxiv.org/abs/2309.01480作者：Ying Ren,Kailai Shen,Zhe Ye,Diqun Yan备注：5 pages, 6 figures,conference【12】 SememeASR: Boosting Performance of End-to-End Speech Recognition against Domain and Long-Tailed Data Shift with Sememe Semantic Knowledge标题：SemomeASR：利用义素语义知识对域和长尾数据转移提高端到端语音识别性能链接：https://arxiv.org/abs/2309.01437作者：Jiaxu Zhu,Changhe Song,Zhiyong Wu,Helen Meng备注：Accepted by INTERSPEECH 2023【13】 MDSC: Towards Evaluating the Style Consistency Between Music and标题：MDSC：评估音乐和音乐之间的风格一致性链接：https://arxiv.org/abs/2309.01340作者：Zixiang Zhou,Baoyuan Wang备注：17 pages, 17 figure【14】 NADiffuSE: Noise-aware Diffusion-based Model for Speech Enhancement标题：NADiffuSE：基于噪声感知扩散的语音增强模型链接：https://arxiv.org/abs/2309.01212作者：Wen Wang,Dongchao Yang,Qichen Ye,Bowen Cao,Yuexian Zou【15】 MAGMA: Music Aligned Generative Motion Autodecoder标题：Magma：音乐对准的生成式运动自动解码器链接：https://arxiv.org/abs/2309.01202作者：Sohan Anisetty,Amit Raj,James Hays【16】 Federated Few-shot Learning for Cough Classification with Edge Devices标题：基于边缘设备的联合小概率学习用于咳嗽分类链接：https://arxiv.org/abs/2309.01076作者：Ngan Dao Hoang,Dat Tran-Anh,Manh Luong,Cong Tran,Cuong Pham备注：21 pages, 5 figures【17】 Timbre-reserved Adversarial Attack in Speaker Identification标题：说话人辨认中保留音色的对抗性攻击链接：https://arxiv.org/abs/2309.00929作者：Qing Wang,Jixun Yao,Li Zhang,Pengcheng Guo,Lei Xie备注：11 pages, 8 figures【18】 BLSP: Bootstrapping Language-Speech Pre-training via Behavior Alignment of Continuation Writing标题：BLSP：通过续写行为对齐的自举语言--言语预训练链接：https://arxiv.org/abs/2309.00916作者：Chen Wang,Minpeng Liao,Zhongqiang Huang,Jinliang Lu,Junhong Wu,Yuchen Liu,Chengqing Zong,Jiajun Zhang【19】 DiCLET-TTS: Diffusion Model based Cross-lingual Emotion Transfer for Text-to-Speech -- A Study between English and Mandarin标题：DiCLET-TTS：基于扩散模型的文语转换跨语言情感传递--基于英语和汉语的对比研究链接：https://arxiv.org/abs/2309.00883作者：Tao Li,Chenxu Hu,Jian Cong,Xinfa Zhu,Jingbei Li,Qiao Tian,Yuping Wang,Lei Xie备注：accepted by TASLP【20】 Pretraining Representations for Bioacoustic Few-shot Detection using Supervised Contrastive Learning标题：基于有监督对比学习的生物声小射击检测前训练表示法链接：https://arxiv.org/abs/2309.00878作者：Ilyass Moummad,Romain Serizel,Nicolas Farrugia【21】 Contextual Biasing of Named-Entities with Large Language Models标题：使用大型语言模型的命名实体的上下文偏向链接：https://arxiv.org/abs/2309.00723作者：Chuanneng Sun,Zeeshan Ahmed,Yingyi Ma,Zhe Liu,Yutong Pang,Ozlem Kalinli备注：5 pages, 4 figures. Conference: ICASSP 2024【22】 Employing Real Training Data for Deep Noise Suppression标题：利用真实训练数据进行深度噪声抑制链接：https://arxiv.org/abs/2309.02432作者：Ziyi Xu,Marvin Sach,Jan Pirklbauer,Tim Fingscheidt【23】 Personalized Adaptation with Pre-trained Speech Encoders for Continuous Emotion Recognition标题：用于连续情感识别的带预训练语音编码器的个性化自适应链接：https://arxiv.org/abs/2309.02418作者：Minh Tran,Yufeng Yin,Mohammad Soleymani备注：Accepted by INTERSPEECH 2023【24】 PromptTTS 2: Describing and Generating Voices with Text Prompt标题：PromptTTS 2：使用文本提示描述和生成语音链接：https://arxiv.org/abs/2309.02285作者：Yichong Leng,Zhifang Guo,Kai Shen,Xu Tan,Zeqian Ju,Yanqing Liu,Yufei Liu,Dongchao Yang,Leying Zhang,Kaitao Song,Lei He,Xiang-Yang Li,Sheng Zhao,Tao Qin,Jiang Bian备注：Demo page: this https URL【25】 PESTO: Pitch Estimation with Self-supervised Transposition-equivariant Objective标题：PESTO：基于自监督换位等变目标的基音估计链接：https://arxiv.org/abs/2309.02265作者：Alain Riou,Stefan Lattner,Gaëtan Hadjeres,Geoffroy Peeters【26】 Single-Channel Speech Enhancement with Deep Complex U-Networks and Probabilistic Latent Space Models标题：基于深度复U-网络和概率潜在空间模型的单通道语音增强链接：https://arxiv.org/abs/2309.01535作者：Eike J. Nustede,Jörn Anemüller备注：None【27】 RGI-Net: 3D Room Geometry Inference from Room Impulse Responses in the Absence of First-order Echoes标题：RGI-Net：无一阶回波情况下房间脉冲响应的三维房间几何推断链接：https://arxiv.org/abs/2309.01513作者：Inmo Yeon,Jung-Woo Choi备注：5 pages, 3 figures, 3 tables【28】 Noise robust speech emotion recognition with signal-to-noise ratio adapting speech enhancement标题：信噪比自适应语音增强的抗噪语音情感识别链接：https://arxiv.org/abs/2309.01164作者：Yu-Wen Chen,Julia Hirschberg,Yu Tsao【29】 MSM-VC: High-fidelity Source Style Transfer for Non-Parallel Voice Conversion by Multi-scale Style Modeling标题：MSM-VC：基于多尺度风格建模的非并行语音转换高保真信源风格转换链接：https://arxiv.org/abs/2309.01142作者：Zhichao Wang,Xinsheng Wang,Qicong Xie,Tao Li,Lei Xie,Qiao Tian,Yuping Wang备注：This work was submitted on April 10, 2022 and accepted on August 29, 2023【30】 Acoustic-to-articulatory inversion for dysarthric speech: Are pre-trained self-supervised representations favorable?标题：构音障碍语音的声学-发音反转：预先训练的自我监督表征是否有利？链接：https://arxiv.org/abs/2309.01108作者：Sarthak Kumar Maharana,Krishna Kamal Adidam,Shoumik Nandi,Ajitesh Srivastava【31】 Improving Small Footprint Few-shot Keyword Spotting with Supervision on Auxiliary Data标题：利用辅助数据监督改进小占地Few-Shot关键词识别链接：https://arxiv.org/abs/2309.00647作者：Seunghan Yang,Byeonggeun Kim,Kyuhong Shim,Simyung Chang备注：Interspeech 2023eess.AS音频处理【1】 Employing Real Training Data for Deep Noise Suppression标题：利用真实训练数据进行深度噪声抑制链接：https://arxiv.org/abs/2309.02432作者：Ziyi Xu,Marvin Sach,Jan Pirklbauer,Tim Fingscheidt【2】 Personalized Adaptation with Pre-trained Speech Encoders for Continuous Emotion Recognition标题：用于连续情感识别的带预训练语音编码器的个性化自适应链接：https://arxiv.org/abs/2309.02418作者：Minh Tran,Yufeng Yin,Mohammad Soleymani备注：Accepted by INTERSPEECH 2023【3】 In-Ear-Voice: Towards Milli-Watt Audio Enhancement With Bone-Conduction Microphones for In-Ear Sensing Platforms标题：入耳语音：用于入耳传感平台的骨传导麦克风的毫瓦级音频增强链接：https://arxiv.org/abs/2309.02393作者：Philipp Schilk,Niccolò Polvani,Andrea Ronco,Milos Cernak,Michele Magno【4】 PromptTTS 2: Describing and Generating Voices with Text Prompt标题：PromptTTS 2：使用文本提示描述和生成语音链接：https://arxiv.org/abs/2309.02285作者：Yichong Leng,Zhifang Guo,Kai Shen,Xu Tan,Zeqian Ju,Yanqing Liu,Yufei Liu,Dongchao Yang,Leying Zhang,Kaitao Song,Lei He,Xiang-Yang Li,Sheng Zhao,Tao Qin,Jiang Bian备注：Demo page: this https URL【5】 PESTO: Pitch Estimation with Self-supervised Transposition-equivariant Objective标题：PESTO：基于自监督换位等变目标的基音估计链接：https://arxiv.org/abs/2309.02265作者：Alain Riou,Stefan Lattner,Gaëtan Hadjeres,Geoffroy Peeters【6】 Single-Channel Speech Enhancement with Deep Complex U-Networks and Probabilistic Latent Space Models标题：基于深度复U-网络和概率潜在空间模型的单通道语音增强链接：https://arxiv.org/abs/2309.01535作者：Eike J. Nustede,Jörn Anemüller备注：None【7】 RGI-Net: 3D Room Geometry Inference from Room Impulse Responses in the Absence of First-order Echoes标题：RGI-Net：无一阶回波情况下房间脉冲响应的三维房间几何推断链接：https://arxiv.org/abs/2309.01513作者：Inmo Yeon,Jung-Woo Choi备注：5 pages, 3 figures, 3 tables【8】 Noise robust speech emotion recognition with signal-to-noise ratio adapting speech enhancement标题：信噪比自适应语音增强的抗噪语音情感识别链接：https://arxiv.org/abs/2309.01164作者：Yu-Wen Chen,Julia Hirschberg,Yu Tsao【9】 MSM-VC: High-fidelity Source Style Transfer for Non-Parallel Voice Conversion by Multi-scale Style Modeling标题：MSM-VC：基于多尺度风格建模的非并行语音转换高保真信源风格转换链接：https://arxiv.org/abs/2309.01142作者：Zhichao Wang,Xinsheng Wang,Qicong Xie,Tao Li,Lei Xie,Qiao Tian,Yuping Wang备注：This work was submitted on April 10, 2022 and accepted on August 29, 2023【10】 Acoustic-to-articulatory inversion for dysarthric speech: Are pre-trained self-supervised representations favorable?标题：构音障碍语音的声学-发音反转：预先训练的自我监督表征是否有利？链接：https://arxiv.org/abs/2309.01108作者：Sarthak Kumar Maharana,Krishna Kamal Adidam,Shoumik Nandi,Ajitesh Srivastava【11】 Improving Small Footprint Few-shot Keyword Spotting with Supervision on Auxiliary Data标题：利用辅助数据监督改进小占地Few-Shot关键词识别链接：https://arxiv.org/abs/2309.00647作者：Seunghan Yang,Byeonggeun Kim,Kyuhong Shim,Simyung Chang备注：Interspeech 2023【12】 Generating Realistic Images from In-the-wild Sounds标题：从野外声音生成逼真的图像链接：https://arxiv.org/abs/2309.02405作者：Taegyeong Lee,Jeonghun Kang,Hyeonyu Kim,Taehwan Kim备注：Accepted to ICCV 2023【13】 Voice Morphing: Two Identities in One Voice标题：声音变形：一个声音中的两个身份链接：https://arxiv.org/abs/2309.02404作者：Sushanta K. Pani,Anurag Chowdhury,Morgan Sandler,Arun Ross备注：Accepted oral paper at BIOSIG 2023【14】 The Batik-plays-Mozart Corpus: Linking Performance to Score to Musicological Annotations标题：蜡染-戏剧-莫扎特语料库：将演奏和配乐与音乐学注释联系起来链接：https://arxiv.org/abs/2309.02399作者：Patricia Hu,Gerhard Widmer备注：To be published in the Proceedings of the International Society for Music Information Retrieval Conference (ISMIR)【15】 Self-Similarity-Based and Novelty-based loss for music structure analysis标题：基于自相似和新奇损失的音乐结构分析链接：https://arxiv.org/abs/2309.02243作者：Geoffroy Peeters【16】 FSD: An Initial Chinese Dataset for Fake Song Detection标题：FSD：一个用于假歌检测的初始中文数据集链接：https://arxiv.org/abs/2309.02232作者：Yuankun Xie,Jingjing Zhou,Xiaolin Lu,Zhenghao Jiang,Yuxin Yang,Haonan Cheng,Long Ye备注：Submitted to ICASSP 2024【17】 Bring the Noise: Introducing Noise Robustness to Pretrained Automatic Speech Recognition标题：带来噪声：将噪声稳健性引入预先训练的自动语音识别链接：https://arxiv.org/abs/2309.02145作者：Patrick Eickhoff,Matthias Möller,Theresa Pekarek Rosin,Johannes Twiefel,Stefan Wermter备注：Submitted and accepted for ICANN 2023 (32nd International Conference on Artificial Neural Networks)【18】 Evaluating Methods for Ground-Truth-Free Foreign Accent Conversion标题：无事实基础的外国口音转换的评价方法链接：https://arxiv.org/abs/2309.02133作者：Wen-Chin Huang,Tomoki Toda备注：Accepted to the 2023 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC). Demo page: this https URL Code: this https URL【19】 Leveraging Label Information for Multimodal Emotion Recognition标题：利用标签信息进行多通道情感识别链接：https://arxiv.org/abs/2309.02106作者：Peiying Wang,Sunlu Zeng,Junqing Chen,Lu Fan,Meng Chen,Youzheng Wu,Xiaodong He备注：Accepted by Interspeech 2023【20】 RADIO: Reference-Agnostic Dubbing Video Synthesis标题：广播：参考不可知的配音视频合成链接：https://arxiv.org/abs/2309.01950作者：Dongyeun Lee,Chaewon Kim,Sangjoon Yu,Jaejun Yoo,Gyeong-Moon Park备注：Under review【21】 TODM: Train Once Deploy Many Efficient Supernet-Based RNN-T Compression For On-device ASR Models标题：TODM：一次训练为设备上ASR模型部署多个高效的基于超网的RNN-T压缩链接：https://arxiv.org/abs/2309.01947作者：Yuan Shangguan,Haichuan Yang,Danni Li,Chunyang Wu,Yassir Fathullah,Dilin Wang,Ayushi Dalmia,Raghuraman Krishnamoorthi,Ozlem Kalinli,Junteng Jia,Jay Mahadeokar,Xin Lei,Mike Seltzer,Vikas Chandra备注：Meta AI; Submitted to ICASSP 2024【22】 A Comparative Analysis of Pretrained Language Models for Text-to-Speech标题：文语转换中几种预训练语言模型的比较分析链接：https://arxiv.org/abs/2309.01576作者：Marcel Granero-Moya,Penny Karanasou,Sri Karlapati,Bastian Schnell,Nicole Peinelt,Alexis Moinet,Thomas Drugman备注：Accepted for presentation at the 12th ISCA Speech Synthesis Workshop (SSW) in Grenoble, France, from 26th to 28th August 2023【23】 BadSQA: Stealthy Backdoor Attacks Using Presence Events as Triggers in Non-Intrusive Speech Quality Assessment标题：BadSQA：在非侵入式语音质量评估中使用在线事件作为触发器的秘密后门攻击链接：https://arxiv.org/abs/2309.01480作者：Ying Ren,Kailai Shen,Zhe Ye,Diqun Yan备注：5 pages, 6 figures,conference【24】 SememeASR: Boosting Performance of End-to-End Speech Recognition against Domain and Long-Tailed Data Shift with Sememe Semantic Knowledge标题：SemomeASR：利用义素语义知识对域和长尾数据转移提高端到端语音识别性能链接：https://arxiv.org/abs/2309.01437作者：Jiaxu Zhu,Changhe Song,Zhiyong Wu,Helen Meng备注：Accepted by INTERSPEECH 2023【25】 MDSC: Towards Evaluating the Style Consistency Between Music and标题：MDSC：评估音乐和音乐之间的风格一致性链接：https://arxiv.org/abs/2309.01340作者：Zixiang Zhou,Baoyuan Wang备注：17 pages, 17 figure【26】 NADiffuSE: Noise-aware Diffusion-based Model for Speech Enhancement标题：NADiffuSE：基于噪声感知扩散的语音增强模型链接：https://arxiv.org/abs/2309.01212作者：Wen Wang,Dongchao Yang,Qichen Ye,Bowen Cao,Yuexian Zou【27】 MAGMA: Music Aligned Generative Motion Autodecoder标题：Magma：音乐对准的生成式运动自动解码器链接：https://arxiv.org/abs/2309.01202作者：Sohan Anisetty,Amit Raj,James Hays【28】 Federated Few-shot Learning for Cough Classification with Edge Devices标题：基于边缘设备的联合小概率学习用于咳嗽分类链接：https://arxiv.org/abs/2309.01076作者：Ngan Dao Hoang,Dat Tran-Anh,Manh Luong,Cong Tran,Cuong Pham备注：21 pages, 5 figures【29】 Timbre-reserved Adversarial Attack in Speaker Identification标题：说话人辨认中保留音色的对抗性攻击链接：https://arxiv.org/abs/2309.00929作者：Qing Wang,Jixun Yao,Li Zhang,Pengcheng Guo,Lei Xie备注：11 pages, 8 figures【30】 BLSP: Bootstrapping Language-Speech Pre-training via Behavior Alignment of Continuation Writing标题：BLSP：通过续写行为对齐的自举语言--言语预训练链接：https://arxiv.org/abs/2309.00916作者：Chen Wang,Minpeng Liao,Zhongqiang Huang,Jinliang Lu,Junhong Wu,Yuchen Liu,Chengqing Zong,Jiajun Zhang【31】 DiCLET-TTS: Diffusion Model based Cross-lingual Emotion Transfer for Text-to-Speech -- A Study between English and Mandarin标题：DiCLET-TTS：基于扩散模型的文语转换跨语言情感传递--基于英语和汉语的对比研究链接：https://arxiv.org/abs/2309.00883作者：Tao Li,Chenxu Hu,Jian Cong,Xinfa Zhu,Jingbei Li,Qiao Tian,Yuping Wang,Lei Xie备注：accepted by TASLP【32】 Pretraining Representations for Bioacoustic Few-shot Detection using Supervised Contrastive Learning标题：基于有监督对比学习的生物声小射击检测前训练表示法链接：https://arxiv.org/abs/2309.00878作者：Ilyass Moummad,Romain Serizel,Nicolas Farrugia【33】 Contextual Biasing of Named-Entities with Large Language Models标题：使用大型语言模型的命名实体的上下文偏向链接：https://arxiv.org/abs/2309.00723作者：Chuanneng Sun,Zeeshan Ahmed,Yingyi Ma,Zhe Liu,Yutong Pang,Ozlem Kalinli备注：5 pages, 4 figures. Conference: ICASSP 2024机器翻译，仅供参考
每日语音速递
