今日论文合集：cs.SD语音3篇，eess.AS音频处理3篇。
cs.SD语音

【1】 Context-aware Coherent Speaking Style Prediction with Hierarchical Transformers for Audiobook Speech Synthesis标题：用于有声读物语音合成的基于分层变换的上下文感知连贯发声风格预测链接：http://arxiv.org/pdf/2304.06359v1

作者：Shun Lei,Yixuan Zhou,Liyang Chen,Zhiyong Wu,Shiyin Kang,Helen Meng机构：Shenzhen International Graduate School, Tsinghua University, Shenzhen , Peng Cheng Lab, Shenzhen, XVerse Inc., Shenzhen , The Chinese University of Hong Kong, Hong Kong SAR备注：Accepted by ICASSP 2023摘要：文本到语音的最新进展显着提高了合成语音的表达能力。然而，它仍然是具有挑战性的生成语音与上下文适当的和连贯的说话风格的多句子文本的有声读物。在本文中，我们提出了一个上下文感知的连贯说话风格预测方法的有声读物语音合成。为了预测当前话语的风格嵌入，设计了一种基于分层变换的上下文感知风格预测器，该预测器具有混合注意掩模，同时考虑文本侧上下文信息和先前语音的语音侧风格信息。在此基础上，我们可以逐句生成风格连贯、韵律清晰的长格式语音。对普通话有声读物数据集的客观和主观评价表明，我们提出的模型可以生成比基线更具表现力和连贯性的说话风格的语音，无论是单句还是多句测试。摘要：Recent advances in text-to-speech have significantly improved the expressiveness of synthesized speech. However, it is still challenging to generate speech with contextually appropriate and coherent speaking style for multi-sentence text in audiobooks. In this paper, we propose a context-aware coherent speaking style prediction method for audiobook speech synthesis. To predict the style embedding of the current utterance, a hierarchical transformer-based context-aware style predictor with a mixture attention mask is designed, considering both text-side context information and speech-side style information of previous speeches. Based on this, we can generate long-form speech with coherent style and prosody sentence by sentence. Objective and subjective evaluations on a Mandarin audiobook dataset demonstrate that our proposed model can generate speech with more expressive and coherent speaking style than baselines, for both single-sentence and multi-sentence test.

【2】 Brain Connectivity Features-based Age Group Classification using Temporal Asynchrony Audio-Visual Integration Task标题：基于脑连通性特征的时态异步性视听整合任务年龄组分类链接：http://arxiv.org/pdf/2304.06315v1

作者：Prerna Singh,Ayush Tripathi,Lalan Kumar,Tapan Kumar Gandhi机构：in 3Lalan Kumar is with the Department of Electrical Engineering, BhartiSchool of Telecommunication Technology and Management, and YardiSchool of Artificial Intelligence, Indian Institute of Technology Delhi摘要：来自人脑中的几种感觉模态的输入的整合过程被称为多感觉整合。与年龄相关的认知能力下降导致大脑丧失构思多感官输入的能力。在研究老年群体的这种认知变化方面已经做了大量的工作。然而，就中年群体而言，这种分析是有限的。受此启发，在目前的工作中，基于脑电图的功能连接性视听时间异步整合任务的中年群体进行了探索。在不同任务期间进行了调查，例如：单峰音频、单峰视觉、以及视听刺激的变化。进行基于相关性的功能连接性分析，不同年龄组之间的变化包括：观察到年轻（18-25岁）、从年轻到中年的过渡（25-33岁）和中等（33-41岁）。此外，从连接图中提取的特征已被用于在不同年龄组之间进行分类。$89.4 %$和$88.4 %$的音频和音频-50-视觉刺激的情况下，与随机森林为基础的分类器，从而验证了所提出的方法的有效性。摘要：The process of integration of inputs from several sensory modalities in the human brain is referred to as multisensory integration. Age-related cognitive decline leads to a loss in the ability of the brain to conceive multisensory inputs. There has been considerable work done in the study of such cognitive changes for the old age groups. However, in the case of middle age groups, such analysis is limited. Motivated by this, in the current work, EEG-based functional connectivity during audiovisual temporal asynchrony integration task for middle-aged groups is explored. Investigation has been carried out during different tasks such as: unimodal audio, unimodal visual, and variations of audio-visual stimulus. A correlation-based functional connectivity analysis is done, and the changes among different age groups including: young (18-25 years), transition from young to middle age (25-33 years), and medium (33-41 years), are observed. Furthermore, features extracted from the connectivity graphs have been used to classify among the different age groups. Classification accuracies of $89.4 %$ and $88.4 %$ are obtained for the Audio and Audio-50-Visual stimuli cases with a Random Forest based classifier, thereby validating the efficacy of the proposed method.

【3】 Acoustic absement in detail: Quantifying acoustic differences across time-series representations of speech data标题：详细的声学缺失：量化语音数据的时间序列表示的声学差异链接：http://arxiv.org/pdf/2304.06183v1

作者：Matthew C. Kelley机构：Department of Linguistics, University of Washington备注：5 pages, 1 figure, accepted for ICPhS 2023摘要：语音信号是时间序列数据的完美示例。信号的声学效果随时间变化，有时变化很大。然而，我们在语音学中执行的最常见类型的比较是瞬时声学测量之间的比较，例如共振峰值。在本文中，我讨论的概念absement作为两个时间序列之间的差异的量化。然后，我提供了一个实验的例子absement应用到语音分析的人类和 或计算机语音识别。实验是基于模板的语音识别任务，使用动态时间规整来比较孤立单词的录音之间的声学效果。识别准确率为57.9%。实验的结果进行了讨论，在使用absement作为一种工具，以及使用声学模型的口语单词识别的影响，作为最小的离散语言单元的单词。摘要：The speech signal is a consummate example of time-series data. The acoustics of the signal change over time, sometimes dramatically. Yet, the most common type of comparison we perform in phonetics is between instantaneous acoustic measurements, such as formant values. In the present paper, I discuss the concept of absement as a quantification of differences between two time-series. I then provide an experimental example of absement applied to phonetic analysis for human and or computer speech recognition. The experiment is a template-based speech recognition task, using dynamic time warping to compare the acoustics between recordings of isolated words. A recognition accuracy of 57.9% was achieved. The results of the experiment are discussed in terms of using absement as a tool, as well as the implications of using acoustics-only models of spoken word recognition with the word as the smallest discrete linguistic unit.
eess.AS音频处理

【1】 Brain Connectivity Features-based Age Group Classification using Temporal Asynchrony Audio-Visual Integration Task标题：基于脑连通性特征的时态异步性视听整合任务年龄组分类链接：http://arxiv.org/pdf/2304.06315v1

*与cs.SD语音【2】为同一篇

作者：Prerna Singh,Ayush Tripathi,Lalan Kumar,Tapan Kumar Gandhi机构：in 3Lalan Kumar is with the Department of Electrical Engineering, BhartiSchool of Telecommunication Technology and Management, and YardiSchool of Artificial Intelligence, Indian Institute of Technology Delhi摘要：来自人脑中的几种感觉模态的输入的整合过程被称为多感觉整合。与年龄相关的认知能力下降导致大脑丧失构思多感官输入的能力。在研究老年群体的这种认知变化方面已经做了大量的工作。然而，就中年群体而言，这种分析是有限的。受此启发，在目前的工作中，基于脑电图的功能连接性视听时间异步整合任务的中年群体进行了探索。在不同任务期间进行了调查，例如：单峰音频、单峰视觉、以及视听刺激的变化。进行基于相关性的功能连接性分析，不同年龄组之间的变化包括：观察到年轻（18-25岁）、从年轻到中年的过渡（25-33岁）和中等（33-41岁）。此外，从连接图中提取的特征已被用于在不同年龄组之间进行分类。$89.4 %$和$88.4 %$的音频和音频-50-视觉刺激的情况下，与随机森林为基础的分类器，从而验证了所提出的方法的有效性。摘要：The process of integration of inputs from several sensory modalities in the human brain is referred to as multisensory integration. Age-related cognitive decline leads to a loss in the ability of the brain to conceive multisensory inputs. There has been considerable work done in the study of such cognitive changes for the old age groups. However, in the case of middle age groups, such analysis is limited. Motivated by this, in the current work, EEG-based functional connectivity during audiovisual temporal asynchrony integration task for middle-aged groups is explored. Investigation has been carried out during different tasks such as: unimodal audio, unimodal visual, and variations of audio-visual stimulus. A correlation-based functional connectivity analysis is done, and the changes among different age groups including: young (18-25 years), transition from young to middle age (25-33 years), and medium (33-41 years), are observed. Furthermore, features extracted from the connectivity graphs have been used to classify among the different age groups. Classification accuracies of $89.4 %$ and $88.4 %$ are obtained for the Audio and Audio-50-Visual stimuli cases with a Random Forest based classifier, thereby validating the efficacy of the proposed method.

【2】 Acoustic absement in detail: Quantifying acoustic differences across time-series representations of speech data标题：详细的声学缺失：量化语音数据的时间序列表示的声学差异链接：http://arxiv.org/pdf/2304.06183v1

*与cs.SD语音【3】为同一篇

作者：Matthew C. Kelley机构：Department of Linguistics, University of Washington备注：5 pages, 1 figure, accepted for ICPhS 2023摘要：语音信号是时间序列数据的完美示例。信号的声学效果随时间变化，有时变化很大。然而，我们在语音学中执行的最常见类型的比较是瞬时声学测量之间的比较，例如共振峰值。在本文中，我讨论的概念absement作为两个时间序列之间的差异的量化。然后，我提供了一个实验的例子absement应用到语音分析的人类和 或计算机语音识别。实验是基于模板的语音识别任务，使用动态时间规整来比较孤立单词的录音之间的声学效果。识别准确率为57.9%。实验的结果进行了讨论，在使用absement作为一种工具，以及使用声学模型的口语单词识别的影响，作为最小的离散语言单元的单词。摘要：The speech signal is a consummate example of time-series data. The acoustics of the signal change over time, sometimes dramatically. Yet, the most common type of comparison we perform in phonetics is between instantaneous acoustic measurements, such as formant values. In the present paper, I discuss the concept of absement as a quantification of differences between two time-series. I then provide an experimental example of absement applied to phonetic analysis for human and or computer speech recognition. The experiment is a template-based speech recognition task, using dynamic time warping to compare the acoustics between recordings of isolated words. A recognition accuracy of 57.9% was achieved. The results of the experiment are discussed in terms of using absement as a tool, as well as the implications of using acoustics-only models of spoken word recognition with the word as the smallest discrete linguistic unit.

【3】 Context-aware Coherent Speaking Style Prediction with Hierarchical Transformers for Audiobook Speech Synthesis标题：用于有声读物语音合成的基于分层变换的上下文感知连贯发声风格预测链接：http://arxiv.org/pdf/2304.06359v1

*与cs.SD语音【1】为同一篇

作者：Shun Lei,Yixuan Zhou,Liyang Chen,Zhiyong Wu,Shiyin Kang,Helen Meng机构：Shenzhen International Graduate School, Tsinghua University, Shenzhen , Peng Cheng Lab, Shenzhen, XVerse Inc., Shenzhen , The Chinese University of Hong Kong, Hong Kong SAR备注：Accepted by ICASSP 2023摘要：文本到语音的最新进展显着提高了合成语音的表达能力。然而，它仍然是具有挑战性的生成语音与上下文适当的和连贯的说话风格的多句子文本的有声读物。在本文中，我们提出了一个上下文感知的连贯说话风格预测方法的有声读物语音合成。为了预测当前话语的风格嵌入，设计了一种基于分层变换的上下文感知风格预测器，该预测器具有混合注意掩模，同时考虑文本侧上下文信息和先前语音的语音侧风格信息。在此基础上，我们可以逐句生成风格连贯、韵律清晰的长格式语音。对普通话有声读物数据集的客观和主观评价表明，我们提出的模型可以生成比基线更具表现力和连贯性的说话风格的语音，无论是单句还是多句测试。摘要：Recent advances in text-to-speech have significantly improved the expressiveness of synthesized speech. However, it is still challenging to generate speech with contextually appropriate and coherent speaking style for multi-sentence text in audiobooks. In this paper, we propose a context-aware coherent speaking style prediction method for audiobook speech synthesis. To predict the style embedding of the current utterance, a hierarchical transformer-based context-aware style predictor with a mixture attention mask is designed, considering both text-side context information and speech-side style information of previous speeches. Based on this, we can generate long-form speech with coherent style and prosody sentence by sentence. Objective and subjective evaluations on a Mandarin audiobook dataset demonstrate that our proposed model can generate speech with more expressive and coherent speaking style than baselines, for both single-sentence and multi-sentence test.

机器翻译，仅供参考