今日论文合集：cs.SD语音6篇，eess.AS音频处理7篇。本文经arXiv每日学术速递授权转载
【1】PromptCodec: High-Fidelity Neural Speech Codec using Disentangled Representation Learning based Adaptive Feature-aware Prompt Encoders
标题：基于解纠缠表示学习的高保真神经语音编解码器
链接：https://arxiv.org/abs/2404.02702
作者：Yu Pan,Lei Ma,Jianjun Zhao
备注：7
摘要：神经语音编解码器最近在生成语音建模领域获得了广泛的关注，如语音转换，文本到语音合成等，然而，确保在高压缩率下的语音编解码器的高保真音频重建仍然是一个开放的和具有挑战性的问题。在本文中，我们提出了一种新的端到端的神经语音编解码器模型，使用基于解纠缠表示学习的特征感知提示编码器。通过结合来自即时编码器的附加特征表示，MPTCodec可以分发需要处理的语音信息并增强其功能。此外，一个简单而有效的自适应特征加权融合方法被引入到不同的编码器集成功能。同时，我们提出了一种新的基于余弦距离的解纠缠表示学习策略，以优化MPTCodec的编码器，以确保其效率，从而进一步提高MPTCodec的性能。LibriTTS上的实验表明，我们提出的LectCodec在所有不同的比特率条件下始终优于最先进的神经语音编解码器模型，同时在低比特率下实现了令人印象深刻的性能。
摘要：Neural speech codec has recently gained widespread attention in generative speech modeling domains, like voice conversion, text-to-speech synthesis, etc. However, ensuring high-fidelity audio reconstruction of speech codecs under high compression rates remains an open and challenging issue. In this paper, we propose PromptCodec, a novel end-to-end neural speech codec model using disentangled representation learning based feature-aware prompt encoders. By incorporating additional feature representations from prompt encoders, PromptCodec can distribute the speech information requiring processing and enhance its capabilities. Moreover, a simple yet effective adaptive feature weighted fusion approach is introduced to integrate features of different encoders. Meanwhile, we propose a novel disentangled representation learning strategy based on cosine distance to optimize PromptCodec's encoders to ensure their efficiency, thereby further improving the performance of PromptCodec. Experiments on LibriTTS demonstrate that our proposed PromptCodec consistently outperforms state-of-the-art neural speech codec models under all different bitrate conditions while achieving impressive performance with low bitrates.
【2】 Leveraging the Interplay Between Syntactic and Acoustic Cues for Optimizing Korean TTS Pause Formation标题：利用句法和声学线索的相互作用优化韩语TTS的结构作者：Yejin Jeon,Yunsu Kim,Gary Geunbae Lee备注：Accepted to LREC-COLING 2024摘要：当代神经语音合成模型确实在合成语音生成方面表现出了非凡的能力，因为它们已经达到了与人类产生的语音相当的质量水平。然而，值得注意的是，这些成就主要是在英语等高资源语言的背景下得到验证的。此外，Tacotron和FastSpeech变体在应用于韩语时显示出大量的停顿错误，这影响了语音感知和自然度。为了解决上述问题，我们提出了一个新的框架，结合了全面的建模与暂停模式相关的句法和声学线索。值得注意的是，我们的框架具有持续生成自然语音的能力，即使是对于更扩展和复杂的域外（OOD）句子，尽管它是在短音频片段上训练的。通过使用主观和客观指标与基线模型和消融研究进行比较，确认了结构设计选择，从而确认了模型性能。摘要：Contemporary neural speech synthesis models have indeed demonstrated remarkable proficiency in synthetic speech generation as they have attained a level of quality comparable to that of human-produced speech. Nevertheless, it is important to note that these achievements have predominantly been verified within the context of high-resource languages such as English. Furthermore, the Tacotron and FastSpeech variants show substantial pausing errors when applied to the Korean language, which affects speech perception and naturalness. In order to address the aforementioned issues, we propose a novel framework that incorporates comprehensive modeling of both syntactic and acoustic cues that are associated with pausing patterns. Remarkably, our framework possesses the capability to consistently generate natural speech even for considerably more extended and intricate out-of-domain (OOD) sentences, despite its training on short audio clips. Architectural design choices are validated through comparisons with baseline models and ablation studies using subjective and objective metrics, thus confirming model performance.
【3】 PhonologyBench: Evaluating Phonological Skills of Large Language Models标题：PhonologyBench：评估大型语言模型的语音技巧作者：Ashima Suvarna,Harshita Khandelwal,Nanyun Peng备注：17 pages, 7 figures, 6 tables摘要：音系学是对语音结构和发音规则的研究，是大型语言模型（LLM）研究中一个重要但经常被忽视的组成部分。LLM被广泛用于利用语音学的各种下游应用，如教育工具和诗歌生成。此外，LLM可以潜在地从训练数据中学习正字法和音韵形式之间的不完美关联。因此，有必要对LLM的语音技能进行基准测试。为此，我们提出了PhonologyBench，一个新的基准，包括三个诊断任务，旨在明确测试的语音技能的LLM在英语：字形到音素转换，音节计数，和押韵词生成。尽管无法访问语音数据，LLM在PhonologyBench任务中表现出了显着的表现。然而，与人类相比，我们观察到押韵词生成和音节计数分别有17%和45%的显着差距。我们的研究结果强调了研究LLM在无意中影响现实世界应用的语音任务上的表现的重要性。此外，我们鼓励研究人员选择在与下游应用程序密切相关的语音任务上表现良好的LLM，因为我们发现没有一个模型在所有任务上都表现得优于其他模型。摘要：Phonology, the study of speech's structure and pronunciation rules, is a critical yet often overlooked component in Large Language Model (LLM) research. LLMs are widely used in various downstream applications that leverage phonology such as educational tools and poetry generation. Moreover, LLMs can potentially learn imperfect associations between orthographic and phonological forms from the training data. Thus, it is imperative to benchmark the phonological skills of LLMs. To this end, we present PhonologyBench, a novel benchmark consisting of three diagnostic tasks designed to explicitly test the phonological skills of LLMs in English: grapheme-to-phoneme conversion, syllable counting, and rhyme word generation. Despite having no access to speech data, LLMs showcased notable performance on the PhonologyBench tasks. However, we observe a significant gap of 17% and 45% on Rhyme Word Generation and Syllable counting, respectively, when compared to humans. Our findings underscore the importance of studying LLM performance on phonological tasks that inadvertently impact real-world applications. Furthermore, we encourage researchers to choose LLMs that perform well on the phonological task that is closely related to the downstream application since we find that no single model consistently outperforms the others on all the tasks.
【4】 A Computational Analysis of Lyric Similarity Perception作者：Haven Kim,Taketo Akama摘要：在包含声乐的音乐作品中，歌词对艺术表现有着重要的作用。因此，先前的研究已经引入了推荐系统的概念，该推荐系统建议与用户的最爱或个性化偏好相似的歌词，以帮助在数百万首曲目中发现歌词。然而，这些系统中的许多系统并没有充分考虑人类对歌词相似性的感知，这主要是由于该领域的研究有限。为了弥合这一差距，我们进行了比较分析的计算方法与人类感知建模抒情相似性。结果表明，基于预训练的基于BERT的模型的嵌入之间的相似性的计算模型，歌词所来源的音频和语音成分指示感知歌词相似性。这一发现强调了语义，风格和语音的相似性，在人类感知的歌词相似性的重要性。我们预计，我们的研究结果将通过为神经网络开发提供伪标签和引入客观的评价指标来促进基于相似性的歌词推荐系统的发展。摘要：In musical compositions that include vocals, lyrics significantly contribute to artistic expression. Consequently, previous studies have introduced the concept of a recommendation system that suggests lyrics similar to a user's favorites or personalized preferences, aiding in the discovery of lyrics among millions of tracks. However, many of these systems do not fully consider human perceptions of lyric similarity, primarily due to limited research in this area. To bridge this gap, we conducted a comparative analysis of computational methods for modeling lyric similarity with human perception. Results indicated that computational models based on similarities between embeddings from pre-trained BERT-based models, the audio from which the lyrics are derived, and phonetic components are indicative of perceptual lyric similarity. This finding underscores the importance of semantic, stylistic, and phonetic similarities in human perception about lyric similarity. We anticipate that our findings will enhance the development of similarity-based lyric recommendation systems by offering pseudo-labels for neural network development and introducing objective evaluation metrics.
【5】 SMITIN: Self-Monitored Inference-Time INtervention for Generative Music Transformers标题：SMITIN：生成音乐变换器的自激励推理时间干扰作者：Junghyun Koo,Gordon Wichern,Francois G. Germain,Sameer Khurana,Jonathan Le Roux摘要：我们介绍了自回归推理时间干预（SMITIN），一种使用分类器探针控制自回归生成音乐Transformer的方法。这些简单的逻辑回归探测器在Transformer中的每个注意力头部的输出上使用表现出和缺失特定音乐特质（例如，鼓的存在/不存在，或真实/合成音乐）。然后，我们将注意力转向探测方向，确保生成模型输出捕捉到所需的音乐特征。此外，我们监控探头输出，以避免在自回归生成中添加过多的干预，这可能导致时间上不连贯的音乐。我们客观和主观地验证了我们的结果，音频延续和文本到音乐的应用程序，展示了添加控制的能力，以大型生成模型的再培训，甚至微调是不切实际的大多数音乐家。 建议的干预方法的音频样本可在我们的演示页面http://tinyurl.com/smitin上获得。摘要：We introduce Self-Monitored Inference-Time INtervention (SMITIN), an approach for controlling an autoregressive generative music transformer using classifier probes. These simple logistic regression probes are trained on the output of each attention head in the transformer using a small dataset of audio examples both exhibiting and missing a specific musical trait (e.g., the presence/absence of drums, or real/synthetic music). We then steer the attention heads in the probe direction, ensuring the generative model output captures the desired musical trait. Additionally, we monitor the probe output to avoid adding an excessive amount of intervention into the autoregressive generation, which could lead to temporally incoherent music. We validate our results objectively and subjectively for both audio continuation and text-to-music applications, demonstrating the ability to add controls to large generative models for which retraining or even fine-tuning is impractical for most musicians. Audio samples of the proposed intervention approach are available on our demo page http://tinyurl.com/smitin .
【6】 CLaM-TTS: Improving Neural Codec Language Model for Zero-Shot Text-to-Speech标题：CLaM—TTS：改进的Zero-Shot文本到语音的神经编解码语言模型作者：Jaehyeon Kim,Keon Lee,Seungjun Chung,Jaewoong Cho摘要：随着神经音频编解码器的出现，其编码来自音频的多个离散令牌流，大型语言模型最近作为zero-shot文本到语音（TTS）合成的有前途的方法而受到关注。尽管不断涌现出可扩展的范例，但具有讽刺意味的是，音频标记化放大了可扩展性的挑战，这源于其长序列长度和对多个序列建模的复杂性。为了缓解这些问题，我们提出了采用概率残差矢量量化的CLaM-TTS，以（1）实现令牌长度的卓越压缩，以及（2）允许语言模型一次生成多个令牌，从而消除了级联建模来处理令牌流数量的需要。我们的实验结果表明，CLaM-TTS是优于或相当于国家的最先进的神经编解码器为基础的TTS模型的自然度，可懂度，说话人的相似性和推理速度。此外，我们还研究了语言模型的预训练程度及其文本标记化策略对性能的影响。摘要：With the emergence of neural audio codecs, which encode multiple streams of discrete tokens from audio, large language models have recently gained attention as a promising approach for zero-shot Text-to-Speech (TTS) synthesis. Despite the ongoing rush towards scaling paradigms, audio tokenization ironically amplifies the scalability challenge, stemming from its long sequence length and the complexity of modelling the multiple sequences. To mitigate these issues, we present CLaM-TTS that employs a probabilistic residual vector quantization to (1) achieve superior compression in the token length, and (2) allow a language model to generate multiple tokens at once, thereby eliminating the need for cascaded modeling to handle the number of token streams. Our experimental results demonstrate that CLaM-TTS is better than or comparable to state-of-the-art neural codec-based TTS models regarding naturalness, intelligibility, speaker similarity, and inference speed. In addition, we examine the impact of the pretraining extent of the language models and their text tokenization strategies on performances.
【1】 CLaM-TTS: Improving Neural Codec Language Model for Zero-Shot Text-to-Speech标题：CLaM—TTS：改进的Zero-Shot文本到语音的神经编解码语言模型作者：Jaehyeon Kim,Keon Lee,Seungjun Chung,Jaewoong Cho摘要：随着神经音频编解码器的出现，其编码来自音频的多个离散令牌流，大型语言模型最近作为zero-shot文本到语音（TTS）合成的有前途的方法而受到关注。尽管不断涌现出可扩展的范例，但具有讽刺意味的是，音频标记化放大了可扩展性的挑战，这源于其长序列长度和对多个序列建模的复杂性。为了缓解这些问题，我们提出了采用概率残差矢量量化的CLaM-TTS，以（1）实现令牌长度的卓越压缩，以及（2）允许语言模型一次生成多个令牌，从而消除了级联建模来处理令牌流数量的需要。我们的实验结果表明，CLaM-TTS是优于或相当于国家的最先进的神经编解码器为基础的TTS模型的自然度，可懂度，说话人的相似性和推理速度。此外，我们还研究了语言模型的预训练程度及其文本标记化策略对性能的影响。摘要：With the emergence of neural audio codecs, which encode multiple streams of discrete tokens from audio, large language models have recently gained attention as a promising approach for zero-shot Text-to-Speech (TTS) synthesis. Despite the ongoing rush towards scaling paradigms, audio tokenization ironically amplifies the scalability challenge, stemming from its long sequence length and the complexity of modelling the multiple sequences. To mitigate these issues, we present CLaM-TTS that employs a probabilistic residual vector quantization to (1) achieve superior compression in the token length, and (2) allow a language model to generate multiple tokens at once, thereby eliminating the need for cascaded modeling to handle the number of token streams. Our experimental results demonstrate that CLaM-TTS is better than or comparable to state-of-the-art neural codec-based TTS models regarding naturalness, intelligibility, speaker similarity, and inference speed. In addition, we examine the impact of the pretraining extent of the language models and their text tokenization strategies on performances.
【2】 The VoicePrivacy 2024 Challenge Evaluation Plan标题：VoicePrivacy 2024挑战评估计划作者：Natalia Tomashenko,Xiaoxiao Miao,Pierre Champion,Sarina Meyer,Xin Wang,Emmanuel Vincent,Michele Panariello,Nicholas Evans,Junichi Yamagishi,Massimiliano Todisco备注：arXiv admin note: substantial text overlap with arXiv:2203.12468摘要：挑战的任务是开发一个语音数据的语音匿名化系统，隐藏说话者的语音身份，同时保护语言内容和情感状态。组织者提供开发和评价数据集和评价脚本，以及基线匿名化系统和根据参与者的请求编制的培训资源清单。参与者应用他们开发的匿名化系统，运行评估脚本，并将评估结果和匿名语音数据提交给组织者。结果将在与Interspeech 2024联合举办的研讨会上展示，邀请所有参与者介绍他们的挑战系统并提交额外的研讨会论文。摘要：The task of the challenge is to develop a voice anonymization system for speech data which conceals the speaker's voice identity while protecting linguistic content and emotional states. The organizers provide development and evaluation datasets and evaluation scripts, as well as baseline anonymization systems and a list of training resources formed on the basis of the participants' requests. Participants apply their developed anonymization systems, run evaluation scripts and submit evaluation results and anonymized speech data to the organizers. Results will be presented at a workshop held in conjunction with Interspeech 2024 to which all participants are invited to present their challenge systems and to submit additional workshop papers.【3】 ART: The Alternating Reading Task Corpus for Speech Entrainment and Imitation作者：Zheng Yuan,Dorina de Jong,Štefan Beňuš,Noël Nguyen,Ruitao Feng,Róbert Sabo,Luciano Fadiga,Alessandro D`Ausilio备注：15 pages, 2 figures, 7 tables, accepted at LREC-COLING 2024 conference摘要：我们介绍了交替阅读任务（ART）语料库，一个收集的二元句子阅读研究夹带和模仿行为在言语交际。ART语料库的特点是三个实验条件-独奏阅读，交替阅读，刻意模仿-以及三个子语料库，包括法语，意大利语和哈萨克口音的英语。这种设计允许在受控和自发性较低的环境中系统地研究语音夹带。除了详细的翻译，它包括英语水平分数，人口统计学，并在实验中的问卷调查，探讨语言，个人和人际关系的影响夹带。我们的演讲涵盖了它的设计，收集，注释过程，初步分析和未来的研究前景。摘要：We introduce the Alternating Reading Task (ART) Corpus, a collection of dyadic sentence reading for studying the entrainment and imitation behaviour in speech communication. The ART corpus features three experimental conditions - solo reading, alternating reading, and deliberate imitation - as well as three sub-corpora encompassing French-, Italian-, and Slovak-accented English. This design allows systematic investigation of speech entrainment in a controlled and less-spontaneous setting. Alongside detailed transcriptions, it includes English proficiency scores, demographics, and in-experiment questionnaires for probing linguistic, personal and interpersonal influences on entrainment. Our presentation covers its design, collection, annotation processes, initial analysis, and future research prospects.
【4】 Leveraging the Interplay Between Syntactic and Acoustic Cues for Optimizing Korean TTS Pause Formation标题：利用句法和声学线索的相互作用优化韩语TTS的结构作者：Yejin Jeon,Yunsu Kim,Gary Geunbae Lee备注：Accepted to LREC-COLING 2024摘要：当代神经语音合成模型确实在合成语音生成方面表现出了非凡的能力，因为它们已经达到了与人类产生的语音相当的质量水平。然而，值得注意的是，这些成就主要是在英语等高资源语言的背景下得到验证的。此外，Tacotron和FastSpeech变体在应用于韩语时显示出大量的停顿错误，这影响了语音感知和自然度。为了解决上述问题，我们提出了一个新的框架，结合了全面的建模与暂停模式相关的句法和声学线索。值得注意的是，我们的框架具有持续生成自然语音的能力，即使是对于更扩展和复杂的域外（OOD）句子，尽管它是在短音频片段上训练的。通过使用主观和客观指标与基线模型和消融研究进行比较，确认了结构设计选择，从而确认了模型性能。摘要：Contemporary neural speech synthesis models have indeed demonstrated remarkable proficiency in synthetic speech generation as they have attained a level of quality comparable to that of human-produced speech. Nevertheless, it is important to note that these achievements have predominantly been verified within the context of high-resource languages such as English. Furthermore, the Tacotron and FastSpeech variants show substantial pausing errors when applied to the Korean language, which affects speech perception and naturalness. In order to address the aforementioned issues, we propose a novel framework that incorporates comprehensive modeling of both syntactic and acoustic cues that are associated with pausing patterns. Remarkably, our framework possesses the capability to consistently generate natural speech even for considerably more extended and intricate out-of-domain (OOD) sentences, despite its training on short audio clips. Architectural design choices are validated through comparisons with baseline models and ablation studies using subjective and objective metrics, thus confirming model performance.
【5】 PhonologyBench: Evaluating Phonological Skills of Large Language Models标题：PhonologyBench：评估大型语言模型的语音技巧作者：Ashima Suvarna,Harshita Khandelwal,Nanyun Peng备注：17 pages, 7 figures, 6 tables摘要：音系学是对语音结构和发音规则的研究，是大型语言模型（LLM）研究中一个重要但经常被忽视的组成部分。LLM被广泛用于利用语音学的各种下游应用，如教育工具和诗歌生成。此外，LLM可以潜在地从训练数据中学习正字法和音韵形式之间的不完美关联。因此，有必要对LLM的语音技能进行基准测试。为此，我们提出了PhonologyBench，一个新的基准，包括三个诊断任务，旨在明确测试的语音技能的LLM在英语：字形到音素转换，音节计数，和押韵词生成。尽管无法访问语音数据，LLM在PhonologyBench任务中表现出了显着的表现。然而，与人类相比，我们观察到押韵词生成和音节计数分别有17%和45%的显着差距。我们的研究结果强调了研究LLM在无意中影响现实世界应用的语音任务上的表现的重要性。此外，我们鼓励研究人员选择在与下游应用程序密切相关的语音任务上表现良好的LLM，因为我们发现没有一个模型在所有任务上都表现得优于其他模型。摘要：Phonology, the study of speech's structure and pronunciation rules, is a critical yet often overlooked component in Large Language Model (LLM) research. LLMs are widely used in various downstream applications that leverage phonology such as educational tools and poetry generation. Moreover, LLMs can potentially learn imperfect associations between orthographic and phonological forms from the training data. Thus, it is imperative to benchmark the phonological skills of LLMs. To this end, we present PhonologyBench, a novel benchmark consisting of three diagnostic tasks designed to explicitly test the phonological skills of LLMs in English: grapheme-to-phoneme conversion, syllable counting, and rhyme word generation. Despite having no access to speech data, LLMs showcased notable performance on the PhonologyBench tasks. However, we observe a significant gap of 17% and 45% on Rhyme Word Generation and Syllable counting, respectively, when compared to humans. Our findings underscore the importance of studying LLM performance on phonological tasks that inadvertently impact real-world applications. Furthermore, we encourage researchers to choose LLMs that perform well on the phonological task that is closely related to the downstream application since we find that no single model consistently outperforms the others on all the tasks.
【6】 A Computational Analysis of Lyric Similarity Perception作者：Haven Kim,Taketo Akama摘要：在包含声乐的音乐作品中，歌词对艺术表现有着重要的作用。因此，先前的研究已经引入了推荐系统的概念，该推荐系统建议与用户的最爱或个性化偏好相似的歌词，以帮助在数百万首曲目中发现歌词。然而，这些系统中的许多系统并没有充分考虑人类对歌词相似性的感知，这主要是由于该领域的研究有限。为了弥合这一差距，我们进行了比较分析的计算方法与人类感知建模抒情相似性。结果表明，基于预训练的基于BERT的模型的嵌入之间的相似性的计算模型，歌词所来源的音频和语音成分指示感知歌词相似性。这一发现强调了语义，风格和语音的相似性，在人类感知的歌词相似性的重要性。我们预计，我们的研究结果将通过为神经网络开发提供伪标签和引入客观的评价指标来促进基于相似性的歌词推荐系统的发展。摘要：In musical compositions that include vocals, lyrics significantly contribute to artistic expression. Consequently, previous studies have introduced the concept of a recommendation system that suggests lyrics similar to a user's favorites or personalized preferences, aiding in the discovery of lyrics among millions of tracks. However, many of these systems do not fully consider human perceptions of lyric similarity, primarily due to limited research in this area. To bridge this gap, we conducted a comparative analysis of computational methods for modeling lyric similarity with human perception. Results indicated that computational models based on similarities between embeddings from pre-trained BERT-based models, the audio from which the lyrics are derived, and phonetic components are indicative of perceptual lyric similarity. This finding underscores the importance of semantic, stylistic, and phonetic similarities in human perception about lyric similarity. We anticipate that our findings will enhance the development of similarity-based lyric recommendation systems by offering pseudo-labels for neural network development and introducing objective evaluation metrics.
【7】 SMITIN: Self-Monitored Inference-Time INtervention for Generative Music Transformers标题：SMITIN：生成音乐变换器的自激励推理时间干扰作者：Junghyun Koo,Gordon Wichern,Francois G. Germain,Sameer Khurana,Jonathan Le Roux摘要：我们介绍了自回归推理时间干预（SMITIN），一种使用分类器探针控制自回归生成音乐Transformer的方法。这些简单的逻辑回归探测器在Transformer中的每个注意力头部的输出上使用表现出和缺失特定音乐特质（例如，鼓的存在/不存在，或真实/合成音乐）。然后，我们将注意力转向探测方向，确保生成模型输出捕捉到所需的音乐特征。此外，我们监控探头输出，以避免在自回归生成中添加过多的干预，这可能导致时间上不连贯的音乐。我们客观和主观地验证了我们的结果，音频延续和文本到音乐的应用程序，展示了添加控制的能力，以大型生成模型的再培训，甚至微调是不切实际的大多数音乐家。 建议的干预方法的音频样本可在我们的演示页面http://tinyurl.com/smitin上获得。摘要：We introduce Self-Monitored Inference-Time INtervention (SMITIN), an approach for controlling an autoregressive generative music transformer using classifier probes. These simple logistic regression probes are trained on the output of each attention head in the transformer using a small dataset of audio examples both exhibiting and missing a specific musical trait (e.g., the presence/absence of drums, or real/synthetic music). We then steer the attention heads in the probe direction, ensuring the generative model output captures the desired musical trait. Additionally, we monitor the probe output to avoid adding an excessive amount of intervention into the autoregressive generation, which could lead to temporally incoherent music. We validate our results objectively and subjectively for both audio continuation and text-to-music applications, demonstrating the ability to add controls to large generative models for which retraining or even fine-tuning is impractical for most musicians. Audio samples of the proposed intervention approach are available on our demo page http://tinyurl.com/smitin .
每日语音速递