今日论文合集：cs.SD语音11篇，eess.AS音频处理11篇。
cs.SD语音
【1】 TranSTYLer: Multimodal Behavioral Style Transfer for Facial and Body  Gestures Generation标题：TranSTYLer：用于面部和身体手势生成的多通道行为风格转换链接：https://arxiv.org/abs/2308.10843
作者：Mireille Fares,Catherine Pelachaud,Nicolas Obin摘要：本文讨论了一个虚拟代理的行为表现力风格转移到另一个，同时保留行为的形状，因为他们进行沟通的意义的挑战。行为表现风格在这里被看作是行为的定性属性。我们提出TranSTYLER，一个多模态Transformer为基础的模型，合成的多模态行为的源扬声器与目标扬声器的风格。我们假设行为表现风格是编码在各种形式的通信，包括文本，语音，身体姿势和面部表情。该模型采用了一种风格和内容的解纠缠模式，以确保传输的风格不干扰源行为所传达的含义。我们的方法消除了对样式标签的需要，并允许泛化到在训练阶段没有看到的样式。我们在PATS语料库上训练我们的模型，我们扩展到包括对话行为和2D面部标志。客观和主观的评价表明，我们的模型优于国家的最先进的模型在训练过程中看到的和看不见的风格的风格转移。为了解决可能出现的样式和内容泄漏的问题，我们提出了一种方法来评估与目标样式相关联的行为和手势被成功转移的程度，同时确保与源内容相关的行为和手势的保存。摘要：This paper addresses the challenge of transferring the behavior expressivity style of a virtual agent to another one while preserving behaviors shape as they carry communicative meaning. Behavior expressivity style is viewed here as the qualitative properties of behaviors. We propose TranSTYLer, a multimodal transformer based model that synthesizes the multimodal behaviors of a source speaker with the style of a target speaker. We assume that behavior expressivity style is encoded across various modalities of communication, including text, speech, body gestures, and facial expressions. The model employs a style and content disentanglement schema to ensure that the transferred style does not interfere with the meaning conveyed by the source behaviors. Our approach eliminates the need for style labels and allows the generalization to styles that have not been seen during the training phase. We train our model on the PATS corpus, which we extended to include dialog acts and 2D facial landmarks. Objective and subjective evaluations show that our model outperforms state of the art models in style transfer for both seen and unseen styles during training. To tackle the issues of style and content leakage that may arise, we propose a methodology to assess the degree to which behavior and gestures associated with the target style are successfully transferred, while ensuring the preservation of the ones related to the source content.
【2】 LibriWASN: A Data Set for Meeting Separation, Diarization, and  Recognition with Asynchronous Recording Devices标题：LibriWASN：一种用于满足异步记录设备的分离、二值化和识别的数据集链接：https://arxiv.org/abs/2308.10682
作者：Joerg Schmalenstroeer,Tobias Gburrek,Reinhold Haeb-Umbach备注：Accepted for presentation at the ITG conference on Speech Communication 2023摘要：我们提出LibriWASN，其设计密切遵循LibriCSS会议识别数据集的数据集，与显着的差异，数据记录的设备随机放置在会议桌上，其采样时钟不同步。九个不同的设备，五个智能手机，一个单一的录音通道和四个麦克风阵列，用于录制总共29个通道。除此之外，数据集严格遵循LibriCSS设计：从围绕会议桌布置的八个扬声器回放相同的LibriSpeech句子，并且数据被组织成具有不同百分比的语音重叠的子集。LibriWASN旨在作为时钟同步算法的测试集，满足ad-hoc无线声学传感器网络上的分离、拨号和转录系统。由于它与LibriCSS相似，为前者开发的会议转录系统可以很容易地在LibriWASN上测试。数据集记录在两个不同的房间，并补充了地面实况日记信息，谁说话时。摘要：We present LibriWASN, a data set whose design follows closely the LibriCSS meeting recognition data set, with the marked difference that the data is recorded with devices that are randomly positioned on a meeting table and whose sampling clocks are not synchronized. Nine different devices, five smartphones with a single recording channel and four microphone arrays, are used to record a total of 29 channels. Other than that, the data set follows closely the LibriCSS design: the same LibriSpeech sentences are played back from eight loudspeakers arranged around a meeting table and the data is organized in subsets with different percentages of speech overlap. LibriWASN is meant as a test set for clock synchronization algorithms, meeting separation, diarization and transcription systems on ad-hoc wireless acoustic sensor networks. Due to its similarity to LibriCSS, meeting transcription systems developed for the former can readily be tested on LibriWASN. The data set is recorded in two different rooms and is complemented with ground-truth diarization information of who speaks when.
【3】 An Anchor-Point Based Image-Model for Room Impulse Response Simulation  with Directional Source Radiation and Sensor Directivity Patterns标题：基于锚点的定向源辐射和传感器方向图室内冲击响应仿真图像模型链接：https://arxiv.org/abs/2308.10543
作者：Chao Pan,Lei Zhang,Yilong Lu,Jilu Jin,Lin Qiu,Jingdong Chen,Jacob Benesty备注：19 pages, 8 figures摘要：图像模型方法已被广泛用于模拟房间的脉冲响应和努力，以适应不同的应用程序，这种方法也激起了极大的兴趣，在过去的几十年。本文试图扩展的图像模型方法，并开发了一个锚点图像模型（APIM）的方法作为一个解决方案，包括源辐射和传感器的方向性图案模拟脉冲响应。为了确定所有虚拟源的取向，将锚点引入到真实源，这随后导致确定虚拟源的取向。通过考虑方向图函数、分数时延以及计算复杂度，提出了一种利用APIM生成房间脉冲响应的算法。所开发的模型和算法可用于各种声学问题，以模拟房间声学和改进和评估处理算法。摘要：The image model method has been widely used to simulate room impulse responses and the endeavor to adapt this method to different applications has also piqued great interest over the last few decades. This paper attempts to extend the image model method and develops an anchor-point-image-model (APIM) approach as a solution for simulating impulse responses by including both the source radiation and sensor directivity patterns. To determine the orientations of all the virtual sources, anchor points are introduced to real sources, which subsequently lead to the determination of the orientations of the virtual sources. An algorithm is developed to generate room impulse responses with APIM by taking into account the directional pattern functions, factional time delays, as well as the computational complexity. The developed model and algorithms can be used in various acoustic problems to simulate room acoustics and improve and evaluate processing algorithms.
【4】 TokenSplit: Using Discrete Speech Representations for Direct, Refined,  and Transcript-Conditioned Speech Separation and Recognition标题：TokenSplit：使用离散语音表示进行直接、细化和文字条件语音分离和识别链接：https://arxiv.org/abs/2308.10415
作者：Hakan Erdogan,Scott Wisdom,Xuankai Chang,Zalán Borsos,Marco Tagliasacchi,Neil Zeghidour,John R. Hershey备注：INTERSPEECH 2023, project webpage with audio demos at this https URL摘要：我们提出TokenSplit，一个语音分离模型，作用于离散令牌序列。该模型同时对多个任务进行训练：分离和转录每个语音源，并从文本生成语音。该模型对转录和音频令牌序列进行操作，并通过输入的掩蔽实现多个任务。该模型是使用Transformer架构的序列到序列编码器-解码器模型。我们还提出了一个“细化”版本的模型，该模型预测增强的音频令牌从语音的音频令牌由传统的分离模型分离。使用客观指标和主观MUSHRA听力测试，我们表明，我们的模型在分离方面取得了优异的性能，无论有或没有转录条件。我们还测量了自动语音识别（ASR）的性能，并提供语音合成的音频样本，以证明我们的模型的额外效用。摘要：We present TokenSplit, a speech separation model that acts on discrete token sequences. The model is trained on multiple tasks simultaneously: separate and transcribe each speech source, and generate speech from text. The model operates on transcripts and audio token sequences and achieves multiple tasks through masking of inputs. The model is a sequence-to-sequence encoder-decoder model that uses the Transformer architecture. We also present a "refinement" version of the model that predicts enhanced audio tokens from the audio tokens of speech separated by a conventional separation model. Using both objective metrics and subjective MUSHRA listening tests, we show that our model achieves excellent performance in terms of separation, both with or without transcript conditioning. We also measure the automatic speech recognition (ASR) performance and provide audio samples of speech synthesis to demonstrate the additional utility of our model.
【5】 Neural Architectures Learning Fourier Transforms, Signal Processing and  Much More....标题：学习傅立叶变换、信号处理和更多知识的神经体系结构...链接：https://arxiv.org/abs/2308.10388
作者：Prateek Verma备注：12 pages, 6 figures. Technical Report at Stanford University; Presented on 14th August 2023摘要：本报告将探讨和回答有关傅立叶变换的基本问题，并将其与人工智能和神经架构的最新进展联系起来。傅里叶变换的一种解释是通过将信号投影到复指数上来将信号分解为其组成分量。存在变体，诸如离散余弦变换，其不在复数域上操作并且将输入信号仅投影到在不同频率处振荡的余弦函数。然而，这是一个基本的限制，它需要更次优。第一个是所有的内核都是正弦的：如果我们可以根据问题调整或学习一些内核，会怎么样？如果我们可以使用神经架构来实现这一点呢？我们展示了如何从头开始学习这些内核的音频信号处理应用程序。我们发现，神经架构不仅学习正弦内核形状，还发现了各种令人难以置信的信号处理特性。例如，在一个示例中，窗口函数、起始检测器、高通滤波器、低通滤波器、调制等。此外，在分析滤波器时，我们发现神经架构在学习的内核之上具有梳状滤波器状结构。允许谐波频率通过的梳状滤波器是与各种传统信号处理算法的高通、低通和带通滤波器类似的核心构建块/类型的滤波器之一。此外，我们还可以使用卷积运算来从头开始学习信号，我们将探索文献中使用这种强大的Transformer架构的论文。此外，我们还将探索使学习的内核的内容自适应，即，针对不同的输入学习不同的内核。摘要：This report will explore and answer fundamental questions about taking Fourier Transforms and tying it with recent advances in AI and neural architecture. One interpretation of the Fourier Transform is decomposing a signal into its constituent components by projecting them onto complex exponentials. Variants exist, such as discrete cosine transform that does not operate on the complex domain and projects an input signal to only cosine functions oscillating at different frequencies. However, this is a fundamental limitation, and it needs to be more suboptimal. The first one is that all kernels are sinusoidal: What if we could have some kernels adapted or learned according to the problem? What if we can use neural architectures for this? We show how one can learn these kernels from scratch for audio signal processing applications. We find that the neural architecture not only learns sinusoidal kernel shapes but discovers all kinds of incredible signal-processing properties. E.g., windowing functions, onset detectors, high pass filters, low pass filters, modulations, etc. Further, upon analysis of the filters, we find that the neural architecture has a comb filter-like structure on top of the learned kernels. Comb filters that allow harmonic frequencies to pass through are one of the core building blocks/types of filters similar to high-pass, low-pass, and band-pass filters of various traditional signal processing algorithms. Further, we can also use the convolution operation with a signal to be learned from scratch, and we will explore papers in the literature that uses this with that robust Transformer architectures. Further, we would also explore making the learned kernel's content adaptive, i.e., learning different kernels for different inputs.
【6】 Spatial Reconstructed Local Attention Res2Net with F0 Subband for Fake  Speech Detection标题：F0子带空间重构局部注意Res2Net用于虚假语音检测链接：https://arxiv.org/abs/2308.09944
作者：Cunhang Fan,Jun Xue,Jianhua Tao,Jiangyan Yi,Chenglong Wang,Chengshi Zheng,Zhao Lv摘要：合成语音的节奏通常过于平滑，这导致合成语音的基频（F0）与真实语音的基频（F0）存在显著差异。预期F0特征包含用于假语音检测（FSD）任务的判别信息。在本文中，我们提出了一种新的F0子带FSD。此外，为了有效地建模F0子带，以提高FSD的性能，空间重构局部注意力Res 2Net（SR-LA Res 2Net）的建议。具体而言，Res 2Net被用作骨干网络，以获得多尺度信息，并增强了空间重构机制，以避免在信道组不断叠加时丢失重要信息。此外，局部注意力的设计，使该模型专注于F0子带的局部信息。在ASVspoof 2019 LA数据集上的实验结果表明，我们提出的方法获得了0.47%的等错误率（EER）和0.0159的最小串联检测成本函数（min t-DCF），在所有单个系统中实现了最先进的性能。摘要：The rhythm of synthetic speech is usually too smooth, which causes that the fundamental frequency (F0) of synthetic speech is significantly different from that of real speech. It is expected that the F0 feature contains the discriminative information for the fake speech detection (FSD) task. In this paper, we propose a novel F0 subband for FSD. In addition, to effectively model the F0 subband so as to improve the performance of FSD, the spatial reconstructed local attention Res2Net (SR-LA Res2Net) is proposed. Specifically, Res2Net is used as a backbone network to obtain multiscale information, and enhanced with a spatial reconstruction mechanism to avoid losing important information when the channel group is constantly superimposed. In addition, local attention is designed to make the model focus on the local information of the F0 subband. Experimental results on the ASVspoof 2019 LA dataset show that our proposed method obtains an equal error rate (EER) of 0.47% and a minimum tandem detection cost function (min t-DCF) of 0.0159, achieving the state-of-the-art performance among all of the single systems.
【7】 Implicit Self-supervised Language Representation for Spoken Language  Diarization标题：口语二元化的隐式自我监督语言表征链接：https://arxiv.org/abs/2308.10470
作者：Jagabandhu Mishra,S. R. Mahadeva Prasanna备注：Planning to Submit in IEEE-JSTSP摘要：在代码转换（CS）的场景中，使用口语diarization（LD）作为预拥有系统是必不可少的。此外，隐式框架的使用优于显式框架，因为它可以容易地适应于处理低/零资源语言。受说话人diarization（SD）文献的启发，提出了三种基于（1）固定分割，（2）基于变化点的分割和（3）E2 E的框架来执行LD。对合成TTSF-LD数据集的初步探索表明，使用x向量作为隐式语言表示，具有适当的分析窗口长度（$N$），可以实现在显式LD的每性能。最好的隐式LD性能为6.38美元的Jaccard错误率（JER）方面实现了使用E2 E框架。然而，考虑到E2 E框架，隐式LD的性能下降到$60.4$，而使用实际的Microsoft CS（MSCS）数据集。在性能上的差异主要是由于MSCS和TTSF-LD数据集中第二语言的单语段持续时间之间的分布差异。此外，为了避免片段平滑，单语片段的较小持续时间建议使用小值$N$。同时，在小$N$的情况下，由于声学相似性，x向量表示无法捕获所需的语言区分，因为同一说话者正在讲两种语言。因此，为了解决这个问题，本研究提出了一种自我监督的内隐语言表征。与x向量表示相比，所提出的表示提供了63.9美元的相对改进，并实现了21.8美元的JER使用E2 E框架。摘要：In a code-switched (CS) scenario, the use of spoken language diarization (LD) as a pre-possessing system is essential. Further, the use of implicit frameworks is preferable over the explicit framework, as it can be easily adapted to deal with low/zero resource languages. Inspired by speaker diarization (SD) literature, three frameworks based on (1) fixed segmentation, (2) change point-based segmentation and (3) E2E are proposed to perform LD. The initial exploration with synthetic TTSF-LD dataset shows, using x-vector as implicit language representation with appropriate analysis window length ($N$) can able to achieve at per performance with explicit LD. The best implicit LD performance of $6.38$ in terms of Jaccard error rate (JER) is achieved by using the E2E framework. However, considering the E2E framework the performance of implicit LD degrades to $60.4$ while using with practical Microsoft CS (MSCS) dataset. The difference in performance is mostly due to the distributional difference between the monolingual segment duration of secondary language in the MSCS and TTSF-LD datasets. Moreover, to avoid segment smoothing, the smaller duration of the monolingual segment suggests the use of a small value of $N$. At the same time with small $N$, the x-vector representation is unable to capture the required language discrimination due to the acoustic similarity, as the same speaker is speaking both languages. Therefore, to resolve the issue a self-supervised implicit language representation is proposed in this study. In comparison with the x-vector representation, the proposed representation provides a relative improvement of $63.9\%$ and achieved a JER of $21.8$ using the E2E framework.
【8】 Multi-GradSpeech: Towards Diffusion-based Multi-Speaker Text-to-speech  Using Consistent Diffusion Models标题：多等级语音：基于一致扩散模型的多说话人文语转换链接：https://arxiv.org/abs/2308.10428
作者：Heyang Xue,Shuai Guo,Pengcheng Zhu,Mengxiao Bi摘要：基于扩散的声学模型的最新进展已经彻底改变了数据充足的单说话人文本到语音（TTS）方法，其中Grad-TTS是一个主要的例子。然而，扩散模型遭受漂移的训练和采样分布由于不完善的分数匹配。采样漂移问题导致这些方法在实践中在多说话人场景中挣扎。在本文中，我们提出了多梯度语音，一个多说话人扩散为基础的声学模型，它引入了一致扩散模型（CDM）作为一种生成的建模方法。在训练过程中，我们加强了CDM的一致性，以缓解推理阶段的采样漂移问题，从而在多说话人TTS的性能显着提高。我们的实验结果证实，我们提出的方法可以提高不同的扬声器参与多扬声器TTS的性能相比，梯度TTS，甚至优于微调方法。音频样本可在https://welkinyang.github.io/multi-gradspeech/上获得摘要：Recent advancements in diffusion-based acoustic models have revolutionized data-sufficient single-speaker Text-to-Speech (TTS) approaches, with Grad-TTS being a prime example. However, diffusion models suffer from drift in training and sampling distributions due to imperfect score-matching. The sampling drift problem leads to these approaches struggling in multi-speaker scenarios in practice. In this paper, we present Multi-GradSpeech, a multi-speaker diffusion-based acoustic models which introduces the Consistent Diffusion Model (CDM) as a generative modeling approach. We enforce the consistency property of CDM during the training process to alleviate the sampling drift problem in the inference stage, resulting in significant improvements in multi-speaker TTS performance. Our experimental results corroborate that our proposed approach can improve the performance of different speakers involved in multi-speaker TTS compared to Grad-TTS, even outperforming the fine-tuning approach. Audio samples are available at https://welkinyang.github.io/multi-gradspeech/
【9】 Local Periodicity-Based Beat Tracking for Expressive Classical Piano  Music标题：基于局部周期性的表现力古典钢琴节拍跟踪链接：https://arxiv.org/abs/2308.10355
作者：Ching-Yu Chiu,Meinard Müller,Matthew E. P. Davies,Alvin Wen-Yu Su,Yi-Hsuan Yang备注：Accepted to IEEE/ACM Transactions on Audio, Speech, and Language Processing (July 2023)摘要：为了对节拍的周期性进行建模，现有技术的节拍跟踪系统使用“后处理跟踪器”（PPT），其依赖于用于节奏转变的若干经验确定的全局假设，这对于具有稳定节奏的音乐很好地工作。然而，对于富于表现力的古典音乐来说，这些假设可能过于严格。西方古典钢琴音乐的两个大的数据集，即对齐的分数和性能（ASAP）数据集和肖邦的马祖卡（Maz-5）的数据集，我们报告的实验表明，现有的PPT失败，以应付当地的节奏变化，从而呼吁新的方法。在本文中，我们提出了一个新的本地周期性为基础的PPT，主要的本地脉冲为基础的动态规划（PLPDP）跟踪，允许更灵活的节奏过渡。具体地，新PPT结合了称为“主导局部脉冲”（PLP）的方法与动态编程（DP）组件的组合，以在每个时刻联合考虑局部检测到的周期性和搏动激活强度。因此，PLPDP考虑了局部周期性，而不是依赖于全局节奏假设。与现有的PPT相比，PLPDP以较低的精度为代价特别增强了召回值，从而导致ASAP（从0.473到0.493）和Maz-5（从0.595到0.838）中用于搏动跟踪的F1分数的总体改善。摘要：To model the periodicity of beats, state-of-the-art beat tracking systems use "post-processing trackers" (PPTs) that rely on several empirically determined global assumptions for tempo transition, which work well for music with a steady tempo. For expressive classical music, however, these assumptions can be too rigid. With two large datasets of Western classical piano music, namely the Aligned Scores and Performances (ASAP) dataset and a dataset of Chopin's Mazurkas (Maz-5), we report on experiments showing the failure of existing PPTs to cope with local tempo changes, thus calling for new methods. In this paper, we propose a new local periodicity-based PPT, called predominant local pulse-based dynamic programming (PLPDP) tracking, that allows for more flexible tempo transitions. Specifically, the new PPT incorporates a method called "predominant local pulses" (PLP) in combination with a dynamic programming (DP) component to jointly consider the locally detected periodicity and beat activation strength at each time instant. Accordingly, PLPDP accounts for the local periodicity, rather than relying on a global tempo assumption. Compared to existing PPTs, PLPDP particularly enhances the recall values at the cost of a lower precision, resulting in an overall improvement of F1-score for beat tracking in ASAP (from 0.473 to 0.493) and Maz-5 (from 0.595 to 0.838).
【10】 The DKU-DUKEECE System for the Manipulation Region Location Task of ADD  2023标题：ADD 2023操作区定位任务的DKU-DUKECES系统链接：https://arxiv.org/abs/2308.10281
作者：Zexin Cai,Weiqing Wang,Yikang Wang,Ming Li备注：The DKU-DukeECE system description to Task 2 of Audio Deepfake Detection Challenge (ADD 2023)摘要：本文介绍了我们为Track 2设计的系统，该系统专注于定位操纵区域，在第二届Audio Deepfake Detection Challenge（ADD 2023）。我们的方法涉及利用多个检测系统来识别剪接区并确定其真实性。具体来说，我们训练和集成两个帧级系统：一个用于边界检测，另一个用于deepfake检测。此外，我们采用第三个VAE模型专门训练真实的数据，以确定给定的音频剪辑的真实性。通过这三个系统的融合，我们针对ADD挑战的最佳性能解决方案实现了令人印象深刻的82.23%的句子准确率和60.66%的F1得分。这导致最终ADD得分为0.6713，确保ADD 2023第二轨道中的第一名。摘要：This paper introduces our system designed for Track 2, which focuses on locating manipulated regions, in the second Audio Deepfake Detection Challenge (ADD 2023). Our approach involves the utilization of multiple detection systems to identify splicing regions and determine their authenticity. Specifically, we train and integrate two frame-level systems: one for boundary detection and the other for deepfake detection. Additionally, we employ a third VAE model trained exclusively on genuine data to determine the authenticity of a given audio clip. Through the fusion of these three systems, our top-performing solution for the ADD challenge achieves an impressive 82.23% sentence accuracy and an F1 score of 60.66%. This results in a final ADD score of 0.6713, securing the first rank in Track 2 of ADD 2023.
【11】 Effects of Convolutional Autoencoder Bottleneck Width on StarGAN-based  Singing Technique Conversion标题：卷积自动编码器瓶颈宽度对基于StarGan的演唱技术转换的影响链接：https://arxiv.org/abs/2308.10021
作者：Tung-Cheng Su,Yung-Chuan Chang,Yi-Wen Liu备注：The original edition of this paper will be published in the CMMR 2023 Proceedings. This ArXiv publication is a copy摘要：歌唱技术转换是指在保持原有歌唱者身份、旋律和语言成分不变的情况下，从一种发声技术转换为另一种发声技术的过程。以前的STC研究，以及歌唱声音转换研究一般，利用卷积自动编码器（CAE）的转换，但如何瓶颈宽度的CAE影响合成质量尚未得到彻底的评估。为此，我们构建了一个基于GAN的多域STC系统，该系统利用了WORLD声码器表示和CAE架构。我们改变了CAE的瓶颈宽度，并主观评价转换结果。该模型是在普通话数据集上训练的，该数据集具有四个歌手和四种歌唱技巧：胸部的声音，假声，刺耳的声音，和口哨声。结果表明，更宽的瓶颈对应于更好的清晰度，但不一定导致更高的相似性目标技术。在这四种技术中，我们还发现，口哨声是最容易转换的目标，而其他三种技术作为来源产生更令人信服的转换结果比口哨。摘要：Singing technique conversion (STC) refers to the task of converting from one voice technique to another while leaving the original singer identity, melody, and linguistic components intact. Previous STC studies, as well as singing voice conversion research in general, have utilized convolutional autoencoders (CAEs) for conversion, but how the bottleneck width of the CAE affects the synthesis quality has not been thoroughly evaluated. To this end, we constructed a GAN-based multi-domain STC system which took advantage of the WORLD vocoder representation and the CAE architecture. We varied the bottleneck width of the CAE, and evaluated the conversion results subjectively. The model was trained on a Mandarin dataset which features four singers and four singing techniques: the chest voice, the falsetto, the raspy voice, and the whistle voice. The results show that a wider bottleneck corresponds to better articulation clarity but does not necessarily lead to higher likeness to the target technique. Among the four techniques, we also found that the whistle voice is the easiest target for conversion, while the other three techniques as a source produce more convincing conversion results than the whistle.
eess.AS音频处理
【1】 Implicit Self-supervised Language Representation for Spoken Language  Diarization标题：口语二元化的隐式自我监督语言表征链接：https://arxiv.org/abs/2308.10470
作者：Jagabandhu Mishra,S. R. Mahadeva Prasanna备注：Planning to Submit in IEEE-JSTSP摘要：在代码转换（CS）的场景中，使用口语diarization（LD）作为预拥有系统是必不可少的。此外，隐式框架的使用优于显式框架，因为它可以容易地适应于处理低/零资源语言。受说话人diarization（SD）文献的启发，提出了三种基于（1）固定分割，（2）基于变化点的分割和（3）E2 E的框架来执行LD。对合成TTSF-LD数据集的初步探索表明，使用x向量作为隐式语言表示，具有适当的分析窗口长度（$N$），可以实现在显式LD的每性能。最好的隐式LD性能为6.38美元的Jaccard错误率（JER）方面实现了使用E2 E框架。然而，考虑到E2 E框架，隐式LD的性能下降到$60.4$，而使用实际的Microsoft CS（MSCS）数据集。在性能上的差异主要是由于MSCS和TTSF-LD数据集中第二语言的单语段持续时间之间的分布差异。此外，为了避免片段平滑，单语片段的较小持续时间建议使用小值$N$。同时，在小$N$的情况下，由于声学相似性，x向量表示无法捕获所需的语言区分，因为同一说话者正在讲两种语言。因此，为了解决这个问题，本研究提出了一种自我监督的内隐语言表征。与x向量表示相比，所提出的表示提供了63.9美元的相对改进，并实现了21.8美元的JER使用E2 E框架。摘要：In a code-switched (CS) scenario, the use of spoken language diarization (LD) as a pre-possessing system is essential. Further, the use of implicit frameworks is preferable over the explicit framework, as it can be easily adapted to deal with low/zero resource languages. Inspired by speaker diarization (SD) literature, three frameworks based on (1) fixed segmentation, (2) change point-based segmentation and (3) E2E are proposed to perform LD. The initial exploration with synthetic TTSF-LD dataset shows, using x-vector as implicit language representation with appropriate analysis window length ($N$) can able to achieve at per performance with explicit LD. The best implicit LD performance of $6.38$ in terms of Jaccard error rate (JER) is achieved by using the E2E framework. However, considering the E2E framework the performance of implicit LD degrades to $60.4$ while using with practical Microsoft CS (MSCS) dataset. The difference in performance is mostly due to the distributional difference between the monolingual segment duration of secondary language in the MSCS and TTSF-LD datasets. Moreover, to avoid segment smoothing, the smaller duration of the monolingual segment suggests the use of a small value of $N$. At the same time with small $N$, the x-vector representation is unable to capture the required language discrimination due to the acoustic similarity, as the same speaker is speaking both languages. Therefore, to resolve the issue a self-supervised implicit language representation is proposed in this study. In comparison with the x-vector representation, the proposed representation provides a relative improvement of $63.9\%$ and achieved a JER of $21.8$ using the E2E framework.
【2】 Multi-GradSpeech: Towards Diffusion-based Multi-Speaker Text-to-speech  Using Consistent Diffusion Models标题：多等级语音：基于一致扩散模型的多说话人文语转换链接：https://arxiv.org/abs/2308.10428
作者：Heyang Xue,Shuai Guo,Pengcheng Zhu,Mengxiao Bi摘要：基于扩散的声学模型的最新进展已经彻底改变了数据充足的单说话人文本到语音（TTS）方法，其中Grad-TTS是一个主要的例子。然而，扩散模型遭受漂移的训练和采样分布由于不完善的分数匹配。采样漂移问题导致这些方法在实践中在多说话人场景中挣扎。在本文中，我们提出了多梯度语音，一个多说话人扩散为基础的声学模型，它引入了一致扩散模型（CDM）作为一种生成的建模方法。在训练过程中，我们加强了CDM的一致性，以缓解推理阶段的采样漂移问题，从而在多说话人TTS的性能显着提高。我们的实验结果证实，我们提出的方法可以提高不同的扬声器参与多扬声器TTS的性能相比，梯度TTS，甚至优于微调方法。音频样本可在https://welkinyang.github.io/multi-gradspeech/上获得摘要：Recent advancements in diffusion-based acoustic models have revolutionized data-sufficient single-speaker Text-to-Speech (TTS) approaches, with Grad-TTS being a prime example. However, diffusion models suffer from drift in training and sampling distributions due to imperfect score-matching. The sampling drift problem leads to these approaches struggling in multi-speaker scenarios in practice. In this paper, we present Multi-GradSpeech, a multi-speaker diffusion-based acoustic models which introduces the Consistent Diffusion Model (CDM) as a generative modeling approach. We enforce the consistency property of CDM during the training process to alleviate the sampling drift problem in the inference stage, resulting in significant improvements in multi-speaker TTS performance. Our experimental results corroborate that our proposed approach can improve the performance of different speakers involved in multi-speaker TTS compared to Grad-TTS, even outperforming the fine-tuning approach. Audio samples are available at https://welkinyang.github.io/multi-gradspeech/
【3】 Local Periodicity-Based Beat Tracking for Expressive Classical Piano  Music标题：基于局部周期性的表现力古典钢琴节拍跟踪链接：https://arxiv.org/abs/2308.10355
作者：Ching-Yu Chiu,Meinard Müller,Matthew E. P. Davies,Alvin Wen-Yu Su,Yi-Hsuan Yang备注：Accepted to IEEE/ACM Transactions on Audio, Speech, and Language Processing (July 2023)摘要：为了对节拍的周期性进行建模，现有技术的节拍跟踪系统使用“后处理跟踪器”（PPT），其依赖于用于节奏转变的若干经验确定的全局假设，这对于具有稳定节奏的音乐很好地工作。然而，对于富于表现力的古典音乐来说，这些假设可能过于严格。西方古典钢琴音乐的两个大的数据集，即对齐的分数和性能（ASAP）数据集和肖邦的马祖卡（Maz-5）的数据集，我们报告的实验表明，现有的PPT失败，以应付当地的节奏变化，从而呼吁新的方法。在本文中，我们提出了一个新的本地周期性为基础的PPT，主要的本地脉冲为基础的动态规划（PLPDP）跟踪，允许更灵活的节奏过渡。具体地，新PPT结合了称为“主导局部脉冲”（PLP）的方法与动态编程（DP）组件的组合，以在每个时刻联合考虑局部检测到的周期性和搏动激活强度。因此，PLPDP考虑了局部周期性，而不是依赖于全局节奏假设。与现有的PPT相比，PLPDP以较低的精度为代价特别增强了召回值，从而导致ASAP（从0.473到0.493）和Maz-5（从0.595到0.838）中用于搏动跟踪的F1分数的总体改善。摘要：To model the periodicity of beats, state-of-the-art beat tracking systems use "post-processing trackers" (PPTs) that rely on several empirically determined global assumptions for tempo transition, which work well for music with a steady tempo. For expressive classical music, however, these assumptions can be too rigid. With two large datasets of Western classical piano music, namely the Aligned Scores and Performances (ASAP) dataset and a dataset of Chopin's Mazurkas (Maz-5), we report on experiments showing the failure of existing PPTs to cope with local tempo changes, thus calling for new methods. In this paper, we propose a new local periodicity-based PPT, called predominant local pulse-based dynamic programming (PLPDP) tracking, that allows for more flexible tempo transitions. Specifically, the new PPT incorporates a method called "predominant local pulses" (PLP) in combination with a dynamic programming (DP) component to jointly consider the locally detected periodicity and beat activation strength at each time instant. Accordingly, PLPDP accounts for the local periodicity, rather than relying on a global tempo assumption. Compared to existing PPTs, PLPDP particularly enhances the recall values at the cost of a lower precision, resulting in an overall improvement of F1-score for beat tracking in ASAP (from 0.473 to 0.493) and Maz-5 (from 0.595 to 0.838).
【4】 The DKU-DUKEECE System for the Manipulation Region Location Task of ADD  2023标题：ADD 2023操作区定位任务的DKU-DUKECES系统链接：https://arxiv.org/abs/2308.10281
作者：Zexin Cai,Weiqing Wang,Yikang Wang,Ming Li备注：The DKU-DukeECE system description to Task 2 of Audio Deepfake Detection Challenge (ADD 2023)摘要：本文介绍了我们为Track 2设计的系统，该系统专注于定位操纵区域，在第二届Audio Deepfake Detection Challenge（ADD 2023）。我们的方法涉及利用多个检测系统来识别剪接区并确定其真实性。具体来说，我们训练和集成两个帧级系统：一个用于边界检测，另一个用于deepfake检测。此外，我们采用第三个VAE模型专门训练真实的数据，以确定给定的音频剪辑的真实性。通过这三个系统的融合，我们针对ADD挑战的最佳性能解决方案实现了令人印象深刻的82.23%的句子准确率和60.66%的F1得分。这导致最终ADD得分为0.6713，确保ADD 2023第二轨道中的第一名。摘要：This paper introduces our system designed for Track 2, which focuses on locating manipulated regions, in the second Audio Deepfake Detection Challenge (ADD 2023). Our approach involves the utilization of multiple detection systems to identify splicing regions and determine their authenticity. Specifically, we train and integrate two frame-level systems: one for boundary detection and the other for deepfake detection. Additionally, we employ a third VAE model trained exclusively on genuine data to determine the authenticity of a given audio clip. Through the fusion of these three systems, our top-performing solution for the ADD challenge achieves an impressive 82.23% sentence accuracy and an F1 score of 60.66%. This results in a final ADD score of 0.6713, securing the first rank in Track 2 of ADD 2023.
【5】 Effects of Convolutional Autoencoder Bottleneck Width on StarGAN-based  Singing Technique Conversion标题：卷积自动编码器瓶颈宽度对基于StarGan的演唱技术转换的影响链接：https://arxiv.org/abs/2308.10021
作者：Tung-Cheng Su,Yung-Chuan Chang,Yi-Wen Liu备注：The original edition of this paper will be published in the CMMR 2023 Proceedings. This ArXiv publication is a copy摘要：歌唱技术转换是指在保持原有歌唱者身份、旋律和语言成分不变的情况下，从一种发声技术转换为另一种发声技术的过程。以前的STC研究，以及歌唱声音转换研究一般，利用卷积自动编码器（CAE）的转换，但如何瓶颈宽度的CAE影响合成质量尚未得到彻底的评估。为此，我们构建了一个基于GAN的多域STC系统，该系统利用了WORLD声码器表示和CAE架构。我们改变了CAE的瓶颈宽度，并主观评价转换结果。该模型是在普通话数据集上训练的，该数据集具有四个歌手和四种歌唱技巧：胸部的声音，假声，刺耳的声音，和口哨声。结果表明，更宽的瓶颈对应于更好的清晰度，但不一定导致更高的相似性目标技术。在这四种技术中，我们还发现，口哨声是最容易转换的目标，而其他三种技术作为来源产生更令人信服的转换结果比口哨。摘要：Singing technique conversion (STC) refers to the task of converting from one voice technique to another while leaving the original singer identity, melody, and linguistic components intact. Previous STC studies, as well as singing voice conversion research in general, have utilized convolutional autoencoders (CAEs) for conversion, but how the bottleneck width of the CAE affects the synthesis quality has not been thoroughly evaluated. To this end, we constructed a GAN-based multi-domain STC system which took advantage of the WORLD vocoder representation and the CAE architecture. We varied the bottleneck width of the CAE, and evaluated the conversion results subjectively. The model was trained on a Mandarin dataset which features four singers and four singing techniques: the chest voice, the falsetto, the raspy voice, and the whistle voice. The results show that a wider bottleneck corresponds to better articulation clarity but does not necessarily lead to higher likeness to the target technique. Among the four techniques, we also found that the whistle voice is the easiest target for conversion, while the other three techniques as a source produce more convincing conversion results than the whistle.
【6】 TranSTYLer: Multimodal Behavioral Style Transfer for Facial and Body  Gestures Generation标题：TranSTYLer：用于面部和身体手势生成的多通道行为风格转换链接：https://arxiv.org/abs/2308.10843
作者：Mireille Fares,Catherine Pelachaud,Nicolas Obin摘要：本文讨论了一个虚拟代理的行为表现力风格转移到另一个，同时保留行为的形状，因为他们进行沟通的意义的挑战。行为表现风格在这里被看作是行为的定性属性。我们提出TranSTYLER，一个多模态Transformer为基础的模型，合成的多模态行为的源扬声器与目标扬声器的风格。我们假设行为表现风格是编码在各种形式的通信，包括文本，语音，身体姿势和面部表情。该模型采用了一种风格和内容的解纠缠模式，以确保传输的风格不干扰源行为所传达的含义。我们的方法消除了对样式标签的需要，并允许泛化到在训练阶段没有看到的样式。我们在PATS语料库上训练我们的模型，我们扩展到包括对话行为和2D面部标志。客观和主观的评价表明，我们的模型优于国家的最先进的模型在训练过程中看到的和看不见的风格的风格转移。为了解决可能出现的样式和内容泄漏的问题，我们提出了一种方法来评估与目标样式相关联的行为和手势被成功转移的程度，同时确保与源内容相关的行为和手势的保存。摘要：This paper addresses the challenge of transferring the behavior expressivity style of a virtual agent to another one while preserving behaviors shape as they carry communicative meaning. Behavior expressivity style is viewed here as the qualitative properties of behaviors. We propose TranSTYLer, a multimodal transformer based model that synthesizes the multimodal behaviors of a source speaker with the style of a target speaker. We assume that behavior expressivity style is encoded across various modalities of communication, including text, speech, body gestures, and facial expressions. The model employs a style and content disentanglement schema to ensure that the transferred style does not interfere with the meaning conveyed by the source behaviors. Our approach eliminates the need for style labels and allows the generalization to styles that have not been seen during the training phase. We train our model on the PATS corpus, which we extended to include dialog acts and 2D facial landmarks. Objective and subjective evaluations show that our model outperforms state of the art models in style transfer for both seen and unseen styles during training. To tackle the issues of style and content leakage that may arise, we propose a methodology to assess the degree to which behavior and gestures associated with the target style are successfully transferred, while ensuring the preservation of the ones related to the source content.
【7】 LibriWASN: A Data Set for Meeting Separation, Diarization, and  Recognition with Asynchronous Recording Devices标题：LibriWASN：一种用于满足异步记录设备的分离、二值化和识别的数据集链接：https://arxiv.org/abs/2308.10682
作者：Joerg Schmalenstroeer,Tobias Gburrek,Reinhold Haeb-Umbach备注：Accepted for presentation at the ITG conference on Speech Communication 2023摘要：我们提出LibriWASN，其设计密切遵循LibriCSS会议识别数据集的数据集，与显着的差异，数据记录的设备随机放置在会议桌上，其采样时钟不同步。九个不同的设备，五个智能手机，一个单一的录音通道和四个麦克风阵列，用于录制总共29个通道。除此之外，数据集严格遵循LibriCSS设计：从围绕会议桌布置的八个扬声器回放相同的LibriSpeech句子，并且数据被组织成具有不同百分比的语音重叠的子集。LibriWASN旨在作为时钟同步算法的测试集，满足ad-hoc无线声学传感器网络上的分离、拨号和转录系统。由于它与LibriCSS相似，为前者开发的会议转录系统可以很容易地在LibriWASN上测试。数据集记录在两个不同的房间，并补充了地面实况日记信息，谁说话时。摘要：We present LibriWASN, a data set whose design follows closely the LibriCSS meeting recognition data set, with the marked difference that the data is recorded with devices that are randomly positioned on a meeting table and whose sampling clocks are not synchronized. Nine different devices, five smartphones with a single recording channel and four microphone arrays, are used to record a total of 29 channels. Other than that, the data set follows closely the LibriCSS design: the same LibriSpeech sentences are played back from eight loudspeakers arranged around a meeting table and the data is organized in subsets with different percentages of speech overlap. LibriWASN is meant as a test set for clock synchronization algorithms, meeting separation, diarization and transcription systems on ad-hoc wireless acoustic sensor networks. Due to its similarity to LibriCSS, meeting transcription systems developed for the former can readily be tested on LibriWASN. The data set is recorded in two different rooms and is complemented with ground-truth diarization information of who speaks when.
【8】 An Anchor-Point Based Image-Model for Room Impulse Response Simulation  with Directional Source Radiation and Sensor Directivity Patterns标题：基于锚点的定向源辐射和传感器方向图室内冲击响应仿真图像模型链接：https://arxiv.org/abs/2308.10543
作者：Chao Pan,Lei Zhang,Yilong Lu,Jilu Jin,Lin Qiu,Jingdong Chen,Jacob Benesty备注：19 pages, 8 figures摘要：图像模型方法已被广泛用于模拟房间的脉冲响应和努力，以适应不同的应用程序，这种方法也激起了极大的兴趣，在过去的几十年。本文试图扩展的图像模型方法，并开发了一个锚点图像模型（APIM）的方法作为一个解决方案，包括源辐射和传感器的方向性图案模拟脉冲响应。为了确定所有虚拟源的取向，将锚点引入到真实源，这随后导致确定虚拟源的取向。通过考虑方向图函数、分数时延以及计算复杂度，提出了一种利用APIM生成房间脉冲响应的算法。所开发的模型和算法可用于各种声学问题，以模拟房间声学和改进和评估处理算法。摘要：The image model method has been widely used to simulate room impulse responses and the endeavor to adapt this method to different applications has also piqued great interest over the last few decades. This paper attempts to extend the image model method and develops an anchor-point-image-model (APIM) approach as a solution for simulating impulse responses by including both the source radiation and sensor directivity patterns. To determine the orientations of all the virtual sources, anchor points are introduced to real sources, which subsequently lead to the determination of the orientations of the virtual sources. An algorithm is developed to generate room impulse responses with APIM by taking into account the directional pattern functions, factional time delays, as well as the computational complexity. The developed model and algorithms can be used in various acoustic problems to simulate room acoustics and improve and evaluate processing algorithms.
【9】 TokenSplit: Using Discrete Speech Representations for Direct, Refined,  and Transcript-Conditioned Speech Separation and Recognition标题：TokenSplit：使用离散语音表示进行直接、细化和文字条件语音分离和识别链接：https://arxiv.org/abs/2308.10415
作者：Hakan Erdogan,Scott Wisdom,Xuankai Chang,Zalán Borsos,Marco Tagliasacchi,Neil Zeghidour,John R. Hershey备注：INTERSPEECH 2023, project webpage with audio demos at this https URL摘要：我们提出TokenSplit，一个语音分离模型，作用于离散令牌序列。该模型同时对多个任务进行训练：分离和转录每个语音源，并从文本生成语音。该模型对转录和音频令牌序列进行操作，并通过输入的掩蔽实现多个任务。该模型是使用Transformer架构的序列到序列编码器-解码器模型。我们还提出了一个“细化”版本的模型，该模型预测增强的音频令牌从语音的音频令牌由传统的分离模型分离。使用客观指标和主观MUSHRA听力测试，我们表明，我们的模型在分离方面取得了优异的性能，无论有或没有转录条件。我们还测量了自动语音识别（ASR）的性能，并提供语音合成的音频样本，以证明我们的模型的额外效用。摘要：We present TokenSplit, a speech separation model that acts on discrete token sequences. The model is trained on multiple tasks simultaneously: separate and transcribe each speech source, and generate speech from text. The model operates on transcripts and audio token sequences and achieves multiple tasks through masking of inputs. The model is a sequence-to-sequence encoder-decoder model that uses the Transformer architecture. We also present a "refinement" version of the model that predicts enhanced audio tokens from the audio tokens of speech separated by a conventional separation model. Using both objective metrics and subjective MUSHRA listening tests, we show that our model achieves excellent performance in terms of separation, both with or without transcript conditioning. We also measure the automatic speech recognition (ASR) performance and provide audio samples of speech synthesis to demonstrate the additional utility of our model.
【10】 Neural Architectures Learning Fourier Transforms, Signal Processing and  Much More....标题：学习傅立叶变换、信号处理和更多知识的神经体系结构...链接：https://arxiv.org/abs/2308.10388
作者：Prateek Verma备注：12 pages, 6 figures. Technical Report at Stanford University; Presented on 14th August 2023摘要：本报告将探讨和回答有关傅立叶变换的基本问题，并将其与人工智能和神经架构的最新进展联系起来。傅里叶变换的一种解释是通过将信号投影到复指数上来将信号分解为其组成分量。存在变体，诸如离散余弦变换，其不在复数域上操作并且将输入信号仅投影到在不同频率处振荡的余弦函数。然而，这是一个基本的限制，它需要更次优。第一个是所有的内核都是正弦的：如果我们可以根据问题调整或学习一些内核，会怎么样？如果我们可以使用神经架构来实现这一点呢？我们展示了如何从头开始学习这些内核的音频信号处理应用程序。我们发现，神经架构不仅学习正弦内核形状，还发现了各种令人难以置信的信号处理特性。例如，在一个示例中，窗口函数、起始检测器、高通滤波器、低通滤波器、调制等。此外，在分析滤波器时，我们发现神经架构在学习的内核之上具有梳状滤波器状结构。允许谐波频率通过的梳状滤波器是与各种传统信号处理算法的高通、低通和带通滤波器类似的核心构建块/类型的滤波器之一。此外，我们还可以使用卷积运算来从头开始学习信号，我们将探索文献中使用这种强大的Transformer架构的论文。此外，我们还将探索使学习的内核的内容自适应，即，针对不同的输入学习不同的内核。摘要：This report will explore and answer fundamental questions about taking Fourier Transforms and tying it with recent advances in AI and neural architecture. One interpretation of the Fourier Transform is decomposing a signal into its constituent components by projecting them onto complex exponentials. Variants exist, such as discrete cosine transform that does not operate on the complex domain and projects an input signal to only cosine functions oscillating at different frequencies. However, this is a fundamental limitation, and it needs to be more suboptimal. The first one is that all kernels are sinusoidal: What if we could have some kernels adapted or learned according to the problem? What if we can use neural architectures for this? We show how one can learn these kernels from scratch for audio signal processing applications. We find that the neural architecture not only learns sinusoidal kernel shapes but discovers all kinds of incredible signal-processing properties. E.g., windowing functions, onset detectors, high pass filters, low pass filters, modulations, etc. Further, upon analysis of the filters, we find that the neural architecture has a comb filter-like structure on top of the learned kernels. Comb filters that allow harmonic frequencies to pass through are one of the core building blocks/types of filters similar to high-pass, low-pass, and band-pass filters of various traditional signal processing algorithms. Further, we can also use the convolution operation with a signal to be learned from scratch, and we will explore papers in the literature that uses this with that robust Transformer architectures. Further, we would also explore making the learned kernel's content adaptive, i.e., learning different kernels for different inputs.
【11】 Spatial Reconstructed Local Attention Res2Net with F0 Subband for Fake  Speech Detection标题：F0子带空间重构局部注意Res2Net用于虚假语音检测链接：https://arxiv.org/abs/2308.09944
作者：Cunhang Fan,Jun Xue,Jianhua Tao,Jiangyan Yi,Chenglong Wang,Chengshi Zheng,Zhao Lv摘要：合成语音的节奏通常过于平滑，这导致合成语音的基频（F0）与真实语音的基频（F0）存在显著差异。预期F0特征包含用于假语音检测（FSD）任务的判别信息。在本文中，我们提出了一种新的F0子带FSD。此外，为了有效地建模F0子带，以提高FSD的性能，空间重构局部注意力Res 2Net（SR-LA Res 2Net）的建议。具体而言，Res 2Net被用作骨干网络，以获得多尺度信息，并增强了空间重构机制，以避免在信道组不断叠加时丢失重要信息。此外，局部注意力的设计，使该模型专注于F0子带的局部信息。在ASVspoof 2019 LA数据集上的实验结果表明，我们提出的方法获得了0.47%的等错误率（EER）和0.0159的最小串联检测成本函数（min t-DCF），在所有单个系统中实现了最先进的性能。摘要：The rhythm of synthetic speech is usually too smooth, which causes that the fundamental frequency (F0) of synthetic speech is significantly different from that of real speech. It is expected that the F0 feature contains the discriminative information for the fake speech detection (FSD) task. In this paper, we propose a novel F0 subband for FSD. In addition, to effectively model the F0 subband so as to improve the performance of FSD, the spatial reconstructed local attention Res2Net (SR-LA Res2Net) is proposed. Specifically, Res2Net is used as a backbone network to obtain multiscale information, and enhanced with a spatial reconstruction mechanism to avoid losing important information when the channel group is constantly superimposed. In addition, local attention is designed to make the model focus on the local information of the F0 subband. Experimental results on the ASVspoof 2019 LA dataset show that our proposed method obtains an equal error rate (EER) of 0.47% and a minimum tandem detection cost function (min t-DCF) of 0.0159, achieving the state-of-the-art performance among all of the single systems.
机器翻译，仅供参考