维吾尔语中疑问和焦点对语调的共同调节作用

通过严格控制的语音实验,系统研究了维吾尔语焦点和疑问语气对语调的调节作用。结果发现维吾尔语疑问句中焦点对音高和时长都有调节作用:1)焦点词音高上升、音域扩大,焦点后音高仍保持高,但焦点前的音高基本不变; 2)疑问语调的重要特征是句末音高出现大幅度上扬,另外疑问句中焦点词后的音高高于其在陈述句条件下的表现; 3)焦点成分时长延长,而焦点前后成分的时长变化不明显; 4)与陈述句相比,疑问句的总时长要长于陈述句的,且主要表现在句末成分的时长上。总之,维吾尔语和汉语、英语一样支持平行编码目标逼近(PENTA)语调模型,但是三种语言在疑问句中焦点后的语调变化方式上并不相同。     

汉语韵律短语的时长与音高研究

语句和篇章的韵律结构和信息结构的分析及模型化是提高语音合成的自然度、降低自然语言识别错误率的关键。该文在带有韵律标注ASCCD语料库的基础上对韵律短语的时长和音高特性进行了研究,得到并验证了如下一些结论:(1)韵律短语边界对音节时长有明显的延长作用,不同声调对音节的时长延长作用不同,并且不同的重音级别对音节时长的延长作用也不同。(2)韵律短语边界处中断的时长在较小的韵律边界表现的更为明显。韵律短语的边界处发生了明显的音高重置现象,韵律短语的音高低线总是下降的,而音高高线只是在重音后下降,并且重音处的音域大而且音高高线的位置高。     

汉语朗读话语重音自动分类研究

汉语的重音由于受到声调、语调以及韵律单元层级的干扰和制约,对于重音的自动感知一直是比较困难的问题。针对标准的朗读普通话语,本文在广义韵律结构的框架下研究了重音的声学表现,设计并实现了重音的自动感知模型。本文提出的基于分类树结构的区分度模型能有效地结合韵律单元结构对重音的制约。研究结果表明,音高高线、调域、音长是表达重音最重要线索,利用这些线索能有效地实现对重音的自动感知。我们的模型能一般能达到80 %左右的重音检出水平。     

基于语法信息的汉语韵律结构预测

韵律结构的预测, 主要包括短语的自动切分和重音的等级分布两个大的方面。本文在概述汉语韵律结构的基础上, 根据从自然话语中获得的韵律结构与句法结构和词性的关系, 用一种新的方法,通过文本分析,全面地预测韵律边界的位置分布及其等级差异,并进一步预测重音的位置分布及其等级差异。     

支持重音合成的汉语语音合成系统

针对基于单元挑选的汉语语音合成系统中重音预测及实现,本文采用了知识指导下的数据驱动建模策略。首先,采用经过感知结果优化的重音检测器,实现了语音数据库的自动标注;其次,利用重音标注数据库,训练得到支持重音预测的韵律预测模型;用重音韵律预测模型替代原语音合成系统中的相应模型,从而构成了支持重音合成的语音合成系统。实验结果分析表明,基于感知结果优化的重音检测器的标注结果是可靠的;支持重音的韵律声学预测模型是合理的;新的合成系统能够合成出带有轻重变化的语音。     

普通话焦点声学特征的实验研究

以自然语流中出现的焦点为对象,对汉语中焦点的声学特征表现进行了研究.研究结果表明:(1)焦点对音节韵律特征的影响与音节所在的高层韵律环境(上下文相关信息)密切相关.处于不同高层韵律环境的音节,其韵律特征受焦点影响改变的幅度和方向是不同的.(2)焦点的轻重感知一定程度上可以通过线性调节语音声学参数增量来表现出来.(3)在语音合成中,焦点的韵律特征可分为两步来进行预测.实验证实,在焦点位置已知的情况下该方法能够合成自然度很高的汉语语句焦点.     

焦点的韵律表达及认知加工研究综览

焦点是语言学界广泛关注的问题。随着实验语音学与心理语言学的发展,国内外对焦点的韵律表达及认知加工方面的研究发展迅速,主要涉及焦点的语音与音系表征、焦点与重音的对应关系,以及句子理解中焦点加工与韵律加工的大脑机制等问题。该文从这一角度对相关研究进行回顾与总结,介绍该领域的发展状况及主要研究方向并提出见解和评论,以期对今后的研究有所启发。     

基于互补模型的汉语重音检测

针对现有汉语重音检测方法正确率较低的问题,利用声学、词典和语法相关特征的不同分类器组合,基于Boosting分类回归树+条件随机场的互补模型,提出一种改进的汉语重音检测方法.在ASCCD语料库上的实验结果表明,该方法能获得84.9％的重音检测正确率,相比基于神经网络+决策树的基线系统提高2.7％.     

汉语名物性短语句法位置语料库的设计

汉语句物性短语(NP)在汉语信息检索中有重要价值。本文以非统计的信息处理方法为出发点,介绍一个汉语名物性短语句法位置语料库的设计思想、所使用的句法位置标记集以及标记加工规范,并指出了这样一个语料库的潜在价值。目前正在以此为出发点建立一个汉语名物性短语句法位置语料库。     

自然语流中二音节组重读的表现

1 引言在汉语文-语转换系统(CTTS)的研究中,要求系统合成的语音应尽量模拟自然语音,体现出语言抑扬顿挫的美感。这需要正确地界定韵律短语,细微地模拟自然语音的韵律模式。自然语音在字调、语调、重音等方面的变化,其声学特征的表现为音长、基频和幅度等时域参数。因此,在基于时域PSOLA的CTTS研究     

维吾尔语的重音检测

根据维吾尔语词重音的位置规律进行音节级标注,提取不同的特征参数（能量、基频等）,对各个特征参数作单流、分流及特征级融合识别实验,对比分析各特征参数对重音检测识别率的影响。对融合后的高维单流特征采用主成分分析作降维、去冗余处理,并作识别实验。参照识别精确率结合语音语言学知识对实验结果进行分析。     

话题转换方式和句子长度对边界声学参数的影响

该文研究了不同话题转换方式和句子长度对边界处停顿、边界前延长量及音高重置的影响。语料是由两个句子构成的小语篇,通过改变第二个句子控制两种句子长度(短和长)和三种话题转换方式(延续、精述和转折)。20位发音人的语音分析结果显示: (1) 话题转换方式和句子长度对停顿及音高重置都有调节作用,但对边界前词的时长延长量没有显著影响。另外,两因素间没有交互作用。主要表现为: 边界后句子越长,句间停顿越长,且边界处的音高重置越大。从话题延续、话题精述到话题转折,停顿时长呈增长趋势,且音高重置度增大; (2) 停顿时长与边界前延长量存在较弱的负相关,与音高重置则存在较弱的正相关; (3) 相较于男性发音人,女性发音人对话题转换方式更为敏感,且更倾向于用停顿和音高两种声学线索标记话题转换方式。句长效应则在男女发音人中都稳定存在。以上结果表明,句长对边界处声学参数的影响基于底层发音机制,而话题转换方式的影响则是语言中信息传递的需要。     

Prosody dependent speech recognition on radio news corpus of American English

Does prosody help word recognition? This paper proposes a novel probabilistic framework in which word and phoneme are dependent on prosody in a way that reduces word error rates (WER) relative to a prosody-independent recognizer with comparable parameter count. In the proposed prosody-dependent speech recognizer, word and phoneme models are conditioned on two important prosodic variables: the intonational phrase boundary and the pitch accent. An information-theoretic analysis is provided to show that prosody dependent acoustic and language modeling can increase the mutual information between the true word hypothesis and the acoustic observation by exciting the interaction between prosody dependent acoustic model and prosody dependent language model. Empirically, results indicate that the influence of these prosodic variables on allophonic models are mainly restricted to a small subset of distributions: the duration PDFs (modeled using an explicit duration hidden Markov model or EDHMM) and the acoustic-prosodic observation PDFs (normalized pitch frequency). Influence of prosody on cepstral features is limited to a subset of phonemes: for example, vowels may be influenced by both accent and phrase position, but phrase-initial and phrase-final consonants are independent of accent. Leveraging these results, effective prosody dependent allophonic models are built with minimal increase in parameter count. These prosody dependent speech recognizers are able to reduce word error rates by up to 11% relative to prosody independent recognizers with comparable parameter count, in experiments based on the prosodically-transcribed Boston Radio News corpus.     

Pragmatic determinants of intonation contours for dialogue systems

This paper describes an implemented computational model that generates intonation contours for dialogue systems. We concentrate on the relationship between pragmatics and two aspects of intonation: pitch range and pitch accent placement. Pitch range is computed based on the position of an utterance in the discourse structure: utterances that introduce a new topic have an expanded register compared to utterances that continue a topic. Pitch accent placement is based on two pragmatic factors: cognitive status (what the speaker assumes the hearer is attending to) and informativeness (what the speaker assumes to be the interesting or informative component of a phrase). This work suggests that even simple models of discourse topic structure, cognitive status, and informativeness will lead to improved register determination and pitch accent placement in practical conversational systems.     

An Automatic System for Detecting Prosodic Prominence in American English Continuous Speech

A precise identification of prosodic phenomena and the construction of tools able to properly manage such phenomena are essential steps to disambiguate the meaning of certain utterances. In particular they are useful for a wide variety of tasks: automatic recognition of spontaneous speech, automatic enhancement of speech-generation systems, solving ambiguities in natural language interpretation, the construction of large annotated language resources, such as prosodically tagged speech corpora, and teaching languages to foreign students using Computer Aided Language Learning (CALL) systems. This paper presents a study on the automatic detection of prosodic prominence in continuous speech, with particular reference to American English, but with good prospects of application to other languages. Prosodic prominence involves two different prosodic features: pitch accent and stress accent. Pitch accent is acoustically connected with fundamental frequency (F0) movements and overall syllable energy, whereas stress exhibits a strong correlation with syllable nuclei duration and mid-to-high-frequency emphasis. This paper shows that a careful measurement of these acoustic parameters, as well as the identification of their connection to prosodic parameters, makes it possible to build an automatic system capable of identifying prominent syllables in utterances with performance comparable with the inter-human agreement reported in the literature. Two different prominence detectors were studied and developed: the first uses a training corpus to set up thresholds properly, while the second uses a pure unsupervised method. In both cases, it is worth stressing that only acoustic parameters derived directly from speech waveforms are exploited.     

一种数据驱动的汉语成句语音韵律特征产生模型的研究

针对汉语发音特点，基于对大量自然汉语语句基频轮廓数据的统计和分析，提出一种用于数据驱动生成汉语韵律特征的数学模型。该模型以基频参数为主，辅以时长和增益参数，能表现汉语的语气、短语节奏、韵律词声调及轻重音多层韵律信忠，各层参数可按语言知识分类训练和标注。给出了模型的各种归一化“调素”函数和变调规则。仿真实验表明了该模型的有效性。     

基于数据挖掘的普通话韵律规则学习

普通话韵律规则对于语音合成和语音学研究具有重要意义。为了更有效地进行韵律规则学习,该文利用数据挖掘技术从语料库中的取规则。通过聚类分析进行基频模式提取,并以此进行基频序列的离散化;由语言学分析的结果得出训练句子中每个单节的参数,利用决策树和神经网络学习章节的韵律变化规则。测试表明基于数据挖掘的韵律规则学习取得了较好的结果,证实了方法的有效性。