上海普通话与标准普通话元音系统的声学对比研究

【作者】 于珏；

【导师】 李爱军； 马以容；

【作者基本信息】 浙江大学 ， 外国语言学及应用语言学， 2004， 硕士

【摘要】 普通话是现代汉民族共同语的俗称。它是以“北京语音为标准音，以北方话为基础方言，以典范的现代白话文著作为语法规范的”。但是汉语有9大方言区：官话，晋语，吴语，徽语，湘语，赣语，客家话，粤语和闽语。方言对普通话会有一定的影响，这种影响可以是各个层面的，有语音的，有词汇的，也有语法的，但以语音为最。因此来自不同方言地区的人讲普通话往往会或多或少带有口音。口音的这种影响不仅给普通话教学带来了困难，而且就言语工程(包括语音合成及语音识别)来说也是一个很大的挑战。事实上，口音的差异已经成为语音识别中的瓶颈问题，所以作汉语普通话和地方普通话的对比研究在现阶段是迫在眉睫的事。国内这方面的系统研究还没有，但是面向工程的语音学界已经开始积极参与这一问题的探索中，另外国外的相关研究有很多，这为我们也提供了很好的经验，但是汉语有其自身的特点，不能简单地套用国外的理论，所以还需要一段较长的自我摸索期。 本文结合当前这种实际研究和应用的需要，采用实验语音学的方法，以上海市的地方普通话为考察对象，通过对所用语料库中上海发音人口音的归类，语料的选取和标注工作，立足于上海话与普通话音系上的差异，分别对上海普通话与标准普通话基本元音系统，卷舌元音，二合元音系统，三合元音系统和带鼻尾的复合元音系统进行对比研究，画出各自的声学元音图及共振峰模式图，对比上海普通话与标准普通话各个元音的音值及共振峰模式，从而归纳这一地区发音人在音段上的共同特点及相关规律，为汉语口语处理口音问题提供可靠的声学参数，同时为普通话教学中如何消除口音问题提供借鉴。 论文的创新有四：一是在现有语音技术条件和汉语方言研究的基础上以一些重要城市和经济发达地区的方言点为基础建立地方语音语料库，并在此基础上应用大规模语料库的研究方法，通过对大量语料的统计分析，作地方普通话与标准普通话的对比研究，这是前人研究中所欠缺的；二是采用了语音学和言语工程技术相结合的研究思路，以语音学理论为基础，运用计算机语音处理技术来协助研究，对地方普通话与标准普通话音段层面的声学表现进行了系统的定量对比分析，并得到了一些普遍性较强的规律；三是画出男女声各元音的声学元音图及共振峰模式图，对比上海普通话与标准普通话各个元音的音值及共振峰模式；四是参考了大量国外的相关研究，积累了很好的经验，同时又从汉语自身的特点出发，不简单地套用国外的理论，完成了一个初级阶段的自我摸索期。 论文的不足之处在于只是单纯研究单念时的各元音在上海普通话与标准普通话中的差异，如果考虑自然语流的因素，情况可能会更复杂；另外本文对元音系统的差异性的描述还是非常的粗略，没有对各种声学数据作统计及听辨的测试，以论证其可靠性；而且各种声学数据可能还不够充分，这些在以后的研究中都需要进一步的分析探究。更多还原

【Abstract】 Spoken Chinese comprises many regional varieties, called dialects. There are 9 dialectal areas in China: Guan, Jin, Wu, Hui, Xiang, Gan, Kejia, Yue and Min. Guan (Mandarin) is referred to as a common language which covers a very large regional area from the northeast to the southwest of China, with over 800 million speakers. Most Chinese speak one of the Guan (Mandarin) dialects, which are largely mutually intelligible. The dialect spoken in Beijing constitutes the basis for Standard Chinese. It forms the basis both of the modern written vernacular, Balhua, which supplanted classical Chinese in the schools after 1917, and of the official spoken language, Putonghua, prescribed in 1956 for nationwide use in schools. Nowadays Standard Chinese (Putonghua, hereafter referred to as SC or Mandarin) is widely used throughout China in almost every activity from news broadcasting to commercial trades.People from different dialectal areas might not be able to communicate with each other simply because the differences among the dialects are so significant. Mandarin, or Putonghua, would be a good choice as a means of communication. Most people in China are bilingual Chinese speakers, i.e. they can speak their native dialect and Mandarin. Although lots of people CAN speak Mandarin, they speak it with different accents, depending on how well they grasp the language. The Mandarin they speak is always affected by their native dialects phonetically, lexically and syntactically.Wu dialect is a group of dialects spoken in Shanghai, Zhejlang, southern Jiangsu, and part of Fujian and Anhui. Wu dialect has about 70 million speakers, which makes it the second biggest dialect running after Mandarin. The dialect of interest in this paper is Shanghainese, the native dialect spoken in Shanghai covering more than 11,850,000 populations. Although it is rather young in Wu dialect family, Shanghainese becomes more and more appealing to researchers because of its economical and political importance.Dialectal differences are widely investigated for dialect identification, language (L2)learning and pronunciation modeling for Automatic Speech Recognition (ASR). Especially in Chinese ASR systems, how to deal with and tackle the accent issue becomes a big challenge due to the variability of the language. We hope the contrastive study from the phonetic point of view on regional accented Mandarin will shed a light on the Chinese ASR framework.In this paper we will mainly focus our contrastive study on segmental acoustic features between Standard (SC) and Shanghai-accented Mandarin (ASH). Two corpora were involved. The first database used in this study was the SPEECON Mandarin Chinese speech database collected by Nokia in the framework of SpeeCon (http://www.speecon.com). It covered spoken Chinese in four regional accents. The speech of 50 Beijing speakers and 51 Shanghai speakers recorded in office environment was selected for contrastive study. Each speaker had 321 utterances including phonetically rich words and sentences, application commands, proper names, numbers, time expressions and spontaneous speech.Another corpus used was NOKIA-CASS, a special speech corpus designed and recorded for comparison study for three major regions: Shanghai, Suzhou and Ningbo. Here only the speech corpus for Shanghai part was used. It included: monosyllabic words including all zero-initial monosyllables bearing four tones; disyllabic words with above features as well as covering possible syntax; 1500 long sentences with phonetic balance; and 56 statements and 56 Yes-No question consisting of disyllables with the combination of the low tone and the high tone.All the recorded utterances were phonetically annotated on orthographic and pronunciation tiers. Pronunciation variables or phonemic changes caused by dialects were annotated dedicatedly. Phonetic annotation was made for initials and finals of each syllable with time alignment. Prosodic and stress structure were annotated by C-ToBI (Li, 2002). Those extended annotations were only made for 20更多还原