首页 | 本学科首页   官方微博 | 高级检索  
     

听力正常人对正弦波汉语普通话的识别
引用本文:冯艳梅,徐立,周宁,杨光,殷善开. 听力正常人对正弦波汉语普通话的识别[J]. 中华耳科学杂志, 2012, 10(2): 178-181
作者姓名:冯艳梅  徐立  周宁  杨光  殷善开
作者单位:1. 上海交通大学附属第六人民医院耳鼻咽喉科上海市听力测试中心上海交通大学耳鼻咽喉科研究所 上海200233
2. 美国俄亥俄大学听力言语系 美国Ohio 45701
基金项目:国家自然科学基金(81070778);美国NIH/NIDCD基金(R15-DC009504)
摘    要:目的通过听力正常人对正弦波处理的汉语普通话的音调及句子识别,探讨正弦波汉语普通话的识别机制。方法测试包括正弦波处理的汉语四声调识别及句子识别。45名听力正常的年轻人,首先进行正弦波处理的汉语四声调及句子训练,然后进行测试。在训练过程中,给予受试者反馈,使其适应,学习正弦波处理的信号。正式测试中,不给予受试者反馈。声调识别测试中,受试者需要在四个声调中选择所听到的声调,在训练和测试过程中,每个声信号只能播放一次,以受试者答对的百分数计分。句子识别采用MHINT,受试者可以反复听测试材料,直至受试者完全识别正确或识别不出为止。以受试者答对的百分数进行评分。结果正弦波四声调识别的正确率平均值为32.6%,仅稍高于机率得分(25%)。受试者得分在23.5%~42.3%之间,其中绝大部分得分在25%至35%之间。在四声调识别中,一声的识别正确率最高,为42.3%;三声的识别正确率次之,为34.9%;二声和四声的识别正确率则均分别为23.5%和29.6%。正弦波汉语普通话句子识别的正确率平均值为92.3%,受试者得分在78%至100%之间,绝大部分在90%以上。配对t检验结果表明,正弦波汉语普通话四声调识别和句子识别的差异具有统计学显著性意义(P<0.05)。结论汉语普通话经过正弦波处理后,声调识别较差,句子识别很好,从上向下机制在正弦波汉语普通话句子识别中起重要作用。

关 键 词:正弦波  汉语普通话  言语识别  声调识别

The recognition of sine-wave Mandarin Chinese in normal-hearing listeners
FENG Yan-mei , XU Li , ZHOU Ning , YANG Guang , YIN Shan-kai. The recognition of sine-wave Mandarin Chinese in normal-hearing listeners[J]. Chinese Journal of Otology, 2012, 10(2): 178-181
Authors:FENG Yan-mei    XU Li    ZHOU Ning    YANG Guang    YIN Shan-kai
Affiliation:1 Department of Otolaryngology,Shanghai Sixth People’s Hospital,Institute of Otolaryngology,Shanghai Jiao Tong University,Shanghai,People’s Republic of China 200233 2 School of Rehabilitation and Communication Sciences,Ohio University,Athens,Ohio 45701,USA
Abstract:Objective To study the mechanism in recognition of sine-wave Mandarin Chinese tones and sentences in normal-hearing listeners.Method The testing involved recognition of tones and sentences processed with sine-wave replica.A practice session was given to all 45 normal-hearing subjects before formal test.Feedback was provided to the subjects during the practice session (but not during the formal test) to familiarize the subjects with the sine-wave replica.In tone recognition,the subjects were forced to choose from four tone tokens displayed on the screen when he/she heard a sine-wave tone token.The sine-wave tone tokens were presented only once.The correct percent was recorded from each subject.The Mandarin hearing in noise test (MHINT) was adopted for the sentence recognition of sine-wave replica.The subjects could listen to the sine-wave replica as many times as he/she wanted.The correct percent was recorded.Results The mean score for tone recognition of sine-wave replica was 32.6% which was only slightly higher than the chance level (25%).The scores ranged from 23.5% to 42.3%,with most scores ranging from 25% to 35%.In the tone recognition,tone 1 scored the highest (42.3% correct),followed by tone 3 (34.9% correct).Tone 2 and tone 4 scored 23.5% and 29.6% correct,respectively.The mean score for sentence recognition of sine-wave replica was 92.3%.The scores ranged from 78% to 100% correct,with most scores above 80% correct.Paired t test indicated that the difference between tone and sentence recognition of sine-wave replica was statistically significant (P<0.05).Conclusion The recognition of sine-wave Mandarin Chinese tones was poor but the recognition was nearly perfect for sentences.The top-down processes must play an important role in sine-wave Mandarin sentence recognition.
Keywords:Sine wave  Mandarin Chinese  Speech recognition  Tone recognition
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号