摘要
目的通过探讨时域和频域信息对听力正常人声调识别的影响,试图为人工耳蜗汉语言语编码策略的编制提供理论依据。方法共有13名听力正常人参加本次研究,均以普通话作为日常交流方式。测试材料使用美国Ohio大学提供的声调识别测试系统。测试系统包含10组测试词。每组测试词的声、韵母相同而声调不同。将上述测试词录制成原始声音文件。采用声码器技术处理录制的原始声音。将决定时域信息的低通滤波器的截止频率设为1、2、4、8、16、32、64、128、256和512Hz,共10种;将决定频域信息的通道数设为1、2、4、6、8、12、16和20,共8种。经不同的排列组合,共产生6400个处理后的声音文件。测试在标准隔声室内进行。测试声音通过扬声器发出,强度为65dBSPL。声音的发出和成绩的记录均由Matlab软件自动完成。结果听力正常人的声调识别得分随着低通滤波的截止频率和通道数的增加而增加。当低通滤波的截止频率在2Hz以下时,是否增加通道数对声调识别得分基本没有影响。当通道数在2以下时,是否增加低通滤波的截止频率对声调识别得分基本没有影响。结论听力正常人的声调识别与时域和频域信息均相关。时域信息和频域信息在听力正常人的声调识别中的作用是协同的。
Objective The purpose of this study is to determine the relative contributions of temporal and spectral cues to the perception of the tonal characteristics of Mandarin in normal-hearing adults and to provide the theoretical foundations for the improvement of Chinese speech processing strategies utilized in cochlear implant systems. Methods 13 native Mandarin speakers with normal hearing were recruited in this study. The tone recognition test software by Ohio University was used in this study. The testing material included 10 items with 4 words in each. The vowels and consonants of each testing item were the same but with different tones. The original testing sounds were electronically processed and subsequently recorded in the sound-proof room. Temporal cues were manipulated through the low-pass filters at 1, 2, 4, 8, 16, 32, 64, 128, 256 and 512 Hz while spectral cues were controlled by varying the number of channels （1, 2, 4, 6, 8, 12, 16 and 20）. A total of 6,400 sounds, produced in different combinations, were presented through loudspeaker at 65 dB SPL in sound-proof test room. The presentation of the sounds and the recording of the test results were executed automatically by the Matlab software. Results The scores of the tone recognition for the normal-hearing adults increased as the low-pass filters and the numbers of channels increased. When the low-pass filter was less than 2 Hz, the scores remained the same independent of changes of channels. When the channels were less than 2, the scores were constant despite any changes of low-pass filters. Conclusion The tone recognition is related to the temporal and spectral cues which interact with each other in helping the perception of tonal changes in Mandarin.
出处
《中国听力语言康复科学杂志》
2008年第5期18-20,共3页
Chinese Scientific Journal of Hearing and Speech Rehabilitation
关键词
时域
频域
声调识别
人工耳蜗
Temporal cue
Spectral cue
Tone recognition
Cochlear implant
