Suppr超能文献

多通道电触觉语音处理器传输的音素信息。

Phonemic information transmitted by a multichannel electrotactile speech processor.

作者信息

Blamey P J, Cowan R S, Alcantara J I, Clark G M

机构信息

University of Melbourne, Royal Victorian Eye and Ear Hospital, Australia.

出版信息

J Speech Hear Res. 1988 Dec;31(4):620-9. doi: 10.1044/jshr.3104.620.

Abstract

A wearable electrotactile speech processor was evaluated in a study with seven normally hearing and four hearing-impaired subjects. The processor estimated the fundamental frequency, the second-formant frequency, and amplitude of the acoustic speech signal. These parameters were presented as a pattern of electrical pulses applied to eight electrodes positioned over the digital nerve bundles on one hand. The device was shown to provide useful information for the recognition of phonemes in closed sets of words using tactile information alone. The device also supplemented lipreading to improve the recognition of open-set words. The recognition of duration and first- and second-formant frequencies of vowels and the recognition of voicing and manner of consonants were improved over recognition with lipreading alone. Recognition of final consonants was improved more than recognition of initial consonants. These results indicate that the device may be useful to both severely and profoundly hearing-impaired people.

摘要

在一项针对7名听力正常和4名听力受损受试者的研究中,对一种可穿戴式电触觉语音处理器进行了评估。该处理器可估计声学语音信号的基频、第二共振峰频率和幅度。这些参数以施加到一只手上数字神经束上方八个电极的电脉冲模式呈现。结果表明,该设备仅使用触觉信息就能为识别封闭词汇集中的音素提供有用信息。该设备还辅助唇读以提高对开放词汇集的识别。与仅靠唇读识别相比,元音的时长、第一和第二共振峰频率的识别以及辅音的浊音和发音方式的识别都得到了改善。词尾辅音的识别比词首辅音的识别改善得更多。这些结果表明,该设备对重度和极重度听力受损者可能都有用。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验