北京生物医学工程

___________基于Gammatone滤波器组分解的数字助听器频响补偿算法_________

Loudness compensation algorithm based on Gammatone filter bank decomposition for digital hearing aid

作者：李如玮曹龙涛鲍长春吴水才

单位：           北京工业大学电子信息与控制工程学院(北京100124)    

关键词：数字助听器；频响补偿；Gammatone滤波器组；听觉特性

分类号：           R318.04    

出版年·卷·期（页码）：2016·35·2（143-150）

摘要：

目的针对现有数字助听器中的多通道频响补偿算法容易破坏语音信号的共振峰结构，致使频响补偿后语音信号的可懂度、清晰度和舒适度降低等问题，提出一种基于人耳听觉特性的多通道频响补偿算法。方法首先采用一种可以模拟人工耳蜗模型的Gammatone滤波器组，然后通过滤波处理把输入信号分为32个频带，并依据耳聋者的听力曲线对每个频带进行频响补偿。最后采用语音信号每个通道的声压级，以及从共振峰结构和主观测试对比该算法与基于Bark域的非均匀滤波器组分割频率的算法的优越性。结果通过该算法处理后的语音信号可以很好地满足不同听力损失者的听力要求。相对于比较算法，该算法语音信号的共振峰结构完整性更好，在不同的声压级下，该算法的语句可懂度提高10%以上。结论基于Gammatone滤波器组分解的数字助听器频响补偿算法优于对比算法，并且提高了听力损失者的语句可懂度和舒适度。

Objective The intelligibility, the definition and the comfort for digital hearing aids are reduced because of distorting formant of speech resulted by the existing loudness compensation algorithm for hearing aids. In order to solve these problems, a loudness compensation algorithm based on human auditory is proposed in this paper. Methods Firstly, Input signal is divided into 32 bands by Gammatone filter banks which can simulate the auditory model of human ear cochlea. And then each band does loudness compensation according to the curve of the hearing impaired. Finally, we compare formant, intelligibility and the sound pressure level (SPL) between the proposed method and multichannel loudness compensation based on Bark. Results This algorithm can meet the requirements of people with hearing loss, and formant structure of speech sounds gets well protected. Under different SPL, intelligibility increases more than 10%. Conclusions This algorithm is superior to multichannel loudness compensation based on Bark in sentence intelligibility, definition and comfort.

参考文献：

［1］Nielsen LS, Sparso J. Designing asynchronous circuits for low power: An IFIR filter bank for a digital hearing aid［J］. Proceedings of the IEEE, 1999, 87(2): 268-281.

［2］Li M, McAllister HG, Black ND, et al. Wavelet-based nonlinear AGC method for hearing aid loudness compensation［J］. IEEE Proceedings-Vision, Image and Signal Processing, 2000, 147(6): 502-507.

［3］Chong KS, Gwee BH, Chang JS. A 16-channel low-power nonuniform spaced filter bank core for digital hearing aids［J］. Circuits and Systems II: Express Briefs, IEEE Transactions on, 2006, 53(9): 853-857.

［4］赵毅，尹雪飞，陈克安.一种基于共振峰提取的多通道响度补偿算法［J］.信号处理，2012,28(3):352-360.

Zhao Yi, Yin Xuefei, Chen Kean. A multi-channel loudness compensation algorithm based on formant extraction［J］. Signal Processing, 2012,28(3):352-360.

［5］肖宪波，胡光书，刘春红.基于形态学的数字助听器单频段压缩放大方法［J］.清华大学学报:自然科学版，2005,45(12)：1680-1683.

Xiao Xianbo, Hu Guangshu, Liu Chunhong. Morphology-based single band compression algorithm for digital hearing aids［J］. J Tsinghua Univ : Sci&Tech, 2005, 45(12):1680-1683.

［6］Vikrant NP, Krishna Y, Rajashekhar B, et al. Critical-band based compression—an insight into the future of digital hearing aids［C］//Audio, Language and Image Processing, 2008. ICALIP 2008. International Conference on IEEE, 2008: 1620-1623.

［7］王青云，赵力，赵立业，等.一种数字助听器多通道频响补偿方法［J］.电子与信息学报,2009,31(4):832-835.

Wang Qingyun, Zhao Li, Zhao Liye, et al. A multichannel loudness compensation method for digital hearing aids［J］. Journal of Electronics & Information Technology, 2009, 31(4):832-835.

［8］Greenwood DD. A cochlear frequency-position function for several species—29 years later［J］. The Journal of the Acoustical Society of America, 1990, 87(6): 2592-2605.

［9］Karuppuswamy R, Arumugam K. Folded architecture for digital gammatone filter used in speech processor of cochlear implant［J］. ETRI Journal, 2013, 35(4):697-705.

［10］Yin H, Hohmann V, Nadeu C. Acoustic features for speech recognition based on gammatone filterbank and instantaneous frequency［J］. Speech Communication, 2011, 53(5): 707-715.

［11］Johannesma PIM. The pre-response stimulus ensemble of neurons in the cochlear nucleus［C］. Eindhoven, The Netherlands：IPO Symposium on Hearing Theory， IPO, 1972: 58-69.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录