Multi-band spectral subtraction method for speech enhancement based on human auditory system

Computer Engineering and Applications ›› 2011, Vol. 47 ›› Issue (21): 145-148.

• 数据库、信号与信息处理 • Previous Articles Next Articles

Multi-band spectral subtraction method for speech enhancement based on human auditory system

SHEN Yuanyuan1，GU Jihua1，TAO Zhi1，2，ZHAO Heming2，TAN Xuedan1

1.School of Physical Science and Technology，Soochow University，Suzhou，Jiangsu 215006，China
2.School of Electronics and Information Engineering，Soochow University，Suzhou，Jiangsu 215006，China

Received:1900-01-01 Revised:1900-01-01 Online:2011-07-21 Published:2011-07-21

基于人类听觉系统的多频带语音增强

沈圆圆1，顾济华1，陶智1，2，赵鹤鸣2，谈雪丹1

1.苏州大学物理科学与技术学院，江苏苏州 215006
2.苏州大学电子信息学院，江苏苏州 215006

Abstract

Abstract: A multi-band nonlinear spectral subtraction method is proposed with the demand on improving the quality of speech signal corrupted by colored noise.According to the human auditory system，the noisy speech signals are nonlinearly divided into 24 non-overlapping critical bands.The amount of noise needed to be subtracted is determined by different Signal-to-Noise Ratio（SNR） of each band.Under the same experimental conditions，test results show that compared to the Power Spectral Subtraction method（PSS），Nonlinear Spectral Subtraction method（NSS） and the Multi-Band Spectral Subtraction（MBSS），the proposed method can reduce the residual noise and background noise more effectively.The enhanced speech signals of the proposed algorithm have better clearness and intelligibility besides higher output SNR.

Key words: auditory system, multi-band, spectral subtraction, speech enhancement

摘要： 提出了一种基于人类听觉系统的多频带非线性谱减法来进行语音增强。根据人耳听觉特性，将含噪语音信号分在24个临界频带内，由各频带的不同信噪比来确定对应的谱减参数值。实验结果证明，在相同实验条件下，与功率谱减法（PSS）、非线性谱减法（NSS）和传统多频带谱减法（MBSS）相比，该方法增强后的语音信号具有更高的输出信噪比;能更好地消除背景噪声，抑制残留噪声;增强后的语音具有更好的可懂度和清晰度。

关键词: 听觉系统, 多频带, 谱减法, 语音增强

SHEN Yuanyuan1，GU Jihua1，TAO Zhi1，2，ZHAO Heming2，TAN Xuedan1. Multi-band spectral subtraction method for speech enhancement based on human auditory system[J]. Computer Engineering and Applications, 2011, 47(21): 145-148.

沈圆圆1，顾济华1，陶智1，2，赵鹤鸣2，谈雪丹1. 基于人类听觉系统的多频带语音增强[J]. 计算机工程与应用, 2011, 47(21): 145-148.

[1]	MEI Xue, LIU Yuan, WU Haotian. Study on JPEG steganography based on syndrome-trellis codes and multi-band embedding [J]. Computer Engineering and Applications, 2016, 52(20): 118-121.
[2]	XU Wenchao, WANG Guangyan, GENG Yanxiang, BAI Fang, FEI Teng. Speech enhancement algorithm based on spectral subtraction and variable-step LMS algorithm [J]. Computer Engineering and Applications, 2015, 51(1): 213-217.
[3]	WANG Lulu1, XIA Xu2, FENG Lu1, LIU Guangcan1. New speech endpoint detection algorithm based on spectrum variance and spectral subtraction [J]. Computer Engineering and Applications, 2014, 50(8): 194-197.
[4]	WANG Xi1, LI Ying2. Multi-band spectral subtraction method applied to natural sounds classification [J]. Computer Engineering and Applications, 2014, 50(3): 190-193.
[5]	XIA Lele1, SUN Yongrong1, WANG Yong2. Speech enhancement technology based on adaptive noise estimation [J]. Computer Engineering and Applications, 2014, 50(23): 225-228.
[6]	YAN Yonghong1, CHEN Hao1，2, FAN Xiaohua2, LIU Yu2, ZENG Yun1. Design of multi-band CMOS switching power amplifier with power control [J]. Computer Engineering and Applications, 2014, 50(19): 42-46.
[7]	WANG Lulu1, LIU Guangcan1, XIA Xu2. Speech enhancement based on improved spectral subtraction algorithm [J]. Computer Engineering and Applications, 2014, 50(19): 210-213.
[8]	FENG Lu, CHEN Weibing. Improved speech endpoint detection based on weighing threshold spectral entropy [J]. Computer Engineering and Applications, 2013, 49(9): 207-210.
[9]	SUI Luying, ZHANG Xiongwei, HUANG Jianjun, ZHAO Gaihua. Improved spectral subtraction speech enhancement algorithm under non-stationary noise [J]. Computer Engineering and Applications, 2013, 49(16): 216-220.
[10]	WU Fei1，2, ZHANG Dexiang1，2. Research on fusion of multi-band remote sensing image based on Curvelet transforms [J]. Computer Engineering and Applications, 2012, 48(32): 153-156.
[11]	WEI Lixing, ZHANG Miao, ZHONG Yingchun, HAN Guang. Noise speaker dependent recognition system based on PCNN [J]. Computer Engineering and Applications, 2012, 48(3): 133-136.
[12]	ZHANG Man, TAO Liang, ZHOU Jian. Speech enhacement approach based on real-value discrete Gabor transform and spectral subtraction [J]. Computer Engineering and Applications, 2012, 48(29): 109-113.
[13]	CHA Cheng，YANG Ping，PAN Ping. Research of improved spectral subtraction algorithm utilizing Turbo iteration [J]. Computer Engineering and Applications, 2011, 47(36): 159-161.
[14]	WU Bian1，WANG Zhong2，LIU Xingtao2. Research on speech endpoint detection in strong noise [J]. Computer Engineering and Applications, 2011, 47(33): 137-139.
[15]	YUAN Yi，TIAN Xiaolin，XIA Shaowei. Blind audio digital watermarking method based on power comparison [J]. Computer Engineering and Applications, 2011, 47(1): 131-134.

Multi-band spectral subtraction method for speech enhancement based on human auditory system

基于人类听觉系统的多频带语音增强

PDF

Knowledge

Abstract

Cite this article

share this article

References

Related Articles 15

Recommended Articles

Metrics