机载数字化音响中的降噪技术研究
|
摘要
在强噪音的机载环境中特别是旋转翼的直升机,旋翼产生的气流噪音是是影响人机功效的重要因素,另一方面由于无线电电台和前后舱互通话筒产生的噪音也严重影响了飞行员的舒适性。抗噪声技术的研究及语音处理系统开发,是语音处理非常重要的内容,在各领域上都已有广泛的的应用。但是在机载音响中,由于应用的特殊性和使用上限制,机载音响的降噪一直以来都受到业界的关注。本课题对针对载机噪声环境,对主流的语音增强算法进行仿真,针对机载强噪声抑止效果进行评估,并构建实时处理平台,对算法进行移值。根据课题的应用研究,本文简述了数字音响中的降噪技术,涉及语音增强及噪声抑制相关技术的基本原理,内容包括强背景噪声下的语音有无检测算法VAD、自适应噪声抑制算法ANS、自适应回声抵消算法AEC、及双通道语音增强算法ANC。文中给出了各种算法的MATLAB仿真结果,设计机载数字话音综合处理系统,并利用有关数据给出了VAD、ANS和AEC算法的综合效果。
本课题主要研究内容如下：
1.针对机载音响应用中无线电电台和前后舱互通话筒产生的噪音,采用计算频谱变化大小的语音检测算法VAD,在强背景噪声、较低信噪比的情况下,有音段和无音段也能够被准确的分开,提高无语音时的人耳舒适度。
2.针对传统机载模拟系统中反映强烈的啸叫,对自适应回声抵消技术进行了研究,从根本上消除了模拟音响啸叫产生的物理基础,达到啸叫抑制指标。
3.针对机载环境下有音段的带噪话音,对自适应噪声抑制算法ANS、双通道语音增强算法ANC进行了研究,在强噪声背景下能够较好地提高语音的可听可懂度。
4.构建实时处理平台,研制数字化音响中心实物硬件系统,对算法进行移值利用有关数据给出机载环境下数字话音系统的综合降噪效果。
In the strong noise onboard especially rotary wing helicopter, rotor air flow generated noise is important factors of human effectiveness, on the other hand due to radio noise and interworking of microphone noise also affected the pilot comfort. Anti-noise technology research and development of voice processing systems is a very important content, in various fields have broad application. However in the airborne sound system, due to particularities and restrictions on the use of the application, airborne audio noise reduction has always been seen by industry concern. The issue of carrier aircraft noise environment, the mainstream speech enhancement algorithm simulation for on-board to assess the effect of strong noise rejection, and build real-time processing platform, the algorithm shift value. According to the subject of applied research, this article outlines the noise reduction in digital audio technology, involving speech enhancement and noise suppression associated with the fundamental principles, including strong voice without background noise detection algorithm VAD, adaptive noise suppression algorithm ANS, adaptive echo cancellation algorithm AEC, and dual-channel speech enhancement algorithm ANC. Text is given in a variety of algorithms for MATLAB simulation results, designing airborne digital integrated processing system and use of relevant data gives the combined effect of VAD, ANS and ANC algorithm.
The subject's main research contents are as follows:
1. For the airborne sound application in radio station and around the cabin intercommunication microphone noise, are calculated by using the spectrum changes in the size of the voice detection algorithm in VAD, in the strong background noise, the situation of low SNR, a segment and segment can also be accurately separated, improve the voice ear comfort.
2. For traditional airborne simulation system reflects the strong howling, the adaptive echo cancellation technology is studied, and fundamentally eliminates the simulation of sound whistle the physical foundation, achieve the howling suppression index.
3. For the environment has a segment of the noisy speech, the adaptive noise suppression algorithm ANS, dual channel speech enhancement algorithm of ANC were studied, in strong background noise can be effectively improved by the audible speech intelligibility.
4. Construction of real-time processing platform, the development of digital voice processing system, the algorithm is to shift the value of using relevant data are given in airborne environment digital voice system integrated noise reduction effect.
本课题主要研究内容如下：
1.针对机载音响应用中无线电电台和前后舱互通话筒产生的噪音,采用计算频谱变化大小的语音检测算法VAD,在强背景噪声、较低信噪比的情况下,有音段和无音段也能够被准确的分开,提高无语音时的人耳舒适度。
2.针对传统机载模拟系统中反映强烈的啸叫,对自适应回声抵消技术进行了研究,从根本上消除了模拟音响啸叫产生的物理基础,达到啸叫抑制指标。
3.针对机载环境下有音段的带噪话音,对自适应噪声抑制算法ANS、双通道语音增强算法ANC进行了研究,在强噪声背景下能够较好地提高语音的可听可懂度。
4.构建实时处理平台,研制数字化音响中心实物硬件系统,对算法进行移值利用有关数据给出机载环境下数字话音系统的综合降噪效果。
In the strong noise onboard especially rotary wing helicopter, rotor air flow generated noise is important factors of human effectiveness, on the other hand due to radio noise and interworking of microphone noise also affected the pilot comfort. Anti-noise technology research and development of voice processing systems is a very important content, in various fields have broad application. However in the airborne sound system, due to particularities and restrictions on the use of the application, airborne audio noise reduction has always been seen by industry concern. The issue of carrier aircraft noise environment, the mainstream speech enhancement algorithm simulation for on-board to assess the effect of strong noise rejection, and build real-time processing platform, the algorithm shift value. According to the subject of applied research, this article outlines the noise reduction in digital audio technology, involving speech enhancement and noise suppression associated with the fundamental principles, including strong voice without background noise detection algorithm VAD, adaptive noise suppression algorithm ANS, adaptive echo cancellation algorithm AEC, and dual-channel speech enhancement algorithm ANC. Text is given in a variety of algorithms for MATLAB simulation results, designing airborne digital integrated processing system and use of relevant data gives the combined effect of VAD, ANS and ANC algorithm.
The subject's main research contents are as follows:
1. For the airborne sound application in radio station and around the cabin intercommunication microphone noise, are calculated by using the spectrum changes in the size of the voice detection algorithm in VAD, in the strong background noise, the situation of low SNR, a segment and segment can also be accurately separated, improve the voice ear comfort.
2. For traditional airborne simulation system reflects the strong howling, the adaptive echo cancellation technology is studied, and fundamentally eliminates the simulation of sound whistle the physical foundation, achieve the howling suppression index.
3. For the environment has a segment of the noisy speech, the adaptive noise suppression algorithm ANS, dual channel speech enhancement algorithm of ANC were studied, in strong background noise can be effectively improved by the audible speech intelligibility.
4. Construction of real-time processing platform, the development of digital voice processing system, the algorithm is to shift the value of using relevant data are given in airborne environment digital voice system integrated noise reduction effect.
引文
[1]余浩深,语音增强算法研究与基于TMS320C54XDSP的实时实现,2003年湖南大学硕士论文。
[2]姚峰英,语音增强系统的研究与实现,2001中国科学院博士论文。
[3]J. S. Lim and A. V. Oppenheim, Enhancement and Bandwidth Compression of Noisy Speech, Proceedings of IEEE,1979, vol.67, No.12:1586-1604。
[4]杨行峻、迟惠生等,语音信号数字处理,北京,电子工业出版社,1995。
[5]Steven F. Boll, A Spectral Subtraction Algorithm for Suppression of Acoustic Noise in Speech, ICASSP,1979:200-203。
[6]R. H. Frazier, S. Samsam, L. D. Braida and A. V. Oppenheim, Enhancement of Speech by Adaptive Filtering, ICASSP,1976:251-253。
[7]S. V. Vaseghi, Advanced signal processing and digital noise reduction, John Wiley & Sons. Ltd.,1996。
[8]L. Rabiner and B.-H. Juang, Fundamentals of Speech Recognition, Prentice Hall,1993。
[9]Biing-Hwang Juang and L. R. Rabiner, Signal Restoration by Spectral Mapping, ICASSP, 1987:2368-2371。
[10]Yariv Ephraim, Statistical-Model-Based Speech Enhancement Systems, Proceeding of IEEE, 1992,80(10):1526-1555。
[11]李冲泥、胡光锐,一种改进的子波域语音增强方法,通信学报,1999年,第20卷,第4期,88-91。
[12]Saeed Gazor and Afshin Rezayee, an Adaptive Subspace Approach for Speech Enhancement, ICASSP,2000:1839-1842。
[13]Potamitis, N. D. Fakotakis and G. Kokkinakis, Impulsive Noise Suppression Using Neural Networks, ICASSP,2000:1871-1874。
[14]R. Martin, Spectral subtraction based on minimum statistics, In Proc EUSIPCO, pages 1182-1185, Edinburgh, Sept 1994.
[15]王朋、塔维娜、陈树中,带噪汉语语音识别的端点检测方法,计算机工程,第29卷,第17期,120-122,2003。
[16]缪春波,语音增强及其相关技术的研究,2003大连理工大学硕士论文。
[17]岳子琪,3G中语音端点检测算法及其实现研究,2002西安电子科技大学硕士论文。
[18]于迎霞,基于语音学知识的鲁棒性两级语音起点检测方法,噪声控制,2004,51-54,2004。
[19]王让定、柴佩琪,一个基于谱熵的语音端点检测改进方法,信息与控制,第33卷,第1期,2004,76-81,2004。
[20]杨俊杰、黎福海,G.729 Annex B简介,电声技术,第9期,2002,10-12。
[21]周德俊,话音通信中的非连续传输技术,通信技术,2001第9期,46-48。
[22]姚善化、王艳秋,矿井通信中间断传输研究,煤矿机电,2003年第5期,112-116。
[23]ITU-T Rec. P.800:Methods for subjective determination of transmission quality,1996。
[24]赵明忠,短波数字通信系统中数字AGC的DSP实现,电子工程师,2002年,第28卷,第11期,14-16。
[25]G. S. Kang and Mark L. Lidd, "Automatic Gain Control," IEEE ICASSP Proceeding of the conference,1984, pp 19.6.1-19.6.4。
[26]Javier Tapia and Sen M. Kuo, "An Optimized Automatic Gain Controller for Real Time Recursive LMS Adaptive Filters," in Proc. IEEE Int. Symp. Circuits Syst.,1990; pp. 3154-3157。
[27]何培宇,王永德,自适应声回波对消,电子科技导报,1999,第十期,页码：28-31。
[28]刘建锋,“自适应回音抵消技术分析与探讨,”数字通信,2000,第三期,页码：8-9
[29]S. L. Gay and J. Benesty, "Acoustic Signal Processing for Telecommuniction," Kluwer Academic Publishers。
[30]方元,无参考输入双话筒噪声抵消,南京大学学报：自然科学版,2002,38(1),59-63。
[31]Hui Ding、Jing Lu、Xiaojun Qru、Boling Xu, An adaptive speech enhancement method for siren noise cancellation, Applied Acoustics 65(2004) 385-399
[32]余凯、徐柏龄等,基于差分相位相关收敛准则的迭代语音信号分离,声学学报,2001,Vol.26,No.4,335-342
[33]卢晶,沈春华等,基于听觉特性的迭代维纳滤波在语音分离中的应用,应用声学,2001,20(5)1-5
[34]R. Zelinski, A microphone array with adaptive post-filtering for noise reduction in reverberate rooms, ICASSP,1988,2578-2581。
[35]朱琦,多声道语音增强系统中维纳滤波器的研究及其在语音识别中的应用,信号处理,1997,Vol.13,No.3,241-250
[36]R. Zelinski, A microphone array with adaptive post-filtering for noise reduction in reverberate rooms, ICASSP,1988,2578-2581。
[2]姚峰英,语音增强系统的研究与实现,2001中国科学院博士论文。
[3]J. S. Lim and A. V. Oppenheim, Enhancement and Bandwidth Compression of Noisy Speech, Proceedings of IEEE,1979, vol.67, No.12:1586-1604。
[4]杨行峻、迟惠生等,语音信号数字处理,北京,电子工业出版社,1995。
[5]Steven F. Boll, A Spectral Subtraction Algorithm for Suppression of Acoustic Noise in Speech, ICASSP,1979:200-203。
[6]R. H. Frazier, S. Samsam, L. D. Braida and A. V. Oppenheim, Enhancement of Speech by Adaptive Filtering, ICASSP,1976:251-253。
[7]S. V. Vaseghi, Advanced signal processing and digital noise reduction, John Wiley & Sons. Ltd.,1996。
[8]L. Rabiner and B.-H. Juang, Fundamentals of Speech Recognition, Prentice Hall,1993。
[9]Biing-Hwang Juang and L. R. Rabiner, Signal Restoration by Spectral Mapping, ICASSP, 1987:2368-2371。
[10]Yariv Ephraim, Statistical-Model-Based Speech Enhancement Systems, Proceeding of IEEE, 1992,80(10):1526-1555。
[11]李冲泥、胡光锐,一种改进的子波域语音增强方法,通信学报,1999年,第20卷,第4期,88-91。
[12]Saeed Gazor and Afshin Rezayee, an Adaptive Subspace Approach for Speech Enhancement, ICASSP,2000:1839-1842。
[13]Potamitis, N. D. Fakotakis and G. Kokkinakis, Impulsive Noise Suppression Using Neural Networks, ICASSP,2000:1871-1874。
[14]R. Martin, Spectral subtraction based on minimum statistics, In Proc EUSIPCO, pages 1182-1185, Edinburgh, Sept 1994.
[15]王朋、塔维娜、陈树中,带噪汉语语音识别的端点检测方法,计算机工程,第29卷,第17期,120-122,2003。
[16]缪春波,语音增强及其相关技术的研究,2003大连理工大学硕士论文。
[17]岳子琪,3G中语音端点检测算法及其实现研究,2002西安电子科技大学硕士论文。
[18]于迎霞,基于语音学知识的鲁棒性两级语音起点检测方法,噪声控制,2004,51-54,2004。
[19]王让定、柴佩琪,一个基于谱熵的语音端点检测改进方法,信息与控制,第33卷,第1期,2004,76-81,2004。
[20]杨俊杰、黎福海,G.729 Annex B简介,电声技术,第9期,2002,10-12。
[21]周德俊,话音通信中的非连续传输技术,通信技术,2001第9期,46-48。
[22]姚善化、王艳秋,矿井通信中间断传输研究,煤矿机电,2003年第5期,112-116。
[23]ITU-T Rec. P.800:Methods for subjective determination of transmission quality,1996。
[24]赵明忠,短波数字通信系统中数字AGC的DSP实现,电子工程师,2002年,第28卷,第11期,14-16。
[25]G. S. Kang and Mark L. Lidd, "Automatic Gain Control," IEEE ICASSP Proceeding of the conference,1984, pp 19.6.1-19.6.4。
[26]Javier Tapia and Sen M. Kuo, "An Optimized Automatic Gain Controller for Real Time Recursive LMS Adaptive Filters," in Proc. IEEE Int. Symp. Circuits Syst.,1990; pp. 3154-3157。
[27]何培宇,王永德,自适应声回波对消,电子科技导报,1999,第十期,页码：28-31。
[28]刘建锋,“自适应回音抵消技术分析与探讨,”数字通信,2000,第三期,页码：8-9
[29]S. L. Gay and J. Benesty, "Acoustic Signal Processing for Telecommuniction," Kluwer Academic Publishers。
[30]方元,无参考输入双话筒噪声抵消,南京大学学报：自然科学版,2002,38(1),59-63。
[31]Hui Ding、Jing Lu、Xiaojun Qru、Boling Xu, An adaptive speech enhancement method for siren noise cancellation, Applied Acoustics 65(2004) 385-399
[32]余凯、徐柏龄等,基于差分相位相关收敛准则的迭代语音信号分离,声学学报,2001,Vol.26,No.4,335-342
[33]卢晶,沈春华等,基于听觉特性的迭代维纳滤波在语音分离中的应用,应用声学,2001,20(5)1-5
[34]R. Zelinski, A microphone array with adaptive post-filtering for noise reduction in reverberate rooms, ICASSP,1988,2578-2581。
[35]朱琦,多声道语音增强系统中维纳滤波器的研究及其在语音识别中的应用,信号处理,1997,Vol.13,No.3,241-250
[36]R. Zelinski, A microphone array with adaptive post-filtering for noise reduction in reverberate rooms, ICASSP,1988,2578-2581。