用于立体声回波对消的非线性预处理方法的分析和研究
|
摘要
在电讯系统中使用声回波对消装置消除回声(比如语音通信终端和电话会议中由于扬声器—房间—话筒(LRM)这一声回授装置引起的回声)。声回波对消装置可由传统的单声道实现。传统的单声道回声对消器能在对消回声的同时又能较好的辨识声回授通道，这样无论发送终端发生什么改变(例如，在发送终端说话人的位置或方位等空间信息的改变)，单声道回声对消器都能很好的抑制回声。
可是在实际应用中，为了增加声音的现场感，使用双声道技术是必要的。然而对于双声道情形，要必须唯一性地辨识两个声道，由此带来的双声道回波对消问题相对于单声道来说解决起来更加困难。
本文正是在详细分析论述双声道所遇到的这些困难的基础上，从降低输入信号之间互相关性的角度出发，使用非线性方法对输入通道的两路信号进行去相关预处理。为了正确真实地分析评价每一种非线性方法对语音去相关预处理的效果和对语音失真度的主观影响，在非线性去相关预处理工作中使用LRM声回授通道冲激响应的真实数据是重要和必要的。为此我们通过使用TMS320C542 DSP和TMS320AC01在一个真实房间中实时测试获得了所需要的LRM声回授通道冲激响应的真实数据，
文中我们通过使用LRM声回授通道的真实数据和真实语音对每一种非线性方法进行了去相关预处理分析。谱分析和仿真实验表明：软限幅方法与其它几种非线性预处理方法相比，不仅模型简单，实现起来容易，对声音的听觉效果
影响较小，而且对失配的改善效果较为显著.
Acoustic echo cancelers (AECs) are necessary for communication systems such as teleconferencing to reduce echos that result from coupling between the loudspeaker and microphone. With conventional sing-channel (monophonic) systems, such AECs simultaneously reduce the echo and identify the acoustic path so that the echo remains cancelled no matter what happens at the remote transmission room ( such as the change of spatial information, for example , the change of location or direction of speaker of the remote transmission room).
To enhance the sound realism in application , two-channel audio is necessary. However, in this case ( stereophonic sound ) the acoustic echo cancellation problem is more difficult to solve because of the necessity to uniquely identify two acoustic paths.
In this paper, on the basis of explanation of these problems, we use nonlinear preprocessing to deduce the correlation of the two recorded signals. To analyze and judge correctly the result of de-correlation of every nonlinear preprocessing
solution and the subjective effect of every nonlinear preprocessing solution on speech, it is very important to use factual data of LRM impulsive response. In this paper we test LRM impulsive response in a real room using TMS320C542 and TLC320AC01 in order to get factual data of LRM impulsive response.
Using factual data of LRM impulsive response and real speech we analyze and judge every nonlinear preprocessing solution. Spectrum analyses and experiments show that sof t-limiter solution is better than those nonlinear preprocessing solutions proposed in paper[2]
可是在实际应用中，为了增加声音的现场感，使用双声道技术是必要的。然而对于双声道情形，要必须唯一性地辨识两个声道，由此带来的双声道回波对消问题相对于单声道来说解决起来更加困难。
本文正是在详细分析论述双声道所遇到的这些困难的基础上，从降低输入信号之间互相关性的角度出发，使用非线性方法对输入通道的两路信号进行去相关预处理。为了正确真实地分析评价每一种非线性方法对语音去相关预处理的效果和对语音失真度的主观影响，在非线性去相关预处理工作中使用LRM声回授通道冲激响应的真实数据是重要和必要的。为此我们通过使用TMS320C542 DSP和TMS320AC01在一个真实房间中实时测试获得了所需要的LRM声回授通道冲激响应的真实数据，
文中我们通过使用LRM声回授通道的真实数据和真实语音对每一种非线性方法进行了去相关预处理分析。谱分析和仿真实验表明：软限幅方法与其它几种非线性预处理方法相比，不仅模型简单，实现起来容易，对声音的听觉效果
影响较小，而且对失配的改善效果较为显著.
Acoustic echo cancelers (AECs) are necessary for communication systems such as teleconferencing to reduce echos that result from coupling between the loudspeaker and microphone. With conventional sing-channel (monophonic) systems, such AECs simultaneously reduce the echo and identify the acoustic path so that the echo remains cancelled no matter what happens at the remote transmission room ( such as the change of spatial information, for example , the change of location or direction of speaker of the remote transmission room).
To enhance the sound realism in application , two-channel audio is necessary. However, in this case ( stereophonic sound ) the acoustic echo cancellation problem is more difficult to solve because of the necessity to uniquely identify two acoustic paths.
In this paper, on the basis of explanation of these problems, we use nonlinear preprocessing to deduce the correlation of the two recorded signals. To analyze and judge correctly the result of de-correlation of every nonlinear preprocessing
solution and the subjective effect of every nonlinear preprocessing solution on speech, it is very important to use factual data of LRM impulsive response. In this paper we test LRM impulsive response in a real room using TMS320C542 and TLC320AC01 in order to get factual data of LRM impulsive response.
Using factual data of LRM impulsive response and real speech we analyze and judge every nonlinear preprocessing solution. Spectrum analyses and experiments show that sof t-limiter solution is better than those nonlinear preprocessing solutions proposed in paper[2]
引文
[1] J. Benesty, D. R. Morgan, and M. M. Sondi."A better understanding and an improved solution to the specific problems of stereophonic acoustic echo cancellation. "IEEE Trans. Speech Audio Processing, vol.6,pp.156 -165, Mar. 1998.
[2] Dennis R. Morgan, Senior Member, IEEE, Joseph L. Hall, and Jacob Benesty, Member, IEEE, "Investigation of Several Types of Nonlinearities for Use in Stereo Acoustic Echo Cancellation" IEEE Transactions on speech and audio processing Vol.9.NO.6, SEPTEMBER 2001.
[3] J. Benesty, F. Amand, A. Gilloire, and Y. Grenier,"Adaptive filtering algorithms for stereophonic acoustic echo cancellation." in Proc, IEEE ICASSP, 1995, pp. 3099-3102.
[4] TI, TLC320AC01C Single-Supply Analog Interface Circuit Data Manual, 1996.
[5] 戴明桢编，《TMS320C54x数字信号处理器结构、原理及应用》，TIDSP UNIVERSITY，1999．
[6] TI, TMS320C54X DSP Algebraic Instruction Set, 1996.
[7] TI, TMS320C54X DSP CPU and Peripherals, 1996.
[8] TI, TMS320C54X DSP Language Tools, 1996.
[9] TI, TMS320C54X DSP Enhanced Peripherals. 1996.
[10] TI, TMS320C54x DSKplus DSP Starter Kit User's Guide, 1996.
[11] 王春霞，《快速LMS／Newton算法的实现及其改进》．硕士毕业论文，四川大学电子信息学院，2001
[12] [美]维德罗等著，王永德，龙宪惠较译，《自适应信号处理》，1991．
[13] 陈元亨，胡家冀，鲜长久，《信号处理原理》，四川大学出版社，1989．
[14] 王永德．《随机信号分析基础》．四川大学出版社，1993．
[15] Welch, P. D. "The Use of Fast Fourier Transform for
the Estimation of Power Spectra: A Method Based on Time Averaging Over Short, Modified Perlodograms." IEEE Trans. Audio Electroacoust. Vol. AU-15 (June 1967). Pgs. 70-73.
[16] [美]S．M．凯依著，黄建国，武延祥，杨世兴译，《现代谱估计原理与应用》，科学出版社，1994．
[17] Kay, S. M. Modern Spectral Estimation. Englewood Cliffs, NJ: Prentice-Hall, 1988. Pg. 454.
[18] Rablner, L. R.,and B. Gold. Theory and Application of Digital Signal Processing. Englewood Cliffs, NJ: PrenticeHall, 1975.
[19] Matlab, vertion5.3, Copywrite 1984-1999, the Mathworks, Inc.
[2] Dennis R. Morgan, Senior Member, IEEE, Joseph L. Hall, and Jacob Benesty, Member, IEEE, "Investigation of Several Types of Nonlinearities for Use in Stereo Acoustic Echo Cancellation" IEEE Transactions on speech and audio processing Vol.9.NO.6, SEPTEMBER 2001.
[3] J. Benesty, F. Amand, A. Gilloire, and Y. Grenier,"Adaptive filtering algorithms for stereophonic acoustic echo cancellation." in Proc, IEEE ICASSP, 1995, pp. 3099-3102.
[4] TI, TLC320AC01C Single-Supply Analog Interface Circuit Data Manual, 1996.
[5] 戴明桢编，《TMS320C54x数字信号处理器结构、原理及应用》，TIDSP UNIVERSITY，1999．
[6] TI, TMS320C54X DSP Algebraic Instruction Set, 1996.
[7] TI, TMS320C54X DSP CPU and Peripherals, 1996.
[8] TI, TMS320C54X DSP Language Tools, 1996.
[9] TI, TMS320C54X DSP Enhanced Peripherals. 1996.
[10] TI, TMS320C54x DSKplus DSP Starter Kit User's Guide, 1996.
[11] 王春霞，《快速LMS／Newton算法的实现及其改进》．硕士毕业论文，四川大学电子信息学院，2001
[12] [美]维德罗等著，王永德，龙宪惠较译，《自适应信号处理》，1991．
[13] 陈元亨，胡家冀，鲜长久，《信号处理原理》，四川大学出版社，1989．
[14] 王永德．《随机信号分析基础》．四川大学出版社，1993．
[15] Welch, P. D. "The Use of Fast Fourier Transform for
the Estimation of Power Spectra: A Method Based on Time Averaging Over Short, Modified Perlodograms." IEEE Trans. Audio Electroacoust. Vol. AU-15 (June 1967). Pgs. 70-73.
[16] [美]S．M．凯依著，黄建国，武延祥，杨世兴译，《现代谱估计原理与应用》，科学出版社，1994．
[17] Kay, S. M. Modern Spectral Estimation. Englewood Cliffs, NJ: Prentice-Hall, 1988. Pg. 454.
[18] Rablner, L. R.,and B. Gold. Theory and Application of Digital Signal Processing. Englewood Cliffs, NJ: PrenticeHall, 1975.
[19] Matlab, vertion5.3, Copywrite 1984-1999, the Mathworks, Inc.