VoIP中回声抵消算法的研究与改进
|
摘要
随着宽带技术的迅速发展,互联网语音通信日益普及,与传统电话相比,IP电话以其网络带宽利用率高、通话成本低、提供丰富的增值功能而得到广泛应用。然而,VoIP的语音在与其他数据一起在网络中传送的时候要经过压缩、编码、打包等一系列处理,由于其基于分组交换技术,IP网络存在的数据丢包、延迟和抖动造成其QoS无法保证,其中一个急需解决的的问题就是回声抵消。
在VoIP系统中不存在PSTN中由于二／四线转换器阻抗不匹配引起的线路回声,回声的主要来源是声学回声,它的产生是由于在视频会议免提情况下,听筒扬声器播放出来的声音被话筒拾取后发回远端,加上网络和数据处理等各种延迟的影响,使得远端通话者能听到自己的回声。回声的干扰会严重影响通话的质量,因此,控制和消除回声是VoIP电话的关键技术之
本课题的主要工作是对普遍存在的两种回声(线路回声和声学回声)的抵消算法进行研究和改进,并在24位的AR1688 DSP芯片上实现并优化。课题的实现首先需要对算法进行定点化处理和优化,并编写汇编程序,其次进行仿真和调试。在回声抵消算法通过调试后,为了降低算法的时间和空间复杂度,提高运行速度,适应低速处理器的要求,降低设备功耗,利用了AR1688芯片自身的硬件特点和指令特点,对算法进行优化,最终在以AR1688为核心的IP话机上成功地实现了回声抵消的要求。
With the advent of IP voice technology, VoIP (Voice over IP) receives more and more application in recent years. Compared with the traditional telephone, IP phones are widely used because its high network bandwidth utilization, low-cost and flexibility to provide value-added features. However, the quality of voice can not be guaranteed because Internet is the best-effort model. If IP phone wants to compete with traditional telephone, it must improve its voice quality. One of key technologies is echo cancellation in video conference.
There are two types of echo in telephone network, the first type, line echo can be generated because of the mismatch between the subscriber loop and switch office where two wires and four wires are connected. The second type, acoustic echo arises with the invention of hand-free phones, in which the microphone and loudspeaker are coupled. Acoustic echo is characterized with long delay and complex and unstable environment, and need more genius algorithms.
The core of echo cancellation is adaptive filter theory, according to which residual echo is got by subtracting the echo replica from the actual echo and is further used to adapt the coefficients of the filter dynamically. Traditional adaptive filter algorithm has achieved good performance in dealing with line echo which has short delay. To handle acoustic echo, large-taps filters in frequency domain are often used, which make good use of FFT's efficiency to meet the real time requirements.
This thesis is to investigate the line echo cancellation and acoustic echo cancellation algorithm, multidelay block frequency domain adaptive filter (MDF) and expects to implement and optimization their codec on AR1688, a fixed-point digital signal processor. It is meaningful to realize the AEC codec in AR1688 because of the low cost and high precision. The implementation was in two steps:first, make the float point to fixed-point conversion in C with optimization. Then translate the C codecs to DSP assembly codes, which is compatible with ADSP2181, debug the echo cancellation codec, the next work is the optimization of the echo cancellation codec according to the hardware features and instruction characteristic of AR1688 chip in order to reduce the complexity of algorithm, improve the speed of program, adapt the requirements of processor, and reduce the power of equipment. Eventually, we successfully achieved the requirements of real-time AEC on IP phone which the core element is AR1688 chip.
在VoIP系统中不存在PSTN中由于二／四线转换器阻抗不匹配引起的线路回声,回声的主要来源是声学回声,它的产生是由于在视频会议免提情况下,听筒扬声器播放出来的声音被话筒拾取后发回远端,加上网络和数据处理等各种延迟的影响,使得远端通话者能听到自己的回声。回声的干扰会严重影响通话的质量,因此,控制和消除回声是VoIP电话的关键技术之
本课题的主要工作是对普遍存在的两种回声(线路回声和声学回声)的抵消算法进行研究和改进,并在24位的AR1688 DSP芯片上实现并优化。课题的实现首先需要对算法进行定点化处理和优化,并编写汇编程序,其次进行仿真和调试。在回声抵消算法通过调试后,为了降低算法的时间和空间复杂度,提高运行速度,适应低速处理器的要求,降低设备功耗,利用了AR1688芯片自身的硬件特点和指令特点,对算法进行优化,最终在以AR1688为核心的IP话机上成功地实现了回声抵消的要求。
With the advent of IP voice technology, VoIP (Voice over IP) receives more and more application in recent years. Compared with the traditional telephone, IP phones are widely used because its high network bandwidth utilization, low-cost and flexibility to provide value-added features. However, the quality of voice can not be guaranteed because Internet is the best-effort model. If IP phone wants to compete with traditional telephone, it must improve its voice quality. One of key technologies is echo cancellation in video conference.
There are two types of echo in telephone network, the first type, line echo can be generated because of the mismatch between the subscriber loop and switch office where two wires and four wires are connected. The second type, acoustic echo arises with the invention of hand-free phones, in which the microphone and loudspeaker are coupled. Acoustic echo is characterized with long delay and complex and unstable environment, and need more genius algorithms.
The core of echo cancellation is adaptive filter theory, according to which residual echo is got by subtracting the echo replica from the actual echo and is further used to adapt the coefficients of the filter dynamically. Traditional adaptive filter algorithm has achieved good performance in dealing with line echo which has short delay. To handle acoustic echo, large-taps filters in frequency domain are often used, which make good use of FFT's efficiency to meet the real time requirements.
This thesis is to investigate the line echo cancellation and acoustic echo cancellation algorithm, multidelay block frequency domain adaptive filter (MDF) and expects to implement and optimization their codec on AR1688, a fixed-point digital signal processor. It is meaningful to realize the AEC codec in AR1688 because of the low cost and high precision. The implementation was in two steps:first, make the float point to fixed-point conversion in C with optimization. Then translate the C codecs to DSP assembly codes, which is compatible with ADSP2181, debug the echo cancellation codec, the next work is the optimization of the echo cancellation codec according to the hardware features and instruction characteristic of AR1688 chip in order to reduce the complexity of algorithm, improve the speed of program, adapt the requirements of processor, and reduce the power of equipment. Eventually, we successfully achieved the requirements of real-time AEC on IP phone which the core element is AR1688 chip.
引文
[1]Paulo S. R. Diniz(著),刘郁林等译.自适应滤波算法与实现(第二版)[M].北京：电子工业出版社,2004,7
[2]龚耀寰.自适应滤波器[M].2003.
[3]邱天爽魏东兴唐洪等.通信中的自适应信号处理[M]．北京：电子工业出版社
[4]OSLEC-0.2.http://www.rowetel.com,2009
[5]Recommendation G.165. Echo Canceller.CCITT.1993.
[6]J.-S. Soo and K. Pang, "Multidelay block frequency domain adaptive filter," IEEE Transactions on Acoustics, Speech and Signal Processing, vol.38, no.2, pp.373-376, 1990.
[7]Lee, J.& Chang, S. C. "On the convergence properties of multidelay frequency domain adaptive filter", IEEE International Conference on Acoustics, Speech and Signal Processing,Vol.4,pp.1865-1868,1999.
[8]Jean-Marc Valin. On Adjusting the Learning Rate in Frequency Domain Echo Cancellation with Double-Talk. IEEE Transaction on Acoustic, Speech, and Language processing.2007
[9]袁佳能,于凤芹.一种无双端会话检测的回声抵消算法[J].计算机工程与应用,2008(15)
[10]程佩青.数字信号处理(第二版)[M]．北京：清华大学出版社.2001.
[11]Jean-Marc Valin. The Speex Codec manual(version 1.2-beta 1).Aug 12 2006,.
[12]PA1688/AR1688 Manual. http://www.palmmicro.com.cn,2006
[13]刘洪林,蒋昌茂,张建勇.IP语音通信原理、设计及组网应用[M].北京：电子工业出版社,2009
[14]PalmMicro Communication Inc. AR1688 Datasheet V1.0.2006,.
[15]ADI Corporation. Using the ADSP-2100 Family Volumel (Rev 1.0),1990
[16]ADI Corporation, ADSP-218x DSP instruction Set Reference,2004
[17]谢胜利何昭水高鹰.信号处理的自适应理论[M].北京：科学出版社2006
[18]吴卫.基于DSP的自适应回声抵消器的设计[D].西南交通大学,2006
[19]雷鸣,唐昆,崔慧娟,杜文.一种改进的声回声抵消算法[J].清华大学学报(自然科学版),2001,(01).
[20]GUSTAFSSON S, MARTIN R, JAX P, et al. A psychoaoucustic approach to combined acoustic echo cancellation and noise reduction.IEEE Trans on Speech and Audio Processing,2002,10(5):245-256.
[21]JUEANNES R B, SCALARTP, FAUCON G, et al. Combined noise and echo reduction in hands-free systems:a survey.IEEE Trans on Speech and Audio Processing,2001,9(8):808-820.
[22]Doherty,J.,Porayath,R. A robust echo canceller for acoustic environments.IEEE Trans CAS:Analog and Digital Signal Processing,1997,44 (5):389-396.
[23]GILLOIRE A, VETTERLI M. Adaptive filtering in subbands with critical sampling:analysis, experiments, and application to acoustic echo cancellation.IEEE Trans on Signal Processing,1992,40(8):1862-1875.
[24]Hallack S,Petraglia M R. Performance comparison of adaptive algorithms applied to acoustic echo cancelling[J].IEEE International Symposium on Industrial Electronics.2003,2:1147-1150.
[25]Dahl M,Claesson I. Acoustic noiseand echo canceling[J].IEEE Transactions on Vehicular Technology,1999,48 (5):1518-1526.
[26]杨树堂,韩琪,余胜生,周敬利.电视会议系统中的消回声处理研究[J].电子计算机与外部设备,1998,22(1)55～58
[27]Harteneck M, Weiss S, Stewart R W. Design of near perfect reconstruction oversampled filter banks for subband adaptive filters.IEEE Transactions on Circuits and Systems Part Ⅱ:Analog and Digital Signal Processing,1999,46 (8):1081-1 086.
[28]Junghsi Lee, Sheng-Chieh Chong. On the Convergence Properties of Multidelay Frequency Domain Adaptive Filter. Acoustics, Speech, and Signal Processing. Vol 15-19.1999,3:1865-1868.
[29]Jablon, N.K.; On the Complexity of Frequency-domain Adaptive Filtering Signal Processing.IEEE Transactions on Acoustics, Speech, and Signal Processing; Vol 39-10.1991-10:2331-2334.
[30]Bershad N.J., Feintuch P.L..Analysis of the Frequency Domain Adaptive Filter. Proceedings of the IEEE. Vol 67-12.1979-12:1658-1659.
[31]J. J. Shynk, "Frequency-domain and multirate adaptive filtering," IEEE Signal Processing Mag., vol.9, no.1, pp.14-37,Jan.1992.
[32]A. W. H. Khong, J. Benesty, and P. A. Naylor, "A low delay and fast converging improved proportionate algorithm for sparse system identification," EURASIP Journal on Audio, Speech, and Music Processing, vol.2007,2007.
[33]R. Ahmad, A. W. Khong, and P. A. Naylor, "Proportionate frequency domain adaptive algorithms for blind channel identification," in Proc. IEEE Int. Conf. Acoustics Speech Signal Processing, vol.5, May 2006, pp. V29-V32.
[34]J. Benesty, T. Gansler, D. R. Morgan, M. M. Sondhi, and S. L.Gay, Advances in Network and Acoustic Echo Cancellation. Springer,2001.
[35]J. Radecki, Z. Zilic, and K. Radecka, "Echo cancellation in IP networks," in Proc. Fourty-Fifth Midwest Symposium onCircuits and Systems, vol.2,2002, pp.219-222.
[36]ADI Corporation, VisualDSP++2.0 User's Guide for ADSP21xx DSPs,2001
[2]龚耀寰.自适应滤波器[M].2003.
[3]邱天爽魏东兴唐洪等.通信中的自适应信号处理[M]．北京：电子工业出版社
[4]OSLEC-0.2.http://www.rowetel.com,2009
[5]Recommendation G.165. Echo Canceller.CCITT.1993.
[6]J.-S. Soo and K. Pang, "Multidelay block frequency domain adaptive filter," IEEE Transactions on Acoustics, Speech and Signal Processing, vol.38, no.2, pp.373-376, 1990.
[7]Lee, J.& Chang, S. C. "On the convergence properties of multidelay frequency domain adaptive filter", IEEE International Conference on Acoustics, Speech and Signal Processing,Vol.4,pp.1865-1868,1999.
[8]Jean-Marc Valin. On Adjusting the Learning Rate in Frequency Domain Echo Cancellation with Double-Talk. IEEE Transaction on Acoustic, Speech, and Language processing.2007
[9]袁佳能,于凤芹.一种无双端会话检测的回声抵消算法[J].计算机工程与应用,2008(15)
[10]程佩青.数字信号处理(第二版)[M]．北京：清华大学出版社.2001.
[11]Jean-Marc Valin. The Speex Codec manual(version 1.2-beta 1).Aug 12 2006,.
[12]PA1688/AR1688 Manual. http://www.palmmicro.com.cn,2006
[13]刘洪林,蒋昌茂,张建勇.IP语音通信原理、设计及组网应用[M].北京：电子工业出版社,2009
[14]PalmMicro Communication Inc. AR1688 Datasheet V1.0.2006,.
[15]ADI Corporation. Using the ADSP-2100 Family Volumel (Rev 1.0),1990
[16]ADI Corporation, ADSP-218x DSP instruction Set Reference,2004
[17]谢胜利何昭水高鹰.信号处理的自适应理论[M].北京：科学出版社2006
[18]吴卫.基于DSP的自适应回声抵消器的设计[D].西南交通大学,2006
[19]雷鸣,唐昆,崔慧娟,杜文.一种改进的声回声抵消算法[J].清华大学学报(自然科学版),2001,(01).
[20]GUSTAFSSON S, MARTIN R, JAX P, et al. A psychoaoucustic approach to combined acoustic echo cancellation and noise reduction.IEEE Trans on Speech and Audio Processing,2002,10(5):245-256.
[21]JUEANNES R B, SCALARTP, FAUCON G, et al. Combined noise and echo reduction in hands-free systems:a survey.IEEE Trans on Speech and Audio Processing,2001,9(8):808-820.
[22]Doherty,J.,Porayath,R. A robust echo canceller for acoustic environments.IEEE Trans CAS:Analog and Digital Signal Processing,1997,44 (5):389-396.
[23]GILLOIRE A, VETTERLI M. Adaptive filtering in subbands with critical sampling:analysis, experiments, and application to acoustic echo cancellation.IEEE Trans on Signal Processing,1992,40(8):1862-1875.
[24]Hallack S,Petraglia M R. Performance comparison of adaptive algorithms applied to acoustic echo cancelling[J].IEEE International Symposium on Industrial Electronics.2003,2:1147-1150.
[25]Dahl M,Claesson I. Acoustic noiseand echo canceling[J].IEEE Transactions on Vehicular Technology,1999,48 (5):1518-1526.
[26]杨树堂,韩琪,余胜生,周敬利.电视会议系统中的消回声处理研究[J].电子计算机与外部设备,1998,22(1)55～58
[27]Harteneck M, Weiss S, Stewart R W. Design of near perfect reconstruction oversampled filter banks for subband adaptive filters.IEEE Transactions on Circuits and Systems Part Ⅱ:Analog and Digital Signal Processing,1999,46 (8):1081-1 086.
[28]Junghsi Lee, Sheng-Chieh Chong. On the Convergence Properties of Multidelay Frequency Domain Adaptive Filter. Acoustics, Speech, and Signal Processing. Vol 15-19.1999,3:1865-1868.
[29]Jablon, N.K.; On the Complexity of Frequency-domain Adaptive Filtering Signal Processing.IEEE Transactions on Acoustics, Speech, and Signal Processing; Vol 39-10.1991-10:2331-2334.
[30]Bershad N.J., Feintuch P.L..Analysis of the Frequency Domain Adaptive Filter. Proceedings of the IEEE. Vol 67-12.1979-12:1658-1659.
[31]J. J. Shynk, "Frequency-domain and multirate adaptive filtering," IEEE Signal Processing Mag., vol.9, no.1, pp.14-37,Jan.1992.
[32]A. W. H. Khong, J. Benesty, and P. A. Naylor, "A low delay and fast converging improved proportionate algorithm for sparse system identification," EURASIP Journal on Audio, Speech, and Music Processing, vol.2007,2007.
[33]R. Ahmad, A. W. Khong, and P. A. Naylor, "Proportionate frequency domain adaptive algorithms for blind channel identification," in Proc. IEEE Int. Conf. Acoustics Speech Signal Processing, vol.5, May 2006, pp. V29-V32.
[34]J. Benesty, T. Gansler, D. R. Morgan, M. M. Sondhi, and S. L.Gay, Advances in Network and Acoustic Echo Cancellation. Springer,2001.
[35]J. Radecki, Z. Zilic, and K. Radecka, "Echo cancellation in IP networks," in Proc. Fourty-Fifth Midwest Symposium onCircuits and Systems, vol.2,2002, pp.219-222.
[36]ADI Corporation, VisualDSP++2.0 User's Guide for ADSP21xx DSPs,2001