语音信号端点检测方法综述及展望
|
摘要
端点检测是语音信号处理过程中非常重要的一步,它的准确性直接影响语音信号处理的速度和结果,因此端点检测方法的研究,特别是在噪声环境下端点检测的研究,一直是语音信号处理中的热点。从基于时域参数、频域参数、时频参数、模型匹配等方法的角度,较全面地回顾了端点检测方法的发展历程,对各种方法的优缺点进行了比较分析,并给出了这些方法的改进意见,对端点检测未来的研究方向进行了展望。
Endpoint detection,whose accuracy directly affects the speed and the results,is a very important step for speech signals processing.The research of endpoint detection algorithm is a hotspot of speech signals processing,especially in adverse environment.This paper introduced different methods based on time domain,frequency domain,time-frequency domain parameters and model matching algorithms.Meanwhile,it analyzed the advantages,drawbacks and improvements of those algorithms.It also discussed the prospects of endpoint detection.
Endpoint detection,whose accuracy directly affects the speed and the results,is a very important step for speech signals processing.The research of endpoint detection algorithm is a hotspot of speech signals processing,especially in adverse environment.This paper introduced different methods based on time domain,frequency domain,time-frequency domain parameters and model matching algorithms.Meanwhile,it analyzed the advantages,drawbacks and improvements of those algorithms.It also discussed the prospects of endpoint detection.
引文
[1]LAMEL L F,RABINER L R,ROSENBERG A E,et al.An im-proved endpoint detector for isolated word recognition[J].IEEETrans on Acoust,Voice,Signal Processing,1981,29(8):777-785.
[2]LULie,JIANG Hao,ZHANG Hong-jiang.A robust audio classifica-tion and segmentation method[C]//Proc of the 9th ACM Internatio-nal Conference on Multimedia.2001.
[3]SAVOJI MH.A robust algorithm for accurate endpointing of speech[J].Speech Communications,1989,8(1):45-60.
[4]贾川,张健,陈振标,等.噪声环境下的端点检测算法研究[C]//第六届全国人机语音通信学术会议论文集.2001:441-445.
[5]RABINER L R,SAMBUR M R.An algorithm for determining theendpoints of isolated utterances[J].Bell System Technical Jour-nal,1975,54(2):297-315.
[6]张仁志,崔慧娟.基于短时能量的语音端点检测算法研究[J].电声技术,2005(7):52-54.
[7]肖述才,王作英.端点检测中的一种新的对数能量特征[J].电声技术,2004(6):37-41.
[8]李明远,李建东.基于相关性的语音激活检测器[J].电声技术,1995(11):6-9.
[9]陈斐利,朱杰.一种新的基于自相关相似距离的语音信号端点检测方法[J].上海交通大学学报,1999,33(9):1097-1099.
[10]卢艳玲,侯榆青,王宾,等.一种基于多特征的带噪语音信号端点检测与音节分割算法[J].电声技术,2005(7):60-62.
[11]NEY H.An optimization algorithm for determining the endpoints ofisolated utterances[C]//Proc of ICASSP.1981:720-723.
[12]刘庆升,徐霄鹏,黄文浩.一种语音端点检测方法的探究[J].计算机工程,2003,29(3):120-121.
[13]RABINER L R,SAINBUR M R.Voiced unvoiced silence detectionusing the Itakura LPC distance measure[C]//Proc of ICASSP.1977:323-326.
[14]SHEN J L,HUNG J W,LEE L S.Robust entropy-based endpointdetection for speech recognition in noisy environments[C]//Proc ofInternational Conference on Spoken Language Processing.Sydney:[s.n.],1998:232-238.
[15]于迎霞,史家茂.一种改进的基于倒谱特征的带噪端点检测方法[J].计算机工程,2004,30(19):85-87.
[16]SHANNON C E.A mathematical theory of communication[J].BellSyst Technical Journal,1948,27:379-423.
[17]WUBing-fei,WANG Kun-ching.Robust endpoint detection algorithmbased on the adaptive band-partitioning spectral entropy in adverse en-vironments[J].IEEE Trans on Speech and Audio Processing,2005,13(5):762-775.
[18]WUG D,LINC T.Word boundary detection with Mel-scale frequen-cy bank in noisy environment[J].IEEE Trans on Speech and Au-dio Processing,2000,8(5):541-554.
[19]李晔,张仁智,崔慧娟,等.低信噪比下基于谱熵的语音端点检测算法[J].清华大学学报:自然科学版,2005,45(10):1397-1400.
[20]吴军,王作英.汉语信息熵和语言模型的复杂度[J].电子学报,1996,24(10):69-71.
[21]田野,王作英,陆大金.基于子带能量线性映射的噪声中端点检测算法[J].清华大学学报:自然科学版,2002,42(7):953-956.
[22]王让定,柴佩琪.一种基于谱熵的语音端点检测改进方法[J].信息与控制,2004,33(1):77-81.
[23]陈四根,和应民.一种基于信息熵的语音端点检测方法[J].应用科技,2001,28(3):13-14.
[24]HUANG Liang-sheng,YANG C H.A novel approach to robustspeech endpoint detection in carenvironments[C]//Proc of ICASSP.2000:1751-1754.
[25]JUNQUA J C,MAK B,REAVES B.A robust algorithm for wordboundary detection in the presence of noise[J].IEEE TransSpeech Audio Processing,1994,2(3):406-412.
[26]郭继云,王守觉,刘学刚.一种改进的基于频能比的端点检测算法[J].计算机工程与应用,2005,41(29):91-93.
[27]朱杰,韦晓东.噪声环境中基于HMM模型的语音信号端点检测方法[J].上海交通大学学报,1998,32(10):14-16.
[28]徐筱麟,张兴国.一种基于马可夫过程统计模型的语音激活检测方法[J].解放军理工大学学报:自然科学版,2003,4(1):7-10.
[29]董恩清,赵鹤鸣,周亚同,等.支持向量机在语音激活检测中的应用研究[J].通信学报,2003,24(3):70-77.
[30]范万春,施仁,孙煜,等.应用统计模型的地震信号端点检测方法[J].西安交通大学学报,2001,35(4):365-369.
[31]SOHN J,KIMMN S,SUNG W.A statistical model-based voice ac-tivity detection[J].IEEE Signal Processing Letters,1999,6(1):1-3.
[32]QI Ying-yong,HUNT B R.Voiced-unvoiced-silence classification ofspeech using hybrid features and a network classifier[J].IEEE Transon Speech and Audio Processing,1993,1(2):250-255.
[33]KIA S J,COGHILL G G.A mapping neural network and its applica-tion to voiced-unvoiced-silence classification[C]//Proc of the 1stNew Zealand Int Two-Stream Conf Artificial Neural Networks ExpertSystems.1993:104-108.
[34]GHISELLI-CRIPPA T,EL-JAROUDI A.A fast neural net trainingalgorithm and its application to voiced-unvoiced-silence classificationof speech[C]//Proc of Int Confon Speech Language Processing.1991:441-444.
[35]魏涛,顾涵铮.一种基于声学分类的语音激活检测算法[J].合肥工业大学学报:自然科学版,2001,24(2):222-225.
[36]邝航宇,张军,韦岗.一种基于检测元音的孤立词端点检测算法[J].电声技术,2005(3):40-43.
[37]刘鹏,王作英.多模式语音端点检测[J].清华大学学报:自然科学版,2005,45(7):896-899.
[38]VATIKIOTI-BATESON E,BAILLY G,ERRIER P.Audio visualspeech processing[M].[S.l.]:MITPress,2007.
[39]丁琦,徐望,王炳锡.基于模糊分类器的能量可变噪声环境下的词边界检测[J].电声技术,2003(5):45-49.
[40]BERITELLI F.A robust endpoint detector based on differential pa-rameters and fuzzy pattern recognition[C]//Proc of ICSP.1998:601-604.
[41]赵力.语音信号处理[M].北京:北京机械工业出版社,2003.
[42]张蕾.电脑也能读唇语[EB/OL].http://www.people.com.cn/GB/it/53/142/20030501/983126.html.
[43]英特尔推出读唇语的AVSR软件[EB/OL].(2003-04-30).ht-tp://article.pchome.net/content-6819.htm.
[44]可读唇语手机[EB/OL].(2002-04-11).http://www.zaobao.com/special/newspapers/2002/04/hfwb110402.html.
[45][EB/OL].(2004-04-09).http://computer.online.sh.cn/compu-ter/gb/content/2002-04/09/content_32581 2.htm.
[2]LULie,JIANG Hao,ZHANG Hong-jiang.A robust audio classifica-tion and segmentation method[C]//Proc of the 9th ACM Internatio-nal Conference on Multimedia.2001.
[3]SAVOJI MH.A robust algorithm for accurate endpointing of speech[J].Speech Communications,1989,8(1):45-60.
[4]贾川,张健,陈振标,等.噪声环境下的端点检测算法研究[C]//第六届全国人机语音通信学术会议论文集.2001:441-445.
[5]RABINER L R,SAMBUR M R.An algorithm for determining theendpoints of isolated utterances[J].Bell System Technical Jour-nal,1975,54(2):297-315.
[6]张仁志,崔慧娟.基于短时能量的语音端点检测算法研究[J].电声技术,2005(7):52-54.
[7]肖述才,王作英.端点检测中的一种新的对数能量特征[J].电声技术,2004(6):37-41.
[8]李明远,李建东.基于相关性的语音激活检测器[J].电声技术,1995(11):6-9.
[9]陈斐利,朱杰.一种新的基于自相关相似距离的语音信号端点检测方法[J].上海交通大学学报,1999,33(9):1097-1099.
[10]卢艳玲,侯榆青,王宾,等.一种基于多特征的带噪语音信号端点检测与音节分割算法[J].电声技术,2005(7):60-62.
[11]NEY H.An optimization algorithm for determining the endpoints ofisolated utterances[C]//Proc of ICASSP.1981:720-723.
[12]刘庆升,徐霄鹏,黄文浩.一种语音端点检测方法的探究[J].计算机工程,2003,29(3):120-121.
[13]RABINER L R,SAINBUR M R.Voiced unvoiced silence detectionusing the Itakura LPC distance measure[C]//Proc of ICASSP.1977:323-326.
[14]SHEN J L,HUNG J W,LEE L S.Robust entropy-based endpointdetection for speech recognition in noisy environments[C]//Proc ofInternational Conference on Spoken Language Processing.Sydney:[s.n.],1998:232-238.
[15]于迎霞,史家茂.一种改进的基于倒谱特征的带噪端点检测方法[J].计算机工程,2004,30(19):85-87.
[16]SHANNON C E.A mathematical theory of communication[J].BellSyst Technical Journal,1948,27:379-423.
[17]WUBing-fei,WANG Kun-ching.Robust endpoint detection algorithmbased on the adaptive band-partitioning spectral entropy in adverse en-vironments[J].IEEE Trans on Speech and Audio Processing,2005,13(5):762-775.
[18]WUG D,LINC T.Word boundary detection with Mel-scale frequen-cy bank in noisy environment[J].IEEE Trans on Speech and Au-dio Processing,2000,8(5):541-554.
[19]李晔,张仁智,崔慧娟,等.低信噪比下基于谱熵的语音端点检测算法[J].清华大学学报:自然科学版,2005,45(10):1397-1400.
[20]吴军,王作英.汉语信息熵和语言模型的复杂度[J].电子学报,1996,24(10):69-71.
[21]田野,王作英,陆大金.基于子带能量线性映射的噪声中端点检测算法[J].清华大学学报:自然科学版,2002,42(7):953-956.
[22]王让定,柴佩琪.一种基于谱熵的语音端点检测改进方法[J].信息与控制,2004,33(1):77-81.
[23]陈四根,和应民.一种基于信息熵的语音端点检测方法[J].应用科技,2001,28(3):13-14.
[24]HUANG Liang-sheng,YANG C H.A novel approach to robustspeech endpoint detection in carenvironments[C]//Proc of ICASSP.2000:1751-1754.
[25]JUNQUA J C,MAK B,REAVES B.A robust algorithm for wordboundary detection in the presence of noise[J].IEEE TransSpeech Audio Processing,1994,2(3):406-412.
[26]郭继云,王守觉,刘学刚.一种改进的基于频能比的端点检测算法[J].计算机工程与应用,2005,41(29):91-93.
[27]朱杰,韦晓东.噪声环境中基于HMM模型的语音信号端点检测方法[J].上海交通大学学报,1998,32(10):14-16.
[28]徐筱麟,张兴国.一种基于马可夫过程统计模型的语音激活检测方法[J].解放军理工大学学报:自然科学版,2003,4(1):7-10.
[29]董恩清,赵鹤鸣,周亚同,等.支持向量机在语音激活检测中的应用研究[J].通信学报,2003,24(3):70-77.
[30]范万春,施仁,孙煜,等.应用统计模型的地震信号端点检测方法[J].西安交通大学学报,2001,35(4):365-369.
[31]SOHN J,KIMMN S,SUNG W.A statistical model-based voice ac-tivity detection[J].IEEE Signal Processing Letters,1999,6(1):1-3.
[32]QI Ying-yong,HUNT B R.Voiced-unvoiced-silence classification ofspeech using hybrid features and a network classifier[J].IEEE Transon Speech and Audio Processing,1993,1(2):250-255.
[33]KIA S J,COGHILL G G.A mapping neural network and its applica-tion to voiced-unvoiced-silence classification[C]//Proc of the 1stNew Zealand Int Two-Stream Conf Artificial Neural Networks ExpertSystems.1993:104-108.
[34]GHISELLI-CRIPPA T,EL-JAROUDI A.A fast neural net trainingalgorithm and its application to voiced-unvoiced-silence classificationof speech[C]//Proc of Int Confon Speech Language Processing.1991:441-444.
[35]魏涛,顾涵铮.一种基于声学分类的语音激活检测算法[J].合肥工业大学学报:自然科学版,2001,24(2):222-225.
[36]邝航宇,张军,韦岗.一种基于检测元音的孤立词端点检测算法[J].电声技术,2005(3):40-43.
[37]刘鹏,王作英.多模式语音端点检测[J].清华大学学报:自然科学版,2005,45(7):896-899.
[38]VATIKIOTI-BATESON E,BAILLY G,ERRIER P.Audio visualspeech processing[M].[S.l.]:MITPress,2007.
[39]丁琦,徐望,王炳锡.基于模糊分类器的能量可变噪声环境下的词边界检测[J].电声技术,2003(5):45-49.
[40]BERITELLI F.A robust endpoint detector based on differential pa-rameters and fuzzy pattern recognition[C]//Proc of ICSP.1998:601-604.
[41]赵力.语音信号处理[M].北京:北京机械工业出版社,2003.
[42]张蕾.电脑也能读唇语[EB/OL].http://www.people.com.cn/GB/it/53/142/20030501/983126.html.
[43]英特尔推出读唇语的AVSR软件[EB/OL].(2003-04-30).ht-tp://article.pchome.net/content-6819.htm.
[44]可读唇语手机[EB/OL].(2002-04-11).http://www.zaobao.com/special/newspapers/2002/04/hfwb110402.html.
[45][EB/OL].(2004-04-09).http://computer.online.sh.cn/compu-ter/gb/content/2002-04/09/content_32581 2.htm.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心