[1]谢维波,王永初,戴在平,等.清浊音分段判决的递推最小二乘自适应算法[J].华侨大学学报(自然科学版),2008,29(2):225-228.[doi:10.11830/ISSN.1000-5013.2008.02.0225]
　XIE Wei-bo,WANG Yong-chu,DAI Zai-ping,et al.Adaptive Voiced/Unvoiced Segmentation Based on RLS[J].Journal of Huaqiao University(Natural Science),2008,29(2):225-228.[doi:10.11830/ISSN.1000-5013.2008.02.0225]
点击复制

清浊音分段判决的递推最小二乘自适应算法()
分享到：

《华侨大学学报（自然科学版）》[ISSN:1000-5013/CN:35-1079/N]

卷:
第29卷
期数:
2008年第2期
页码:
225-228
栏目:
出版日期:
2008-04-20

文章信息/Info

Title:
Adaptive Voiced/Unvoiced Segmentation Based on RLS
文章编号:
1000-5013(2008)02-0225-04
作者:
谢维波; 王永初; 戴在平; 吴恬盈
华侨大学信息科学与工程学院; 华侨大学机电工程与自动化学院; 泉州师范学院理工学院 福建泉州362021; 福建泉州362000
Author(s):
XIE Wei-bo1;  WANG Yong-chu2;  DAI Zai-ping1;  WU Tian-ying3
1.College of Information Science and Engineering, Huaqiao University; 2.College of Mechanical Engineering and Automation, Huaqiao University, Quanzhou 362021, China 3.College of Science and Engineering, Quanzhou Normal University, Quanzhou 362000, China
关键词:
清浊音分段;  递推最小二乘;  短时过零率;  短时能量
Keywords:
voiced/unvoiced segmentation;  recursive least square;  short-period cross-zero-rate;  short-period energy
分类号:
TN912.3
DOI:
10.11830/ISSN.1000-5013.2008.02.0225
文献标志码:
A
摘要:
在"零-能"判决法的基础上,结合递推最小二乘(RLS)对非平稳信号的自适应跟踪能力,提出自适应的清浊音分段算法.算法能够快速实现语音信号清浊音的精确分段,不需要通过样本集训练进行参数调整.其自适应能力是在单一话音样本上实现的,由RLS算法在清音段、浊音段及清浊音段交界处不同的跟踪能力来判别清/浊音段.与基于阈值的方法不同,算法基于极值点的识别,避免各种基于样本集训练的自适应学习算法在泛化能力上的缺陷,对于不同采样率、说话人、音量、背景噪声等变化因素,具有较强的自适应处理能力.
Abstract:
An adaptive voiced/unvoiced segmentation based on the traditional short-time analysis,with the adaptive tracking capacity of recursive least square(RLS) to the non-steady signal,has been presented.The algorithm can rapidly realize precise voiced/unvoiced segmentation,without parameter-adjustment by samples training.The adaptability comes from a single pronunciation sample,and deciding voiced/unvoiced segmentation based on the different tracking capacity of RLS in voiced/unvoiced section and the intersection point.It is different from the methods based on threshold,the algorithm based on recognizing the extreme value can avoid the drawback of various adaptive learning algorithms in generalization,which can better adapt for various varying factors of different sampling rate,speaker,volume,background noise.

参考文献/References:

[1] 郭英, 李雪娇. 一种组合参数的语音信号清/浊音判决方法 [J]. 空军工程大学学报(自然科学版), 2002(4):18-20.doi:10.3969/j.issn.1009-3516.2002.04.006.
[2] 王毓芳. 一种自适应的汉语普通话音节清/浊音分段方法 [J]. 北京航空航天大学学报, 2001(4):409-412.doi:10.3969/j.issn.1001-5965.2001.04.009.
[3] 马霓, 韦岗, 应益荣. 基于线性预测和小波变换的语音基音周期检测新算法 [J]. 建筑科学与工程学报, 1997(2):36-42.
[4] QI Ying-youg, HUNT B R. Voiced-unvoiced-silence classifications of speech using hybridfeatures and a network classifier [J]. IEEE Transactions on Speech and Audio Processing, 1993(2):250-255.
[5] ATAL B, RABINER L. A pattern recognition approach to voiced-unvoiced-silence classification with applications to speech recognition [J]. IEEE Transactions on Signal Processing, 1976(3):201-212.doi:10.1109/TASSP.1976.1162800.
[6] BASSON J A L, PREEZ J A. Adaptive estimation of speech parameters [A]. 1994.177-182.
[7] RABINER L R, SAMBUR M R. Application of an LPC distance measure to the voiced-unvoiced-silence detection problem [J]. IEEE Transactions on Signal Processing, 1977(4):338-343.
[8] 周志杰, 胡光锐. 采用非线性网络实现清浊音判决 [J]. 南京航空航天大学学报, 1998(1):47-51.

备注/Memo

备注/Memo:
福建省自然科学基金资助项目(A0540005)
更新日期/Last Update: 2014-03-23