一種基于高階循環(huán)累積量的改進(jìn)型VAD算法
An Improved Higher-Order Cyclic Cumulant Based VAD Algorithm
-
摘要: 根據(jù)語(yǔ)音信號(hào)的短段循環(huán)平穩(wěn)(CycloStationary, CS)特征，該文提出了一種應(yīng)用于復(fù)雜背景噪聲條件下的基于高階循環(huán)累積量的改進(jìn)型VAD(Voice Activity Detection)算法，算法采用MA(Moving Average)模型對(duì)語(yǔ)音信號(hào)建模，并選擇平均幅度差(Average Magnitude Difference Function，AMDF)的方法來(lái)估算循環(huán)頻率以降低算法復(fù)雜度。經(jīng)VoIP(Voice over Internet Protocol)平臺(tái)測(cè)試，算法對(duì)高斯(白色或有色)噪聲以及其它平穩(wěn)噪聲自適應(yīng)能力強(qiáng)、檢測(cè)性能突出, 且恢復(fù)后語(yǔ)音質(zhì)量損失較小，對(duì)于非對(duì)稱(chēng)噪聲也具備可檢測(cè)能力。Abstract: On the basis of cyclostationary feature of speech in short segment, an improved Voice Activity Detection (VAD) algorithm based on higher-order cyclic cumulant is proposed in this paper, which can be applied in complex noise background. In this algorithm, modeling speech signal under MA(Moving Average)model is adopted and estimating cyclic frequency by AMDF(Average Magnitude Difference Function) is chosen to decrease computing complexity as well. Extensive testing on VoIP platform reveals that the algorithm owns good adaptive classification capability in steady noise, such as Gauss (whited and colored) noise, while with a little loss in speech quality, even when it is applied in asymmetric noise conditions.
-
Beritelli F.[J].Casale S, Ruggeri G. New speech processing s in IP telephony. Proc. Of IEEE ICCT2000, Beijing, China.2000,:-[2]Beritelli F, Casale S, Ruggeri G. Performance evaluation and comparison of ITU-T/ETSI voice activity detectors[J].Proc. of IEEE ICASSP2001, SaltLake City, USA.2001, (3):1425-[3]Prasad R.[J].Sangwan A, Jamadagni H S, Chiranth M C, Sah R, Gaurav V. Comparison of voice activity detection algorithms for VoIP. Proc. of IEEE ISCC2002, Taormina, Italy.2002,:-[4]Haighand J A.[J].Mason J S. Robust voice activity detection using cepstral features. Proc. of IEEE TENCON1993, Beijing, China.1993,:-[5]Hoyt J D.[J].Wechsler H. Detection of human speech in structured noise. Proc. of IEEE ICASSP1994, Adelaide, Australia.1994,:-[6]Rec.G.729 Annex B. A silence compression scheme for G.729optimized for terminal conforming, Tech. Rep. ITU-T, Oct. 1996.[7]張賢達(dá). 時(shí)間序列分析 高階統(tǒng)計(jì)量方法. 北京：清華大學(xué)出版社, 1996: 356.380.[8]張賢達(dá), 保錚. 非平穩(wěn)信號(hào)分析和處理. 北京：國(guó)防工業(yè)出版社，2001: 324.358.[9]Mendel J M. Tutorial on higher-order statistics (spectra) in signal processing and system theory: theoretical results and some applications[J].Proc. IEEE.1991, 79(3):278-[10]Nemer E, Goubran R, Mahmoud S. Robust voice activity detection using higher-order statistics in the LPC residual domain[J].Speech and Audio Processing.2001, 9 (3):217-[11]姚天任. 數(shù)字語(yǔ)音處理. 武漢：華中理工大學(xué)出版社, 1994: 14.24. -
計(jì)量
- 文章訪問(wèn)數(shù): 2700
- HTML全文瀏覽量: 129
- PDF下載量: 970
- 被引次數(shù): 0
下載:
