基于HMM状态聚类均值替代的宽带语音ISF参数补偿算法

王仕奎; 周琳; 吴镇扬; 尤红岩

doi:10.3724/SP.J.1146.2008.00014

基于HMM状态聚类均值替代的宽带语音ISF参数补偿算法

doi: 10.3724/SP.J.1146.2008.00014

王仕奎^{①②;周琳},
周琳,
吴镇扬,
尤红岩

基金项目:

东南大学科技基金项目(XJ0704268)和安徽省高校省级自然科学研究项目(KJ2007B088)资助课题

计量
- 文章访问数: 3504
- HTML全文浏览量: 73
- PDF下载量: 1160
- 被引次数: 0
出版历程
- 收稿日期: 2008-01-08
- 修回日期: 2008-07-14
- 刊出日期: 2009-04-19

ISF Parameters Concealment Algorithm Based on the HMM Clustering Means for Wideband Speech

摘要

摘要: 宽带语音在Internet传输中不可避免会出现丢帧现象，由于错误传播的影响，使接收语音质量急剧下降。该文采取大型连续分布隐马尔可夫模型(LCDHMM)对宽带语音ISF参数建模，采用Viterbi算法确定丢失帧之前若干语音帧ISF参数观察值的最佳状态序列。由于状态的冗余度较大，用丢帧前最近接收的正确帧ISF参数的HMM状态对应的聚类均值和真实值的加权，代替丢失帧的ISF参数值。将采取该算法的补偿语音和采取G.722.2标准附件I所提算法的补偿语音进行比较，仿真结果表明该算法具有较好的补偿效果，其波形与谱失真更小。
- 宽带语音;丢帧;隐马尔可夫模型;聚类;补偿
Abstract: There exist inevitably frame losses in the transmission of wideband speech. Due to the influence of error propagation, the quality of the received speech decays rapidly. In this paper, large hidden Markov model is adopted to model the ISF parameter in wideband speech, and the best state sequences of ISF parameters of the several speech frames before the lost frame are determined using Viterbi Algorithm. Because of the large redundancy of the HMM states, the lost ISF parameters are substituted by the weighted values between the clustering means and the real values of the ISF parameters of the nearest received frame. The speech compensated by this algorithm is compared to which by Annex I of G.722.2 specification. Simulation shows that the algorithm in this paper can result in best speech, and smaller waveform and spectrum distortion.

HTML全文

参考文献(1)

Ding Lijing, Radwan Ayman, El-Hennawey, Mohamed Samy,and Goubran Rafik A. Performance study of objective voicequality measures in VoIP[C]. IEEE Symposium onComputers and Communications, Aveiro, Portugal, 1-4 July2007: 197-202.[2]Thyssen J, Zopf R, Chen Juin-Hwey, and Shetty N. Acandidate for the ITU-T G.722 packet loss concealmentstandard[C]. IEEE International Conference on Acoustics,Speech and Signal Processing, Honolulu, USA , Vol. 4, 15-20April 2007: IV-549-IV-552.[3]Lee Moon-Keun, Jung Sung-Kyo, Kang Hong-Goo, ParkYoung-Cheol, and Youn Dae-Hee. A packet loss concealmentalgorithm based on time-scale modification for CELP-typespeech coders [C]. Acoustics, Speech, and Signal Processing,2003. Proceedings. (ICASSP '03). 2003 IEEE InternationalConference, Vol.1, 6-10 April 2003: I-116-I-119.[4]Rbro Christoffer A, Murthi Manohar N, Andersen S V, andJensen S H. Hidden Markov model-based packet lossconcealment for voice over IP [J].IEEE Trans. on Audio,Speech, and Language Processing.2006, 14(5):1609-1623[5][5] Telecommunication Standadization Sector of ITU. Widebandcoding of speech at around 16 kbit/s using AdaptiveMulti-Rate Wideband (AMR-WB) Appendix I: Errorconcealment of erroneous or lost frames [S]. 2002.[6][6] Telecommunication Standadization Sector of ITU. ITU-TDraft Rec G.722.2: Wideband coding of speech at around 16kbit/s using Adaptive Multi-Rate Wideband (AMR-WB) [S].2003.[7]Rabiner L and Juang Bin-Huang. Fundamentals of SpeechRecognition [M]. USA, Prentice Hall, 1993: 321-389.[8]Roucos S, Makhoul J, and Schwartz R. A variable-orderMarkov chain for coding of speech spectra [C]. ICASSP, 1982:582-585.[9]Farges E and Clements M. Hidden Markov models applied tovery low bit rate coding [C]. ICASSP, Tokyo, Japan, 1986:433-436.[10]周琳, 吴镇扬. 基于MS 估计和迭代结构的信源信道联合解码系统性能分析 [J].电子与信息学报.2006, 28(2):257-261浏览[11]Collura J S, McCree A, and Tremain T E. Perceptually baseddistortion measurements for spectrum quantization[C].Proceedings., 1995 IEEE Workshop on Speech Coding forTelecommunications, Sept. 20-22, 1995: 49-50.

施引文献

资源附件(0)

访问统计