The Journal of the Acoustical Society of Korea

The Acoustical Society of Korea (한국음향학회)
권호 표지

Robust Speech Decoding Using Channel-Adaptive Parameter Estimation.

  • Lee, Yun-Keun (Information Technology Lab., LG Corporate Institute of Technology) ;
  • Lee, Hwang-Soo (Department of Electrical Engineering, KAIST)
  • Published : 1999.03.01
Download PDF
⟨ Previous Next ⟩

Abstract

In digital mobile communication system, the transmission errors affect the quality of output speech seriously. There are many error concealment techniques using a posteriori probability which provides information about any transmitted parameter. They need knowledge about channel transition probability as well as the 1st order Markov transition probability of codec parameters for estimation of transmitted parameters. However, in applications of mobile communication systems, the channel transition probability varies depending on nonstationary channel characteristics. The mismatch of designed channel transition probability of the estimator to actual channel transition probability degrades the performance of the estimator. In this paper, we proposed a new parameter estimator which adapts to the channel characteristics using short time average of maximum a posteriori probabilities(MAPs). The proposed scheme, when applied to the LSP parameter estimation, performed better than the conventional estimator which do not adapt to the channel characteristics.

Keywords

References

  1. IEEE Tran. on Comm. v.COM-39 no.6 Use of residual redundancy in the design of joint source/channel coders K.Sayood,;J.C.Borkenhagen,
  2. Proc. of EUROSPEECH Enhancing robustness of coded LPC-spectra to channel errors by use of residual redundancy S.Feldes,
  3. IEEE Tran. Comm. v.COM-43 no.9 Soruce-controlled channel decoding J.Hagenaurer
  4. Proc. of Int.Conf. on Acousitics, Speech, and Signal processing Robust speech decoding: a universal approach to bit error concealment T.Fingscheidt,;P.Vary,
  5. Probability, Random Variables, and Stochastic Processes A.Papoulis,
  6. IEEE Tran. on Speech and Audio Processing v.1 no.1 Efficient Vector Quantization of LPC parameters at 24 bits/frame K.K.Paliwal,;B.S.Atal,
  7. Digtal Speech A.M.Kondoz,