Speech signal transmission rate compression using of the time parameters coding method

Downloads

Authors

  • Cz. BASZTURA Acoustic Signals Analysis and Processing Division, Institute Telecommunications and Acoustics, Technical University of Wrocław, Poland

Abstract

A new method of speech signal encoding and decoding is the subject of the presented research project. This method may find applications in the field of telecommunications (for telephone or radio transmission) as well as in the field of speech synthesis. The concept of the method is based on the extraction and transmission of such time parameters as the intervals between the subsequent speech signal zero-crossings and the amplitude of the signal in these intervals as well as on the subsequent speech signal's reconstruction based on the given parameters and knowledge derived from the analysis of speech. The system is composed of two main processes: the extraction and encoding of the transmission parameters and the original speech signal reconstruction (resynthesis). The method makes it possible of decrease the transmission rate about 10 times, as compared to the original speech signal. The results prove that the synthesized speech signal quality, when the new method is used, may be better than the one obtained by the use of other vocoder methods. The method's low cost and a relatively simple hardware system developed for the parameters' extraction and for the reconstruction of the speech signal are the most important advantages of the described method.

References

[1] D.R. ALLEN, W.J. STRONG and E.P. PALMER, Experiments on the intelligibility of low frequency speech codes, JASA 70, 5, 1248—1255 (1981).

[2] [in Russian].

[31 T.P. BARNWELL, Objective measures for speech quality testing, J. Acoust. Soc. A., 66, 6, 1658—1663 (1979).

[4] Cz. BASZTURA, Sources, signals and acoustical images – processing, analysis and recognition [in Polish], WKiŁ, Warszawa 1988.