Analysis of real time compression algorithm for local speech syllable signals in English pronunciation

Abstract When the current algorithm is used to compress the signal, it will cause serious distortion of the signal. In this paper, a real-time compression algorithm for local speech syllable signals in English pronunciation based on variational Bayesian method is proposed. Firstly, the sparse base of the syllable signal is selected, and the signal observation matrix is established. BP algorithm and OMP algorithm are combined, and the syllable signal is reconstructed by sparse matrix of syllable signal. On this basis, the prior and posteriori functions of syllable signal are used to calculate the variational parameters of syllable signals, so that the real-time compression algorithm of syllable signals in local speech is realized. Experimental results show that the proposed algorithm can guarantee the quality of the signal.