Research on Coding and Enhancement Algorithm for Very Low-Rate Speech Compression of Beidou Short Message

Abstract

In recent years, with the development of broadband communication technology, the bandwidth of voice communication is no longer a problem in the field of wired communication, but in the field of wireless communication, bandwidth is a precious resource. In the case of strict transmission bit rate constraints, low-rate speech coding is of great significance. As a low-rate speech coding algorithm, the MELP algorithm has improved the LPC-10 coding scheme, and introduced mixed excitation, aperiodic pulse, Fourier series, pulse discrete filtering and adaptive spectral filtering. A relatively natural voice quality is achieved at a rate of s. The calculation load of the algorithm is small, and it has strong toughness to severe background noise, and can be transplanted into DSP system very well. This paper firstly introduces the current situation of low-rate speech coding and compares the advantages and disadvantages of several main speech coding algorithms. On this basis, the research is carried out, some formulas are deduced theoretically, and the simulation analysis is carried out under MATLAB7.0. In this paper, the MELP algorithm is improved so that its voice quality can quickly meet the needs of the future large-scale communication system development.