English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Based on the Mixed Excitation Linear Prediction (MELP) model, this paper presents a vocoder to obtain high-quality synthetic speech at 800 bps with 25% redundancy for channel coding. The vocoder parameters are designed and an unequal error protection method is proposed to improve the robustness over random error channel. Several channel coding schemes are compared and the optimal one is then selected. Test results show that the proposed speech coding algorithm could provide satisfactory speech quality and also have strong robustness for channel error. Introduction High quality speech coding algorithm with very low bit rate is one of the most important research topics. Speech coding algorithms with no more than 2.4kbps bit rate are applied widely in secure communications, satellite communications, radio communications, IP phone, etc. Linear prediction coding (LPC) is the most widely used algorithm for low-bit-rate speech coding, such as LPC-10 [1], Code Excited Linear Prediction (CELP) [2], and Mixed Excitation Linear Prediction (MELP) [3]. Due to the better quality, MELP vocoder has been selected as the U.S. Federal Standard for 2400 bps speech compression. 2.4 kbps to 300bps speech coding algorithms based on MELP have achieved great success and widespread applications [4-13]. However, the decoded speech quality is usually vulnerable to channel errors in such low bit rate speech coding algorithms. In this paper, the vocoder parameters are designed and an unequal error protection method is proposed to improve the robustness over random error channel. Several coding schemes are compared and then the optimal one is selected. Test results show that the proposed speech coding algorithm could provide satisfactory speech quality and also have strong robustness for channel error. Principle of the speech coding algorithm Principle of MELP Mixed excitation and multi-band excitation are adopted in MELP algorithm which is based on the traditional LPC parametric model. In order to improve the quality of synthetic speech, several new techniques are adopted in the pitch extraction and excitation signal generation. Mixed excitation, aperiodic pulses, adaptive spectral enhancement, Fourier series and pulse dispersion are mainly included in MELP algorithm. Among them, mixed excitation, aperiodic pulses, adaptive spectral enhancement and Fourier series are introduced to further improve the quality of synthetic speech, while excitation signal and pulse spreading filter are used for sequential treatment of synthetic speech. The algorithm is improved based on the traditional MELP algorithm. The overall structures of the encoder and the decoder are shown in Figure 1 and Figure 2, respectively. Unlike traditional MELP algorithm, the Fourier amplitude parameters and aperiodic logo are abandoned, and only the line spectral frequency (LSF) parameters, band pass voicing coefficients (BPVC), pitch parameters and gain parameters are used for analysis, quantization, encoding and decoding. 5th International Conference on Information Engineering for Mechanics and Materials (ICIMM 2015) © 2015. The authors Published by Atlantis Press 91 Four parameters should be extracted for the speech coding algorithm based on MELP, which contains linear spectral frequency parameters (LSF), band pass voicing coefficients (BPVC), pith parameter and gain parameter, as shown in Figure 1. Gain parameters are calculated twice to once every sub-frame.

An 800 bps vocoder based on Mixed Excitation Linear Prediction