摘要
为满足无线通信的要求,文中在传统的MELP的基础上,提出了一种速率为600b/s的语音编码算法。该算法利用帧间冗余,把连续的4帧构成一个超级帧进行联合量化。对线谱对采用两帧联合量化、双向线性内插技术,对能量参数采用分离均值矢量量化等技术。仿真实验证明该语音编码算法有较好质量。
In order to meet the needs of wireless communication, a new 600b/s speech coding algorithm based on MELP（mixed excitation linear prediction） is proposed and implemented in this paper. In this algorithm the four consecutive MELP frames are grouped into a superframe and jointly quantized by utilizing the inter-frame redundancy properties. The algorithm uses such techniques as two-frame LSP（Linear Spectrum Pair） joint quantization and bidirectional prediction in LSP quantization, mean-removed vector quantization for gain parameters etc. Simulation results indicate that the proposed coder is o~ fine quality.
出处
《通信技术》
2009年第7期268-270,共3页
Communications Technology
关键词
语音编码
多帧联合量化
超级帧
speech coding
multi-frame joint quantization
superframe
