论文部分内容阅读
为了适应无线通信等甚低速语音通信应用,文中提出一种基于子帧联合编码的600b/s语音编码算法。该算法的激励源采用混合激励模型,声道参数使用帧内预测多级矢量量化进行高效量化,在参数编解码时提出了子帧分类联合的思想,并在编码端使用语音增强仰制背景噪声,解码端使用后滤波处理来改善语音质量,这些方面较传统LPC算法有了明显改进。同时,选用TI公司的TMS320VC5416DSP芯片实时实现了该算法。非正式主观试听结果表明,该算法在可懂度、消晰度等方面与传统的2.4kb/sLPC自法相当,而速率仅为LPC算法的1/4,是甚低速率的一种良好的编码方案。
In order to adapt to very low speed voice communication applications such as wireless communication, a 600b / s speech coding algorithm based on sub-frame joint coding is proposed in this paper. The excitation source of this algorithm uses a hybrid excitation model. The channel parameters are efficiently quantized using intra-frame prediction multi-level vector quantization. In the parameter coding and decoding, the idea of sub-frame classification is proposed and the coding method is used to enhance the background noise , The decoder uses post-filtering to improve the voice quality, these aspects have a significant improvement over the traditional LPC algorithm. At the same time, TI’s TMS320VC5416DSP chip is chosen to realize this algorithm in real time. The results of informal subjective listening test show that this algorithm is comparable to the traditional 2.4kb / s LPG self-sufficient method in terms of intelligibility and sharpness, while the rate is only 1/4 of the LPC algorithm, which is a good Coding scheme