摘要
主要介绍了一种新的低比特率MELP (MixedExcitationLinearPrediction)混合激励线性预测语音编码技术 ,其中着重分析了该编码算法所采用的几项关键技术。给出了采用MELP压缩编码算法后的输出比特流在各个参数上的比特分配表 ,并通过计算机仿真 ,对MELP合成语音与原始语音的质量做了比较 ,最后就MELP语音编码技术与现今其它几种不同的低速率语音编码技术的合成语音质量在DRT、DAM及MOS得分三个方面做了比较。
This paper introduces a new low bit rate digital speech coding technique with the MELP (Mixed Excitation Linear Prediction) algorithm. Some critical techniques are in detail analyzed. The assignment of the bit use of the parameters for this algorithm are calculated and listed. The comparison between MELP synthesis speech and original speech is made and simulated in a personal computer. Some different coding algorithms are compared in regard to the quality of synthesis speech on the DRT DAM and MOS scores.
出处
《信息技术》
2003年第8期52-54,共3页
Information Technology
