论文部分内容阅读
应用神经网络和Levinson-Durbin算法,本文提出一种改进的语音信号非线性自适应预测编码方案。用该方案实现了16Kb/s语音信号自适应预测编码器。实验结果表明,与原方案相比,本文提出的方案解码恢复后的语音质量有明显地改善。
Using neural network and Levinson-Durbin algorithm, this paper proposes an improved non-linear speech coding adaptive predictive coding scheme. With this scheme, a 16Kb / s speech signal adaptive predictive encoder is realized. The experimental results show that compared with the original scheme, the speech quality of the scheme proposed in this paper after the restoration is obviously improved.