神经网络矢量量化技术在语音编码中的应用

Application of Neural Network Vector Quantization Technology in Speech Encoding

讨论了在语音编码中,应用神经网络技术进行矢量量化的算法。神经网络矢量量化算法可以压缩码本维数,提高码本搜索速度,从而优化矢量量化的效果。将这种优化的矢量量化算法应用于语音编码中,能降低运算复杂度,提高编码质量。

The application of neural network vector quantization in speech encoding is discussed. This vector quantization by neural network can compress the number of the codebook and make the quality of the codebook better so that it can make the result of vector quanfization better. Thus the application of this vector quantization in speech encoding can improve the quality of speech encoding.