论文部分内容阅读
一个码书中空胞腔的码字越多,则LBG算法的量化效果就越不好,解码得到的信号失真就越大。在分析处理LBG算法中空胞腔的次邻域方法的基础上,针对次邻域法所存在的缺点,文中提出了一个新的空胞腔处理方法。利用仿真实验将其与传统的几种空胞腔处理方法进行比较分析,从类间离散度、矢量量化不均匀度、处理后所剩空胞腔数以及收敛性等性能指标都表明提出的改进方法达到减少量化失真,改善量化性能的目的。
The more codewords in a hollow cell of a codebook, the worse the quantization effect of the LBG algorithm, and the greater the signal distortion obtained by decoding. On the basis of analyzing and processing the sub-neighborhood method of empty cell in LBG algorithm, a new empty cell processing method is proposed in view of the disadvantages of sub-neighborhood method. The simulation experiments were used to compare with traditional empty cell processing methods. The performance indicators such as intersubspectrum dispersion, vector quantization inhomogeneity, empty cells after treatment, and convergence were all indicated by the proposed improvements The method achieves the purpose of reducing the quantization distortion and improving the quantization performance.