MPEG-4 AAC音频编码中量化模块的改进

MPEG-4 AAC音频编码标准中的量化过程采用了Brandenburg提出的双循环模式，取得了较好的编码质量和压缩比例。但在实际编码时，该方法由于迭代次数过多，会出现收敛速度较慢，甚至死锁的情况。在深入分析MPEG-4 AAC量化算法的基础上，提出了一种快速计算每个比例因子频带的改进算法。该算法通过有效减少外循环计算过程，并确保每个比例因子频带的量化失真低于允许值，从而减少量化模块的计算量。实验结果表明，在不影响音频编码质量的前提下，该算法能有效地提高编码效率。

Improvement of quantization module in MPEG-4 AAC audio encoding

The quantization process of MPEG-4 AAC audio coding standard uses double circulation patterns proposed by Brandenburg and a better coding quality and compression ratio can be obtained. But in the process of actual coding, because of too many iteration times of the method, there will be slower convergence speed, and even deadlock situations. Based on the analysis of MPEG-4 AAC quantization algorithm, this paper proposes a fast improved algorithm which can calculate frequency band of each scale factor. By reducing the process of the external circulation effectively and ensuring the quantization distortion of each scale factor band below the allowable value, the algorithm reduces the calculation of the quantization module. The experiment results show that the algorithm can improve the coding efficiency effectively without affecting the audio coding quality.