论文部分内容阅读
矢量量化是低速率语音和图像编码传输的有效手段,但是矢量量化系统很容易受信道误码的影响。通常对信道误码可以用纠错码来解决,但是这需要增加系统的传输码率。该文提出了一种不增加传输码率的矢量量化抗噪声优化算法,即对矢量量化码本的索引值分配进行优化,使得优化后所有Hamming距离为1的二进制码本索引值对应的码矢量的距离尽量小,从而在不增加传输码率的条件下有效地控制信道误码引起的信号失真。软件模拟显示在随机误码的情况下此算法是很有效的,并且成功地应用于低速率声码器算法中,使得声码器在10-2的随机误码的情况下仍然具有良好的性能。
Vector quantization is an effective means of transmitting low-rate speech and image coding, but vector quantization systems are susceptible to channel error. Channel error can usually be used to solve error correction code, but this need to increase the system’s transmission rate. This paper proposes a vector quantization anti-noise optimization algorithm which does not increase the transmission bit rate, that is, optimizes the index assignment of the vector quantization codebook so that all optimized code vectors corresponding to the binary codebook index with a Hamming distance of 1 Of the distance as small as possible, so as to effectively control the signal distortion caused by the channel error code without increasing the transmission bit rate. The software simulation shows that this algorithm is very effective in the case of random error codes and has been successfully applied in the low rate vocoder algorithm so that the vocoder still has good performance under the random error of 10-2 .