摘要
由于算法复杂度和延迟的限制,语音编码参数仍然保留了一定的冗余信息,这些冗余量表现为参数的非等概分布和时间相关性。该文在比特级上利用信源编码参数的冗余信息在软输入、软输出信道解码和信源参数估计之间进行迭代运算,并将此联合解码算法运用于GSM EFR系统。仿真结果表明此算法降低了编码参数的BER,提高了解码语音的信噪比,具有一定的可行性和有效性。
In speech signal transmission system, residual redundancy is still left in the parameters after source encoding. The redundancy includes non-uniform distribution and time correlation of subsequent frame. Exploiting the redundancy on bit-level, the paper is proposed iterative algorithm based on soft-in/soft-out channel and parameter estimation, and applies the algorithm to GSM EFR system. The result shows this method reduces the BER and improves SNR of decoded speech.
出处
《电子与信息学报》
EI
CSCD
北大核心
2005年第6期884-887,共4页
Journal of Electronics & Information Technology
关键词
迭代算法
软输入软输出解码器
参数估计
Iterative decoding, Soft-in/soft-out decoder, Parameter estimation