基于压缩感知的语音信号编码研究

Research on speech signal coding based on compressed sensing

为实现语音信号更高效地传输和存储,设计了压缩感知与脉冲编码调制(PCM)技术相结合的语音压缩编码系统.利用语音信号在离散余弦变换(DCT)域的稀疏性,首先将语音信号用随机高斯矩阵进行观测,得到的观测向量维度远小于原语音信号维度,再对观测向量进行PCM编码将模拟语音信号转化为压缩编码的数字信号,接收端PCM译码后利用正交匹配追踪算法重构出语音信号.利用信噪比(SNR)对重构语音信号的质量进行评估,仿真结果表明,该方法实现了当压缩比为0.3时,仍具有较高的重构质量.

In order to realize more efficient transmission and storage of speech signals,a speech coding system combining compressed sensing with pulse code modulation(PCM)has been designed.Using the sparseness of the speech signal in discrete cosine transform domain,we first,observed the speech signal with a random Gaussian matrix,to obtain an observed vector with a much lower dimension than that of the original speech signal,Then PCM coding was applied to the observation vector to convert the analog speech signal into the compressed digital signal.The received PCM-decoded signal reconstructed the speech signal using orthogonal matching pursuit algorithm and the signal-to-noise ratio was used to evaluate the quality of the reconstructed speech signal.Simulation results show that this method still has high reconstruction quality when the compression ratio is 0.3.