高斯信号的量化分析

Analysis of Quantization of Gaussian Signals

由于器件的约束条件不同,在数字信号处理中通常要对AD采样后的信号进行量化处理,如何提高量化的性能是量化的关键。针对N(μ,σ2)高斯分布的量化信号,使用理论推导的方式,得到量化损耗与限幅门限的关系式,并推导出对应量化损耗最小的最优限幅门限公式,其推导方法同样适用于其他随机信号。仿真试验证明了该公式的正确性,结论可应用于工程实现。

Quantization will be necessary in the digital signal processing because of the hardware＇s resource.It is important to improve the performance of the quantizer.An expression of quantization distortion to clip level was derived theoretical for N（μ,σ2） Gaussian pdf signal.Also the expression of the minimal quantization distortion and optimum clip level was deducted.The result of computer simulation is that the expression can be used to estimate the optimum clip level effectively and exactly in the related applications.