chap.7new新版.pptVIP

In general,the second method is adopted.Let first that the analog signal passes through a compression (non-linear)amplifier Concept of non-uniform quantification 1.The step of quantification is proportional to the input of quantizer. 2.The RMS of quantizing noise power is proportional to the sampling value. To satisfy these conditions,we have two widely used compression laws μ-law and A-law y=ln(1+μx)/ln(1+μ) (μ-law ) y=Ax/(1+lnA) 0≤x≤1/A (A-law) y=(1+lnAx)/(1+lnA) 1/A≤x≤1 where μ=255 and A=87.6 Why have the values of μ=255 and A=87.6 be chosen? For practical reasons,explications