[工程科技]通信原理第四章4 Digitization of Analog Signal.pptVIP

Cha.4 Digitization of Analog Signal Contents 4.1 Introduction 4.1 Introduction Contents 4.1 Introduction 4.2 Sampling of Analog Signal Contents 4.1 Introduction 4.3 Quantization of Sampled Signal Contents 4.1 Introduction 4.4 Pulse Code Modulation Contents 4.1 Introduction 4.5 Differential Pulse Code Modulation Contents 4.1 Introduction 4.6 Delta Modulation Contents 4.1 Introduction 4.7 Brief Summary Since A-law is a continuous smooth curve, it is difficult to be accurately achieved by electronic circuits. In practice, 13-broken-line method is usually used to approximate. 13 broken lines compression characteristic– approximation of A-Law 3/8 5/8 1 1/2 1/4 1/8 1/16 1/32 1/64 1/128 1/8 7/8 1 0 x y Abscissa x is nonuniformly divided into 8 segments, while ordinate y is uniformly divided into 8 segments. The slopes of all segments are shown in following table: Considering the slopes of the 1st and the 2nd segments are same, and there being negative part, the complete compression curve in two quadrants has 13 segments. Segment 1 2 3 4 5 6 7 8 Slope 16 16 8 4 2 1 1/2 1/4 Table 4.3.2 is a comparison of A-law and 13-broken-line. Set 1/16 of the 1st segment as the minimum interval of quantization, then the length of every segment based on it is: There are only 128 quantization intervals during nonuniform quantization, while there are 2048 uniform quantization intervals. Segment 1 2 3 4 5 6 7 8 Length( ) 16 16 32 64 128 256 512 1024 Comparing the A-law and μ-law we can make a conclusion that: SNR for small signal given by 15-broken-line is about twice that given by 13-broken-line. However, for large signal, SNR given by 15-broken-line will be a little lower than that given by 13-broken-line. μ-law and 15 broken lines compression characteristics Therefore, to ensure equal number of quantization intervals for small signal, uniform quantization needs 11 bits while nonuniform quantization needs only 7 bits. 4.2 Sampling of Analog Signal 4.3 Quantization