研究生dsp 第十章课件.pptVIP

* §10 Multirate Digital Signal Processing §10.1 Basic Sample Rate Alteration Devices 10.1.1 Time-Domain Characterization …(10.1) …(10.2) Fig 10.1 Fig 10.3 10.1.2 Frequency-Domain Characterization For the factor-of-L up-sampler, …(10.5) Fig.10.7 Effects of up-sampling in frequency domain Considering down-sampler …(10.6) Let …(10.7) Then, …(10.8) Now can be related to through ….(10.9) A convenient representation is given by(Problem 10.4) where ….(10.10) ….(10.11) Using Eq.(10.8) to Eq. (10.11),we get …(10.12) …(10.15) Example M=2, Fig 10.11 Effect of down-sampling in frequency absence of aliasing Fig 10.12 Effect of down-sampling in frequency Input spectrum output spectrum(M=2) output spectrum(M=3) §10.2 Filters In Sampling Rate Alteration Cascade Equivalences Case 1 Case 2 (a) (b) 10.2.1 Filter Specifications …(10.20) …(10.19) Decimator filter interpolator filter 10.2.2 Filters for Fractional Sampling Rate alteration H(z) is a LPF 10.2.3 Computational Requirements If H(z) is a FIR(length-N) The number of multiplications per second 1. Decimator keeps only every Mth sample of v[n], computational saving factor -M 2. Interpolator Less computational saving If H(z) is a IIR(order-K) It is similar to above. computational saving factor -L If H(z) is a FIR, If H(z) is a IIR, Less computational saving. Fig 10.19 (y=decimate(x,M);M=2) 10.2.4 Matlab Simulation Fig 10.20 (y=interp(x,L); L=2) Fig 10.21 (resample(x,L,M);)L=3,M=5) §10.3 Multistage Design of Decimator and Interpolator (a) (b) L=L1×L 2 Fig 10.22 Two-stage implement (b) interpolator (a) decimator M=M1×M 2 Ex 10.8 decimator implementation (12kHz---400Hz) single-stage Fig 10.25(a) steps in two-stage implementation two-stage G(z): F(z): Total complexity of FIR Fig 10.25(b) steps in two-stage implementation Fig 10.25(c) steps in two-stage implementation based on Fig 10.25(c) Three-stage implementation Fig 10.26 multi-stage implementation S(z): R(z): Total complexity of FIR b