John A. Quinn Lecture ESE 531: Digital Signal Processing Lec 15: March 21, 2017 Review, Generalized Linear Phase Systems Penn ESE 531 Spring 2017 Khanna Lecture Outline!!! 2 Frequency Response of LTI System Review: All Pass Systems Review: Minimum Phase Systems General Linear Phase Systems ( ) ( ) ( ) Y e jω = H e jω X e jω! We can define a magnitude response ( ) ( ) ( ) Y e jω = H e jω X e jω! And a phase response ( ) ( ) ( ) Y e jω = H e jω + X e jω Penn ESE 531 Spring 2017 Khanna 3 4 2nd Order IIR with Complex Poles Example: Zero on Real Axis! Penn ESE 531 Spring 2017 Khanna For θ 0 r=0.9, θ=π/4 phase magnitude group delay Penn ESE 531 Spring 2017 Khanna 5 6 1
3 rd Order IIR Example 3 rd Order IIR Example 7 8 All-Pass Filters All-Pass Systems! A system is an all-pass system if! Its phase response may be non-trivial 9 10 General All-Pass Filter! d k =real pole, e k =complex poles paired w/ conjugate, e k * General All-Pass Filter! d k =real pole, e k =complex poles paired w/ conjugate, e k *! Example: d k = 3 4 e k = 0.8e jπ 4 Real zero/pole Complex zeros/poles 11 12 2
All Pass Filter Phase Response All Pass Filter Phase Response! First order system H(e jω ) = e jω a * 1 ae jω = e jω re jθ 1 re jθ e jω! First order system H(e jω ) = e jω a * 1 ae jω = e jω re jθ 1 re jθ e jω! phase! phase arg e jω re jθ 1 re jθ e jω = arg e jω (1 re jθ e jω ) 1 re jθ e jω = arg(e jω ) + arg(1 re jθ e jω ) arg(1 re jθ e jω ) = ω arg(1 re jθ e jω ) arg(1 re jθ e jω ) = ω 2arg(1 re jθ e jω ) 13 14 First Order Example First Order Example! Magnitude:! Phase: 15 16 First Order Example All Pass Filter Phase Response! Group Delay:! Second order system with poles at z = re jθ,re jθ 17 18 3
Second Order Example! Poles at z = 0.9e ± jπ 4 (zeros at conjugates) All-Pass Properties! Claim: For a stable all-pass system: " arg[h ap (e jω )] 0 " Unwrapped phase always non-positive and decreasing " grd[h ap (e jω )]>0 " Group delay always positive " Intuition " delay is positive # system is causal " Phase negative # phase lag 19 20 Minimum-Phase Systems Minimum-Phase Systems! Definition: A stable and causal system H(z) (i.e. poles inside unit circle) whose inverse 1/H(z) is also stable and causal (i.e. zeros inside unit circle) " All poles and zeros inside unit circle 22 Minimum-Phase Systems All-Pass Min-Phase Decomposition! Definition: A stable and causal system H(z) (i.e. poles inside unit circle) whose inverse 1/H(z) is also stable and causal (i.e. zeros inside unit circle)! Any stable, causal system can be decomposed to: H(z) = H min (z) H ap (z) " All poles and zeros inside unit circle H(z) 1/H(z)! Approach: " (1) First construct H ap with all zeros outside unit circle " (2) Compute H min (z) = H(z) H ap (z) 23 24 4
Min-Phase Decomposition Example Min-Phase Decomposition Example H(z) = 1 3z 1 H(z) H(z) = 1 3z 1 H(z) 1 1 2 z 1 1 1 2 z 1 1/2 3 1/2 3! Set z 1 1 H ap (z) = 3 1 1 3 z 1 H ap (z) H min (z) 1/2 1 1 H min (z) = 3 3 z 1 1 1 2 z 1 1/3 3 25 26 Min-Phase Decomposition Purpose Min-Phase Decomposition Purpose! Have some distortion that we want to compensate for:! Have some distortion that we want to compensate for:! If H d (z) is min phase, easy: " H c (z)=1/h d (z) $ also stable and causal 27 28 Min-Phase Decomposition Purpose Minimum Energy-Delay Property! Have some distortion that we want to compensate for: Min phase! If H d (z) is min phase, easy: " H c (z)=1/h d (z) $ also stable and causal! Else, decompose H d (z)=h d,min (z) H d,ap (z) " H c (z)=1/h d,min (z) #H d (z)h c (z)=h d,ap (z) " Compensate for magnitude distortion 29 30 5
Minimum Energy-Delay Property Min phase Max phase Generalized Linear Phase Systems 31 32 Generalized Linear Phase! An LTI system has generalized linear phase if frequency response can be expressed as: Generalized Linear Phase! An LTI system has generalized linear phase if frequency response can be expressed as:! Where A(ω) is a real function.! Where A(ω) is a real function.! What is the group delay? 33 34 Causal FIR Systems Causal FIR Systems! Causal FIR systems have generalized linear phase if they have impulse response length (M+1)! It can be shown if! then 35 36 6
Example: Moving Average! Moving Average Filter " Causal: M 1 =0, M 2 =M Impulse response y[n] = x[n M ]+...+ x[n] M +1 Example: Moving Average! Moving Average Filter " Causal: M 1 =0, M 2 =M Impulse response y[n] = x[n M ]+...+ x[n] M +1 Scaled &Time Shifted window 37 38 Example: Moving Average Example: Moving Average ( ) sin( ω 2) sin (N +1 2)ω w[n] W (e jω ) = ( ) sin( ω 2) 1 M +1 w[n M 2] W (e jω ) = e jωm 2 sin (M 2 +1 2)ω M +1 39 40 Causal FIR Systems FIR GLP: Type I! Causal FIR systems have generalized linear phase if they have impulse response length (M+1)! It can be shown if! Then! Sufficient conditions to guarantee GLP, not necessary " Causal IIR can also have GLP 41 42 7
FIR GLP: Type I FIR GLP: Type I 43 44 FIR GLP: Type I Example, M=4 FIR GLP: Type II 45 46 FIR GLP: Type II FIR GLP: Type II 47 48 8
FIR GLP: Type II Example, M=3 FIR GLP: Type I and II 49 50 FIR GLP: Type III FIR GLP: Type III 51 52 FIR GLP: Type III FIR GLP: Type III Example, M=4 53 54 9
FIR GLP: Type IV FIR GLP: Type IV 55 56 FIR GLP: Type IV FIR GLP: Type IV Example, M=3 57 58 FIR GLP: Type III and IV Zeros of GLP System Type I and II 59 60 10
Zeros of GLP System Type I and II Zeros of GLP System Type I and II! If h[n] is real, * 61 62 Zeros of GLP System Type I and II Zeros of GLP System Type I and II! If zero is on unit circle (r=1)! If zero is real and not on unit circle (θ=0) 63 64 Zeros of GLP System Type I and II Zeros of GLP System Type II 65 66 11
Zeros of GLP System Type I and II FIR GLP: Type I and II Type I Type II 67 68 Zeros of GLP System Type III and IV Zeros of GLP System Type III and IV 69 70 Zeros of GLP System Type III and IV Zeros of GLP System Type III and IV! If zero is on unit circle (r=1)! If zero is real and not on unit circle (θ=0) 71 72 12
Zeros of GLP System Type III and IV Zeros of GLP System Type III and IV Type III Type IV 73 74 FIR GLP: Type III and IV GLP and Min Phase Systems! Any FIR linear-phase system can be decomposed into:! A min phase system, system containing only zeros on unit circle, and max phase system 75 76 GLP and Min Phase Systems Big Ideas! Any FIR linear-phase system can be decomposed into:! Frequency Response of LTI Systems " Magnitude Response, Phase Response, Group Delay! All Pass Systems " Used for delay compensation! Minimum Phase Systems! A min phase system, system containing only zeros on unit circle, and max phase system " Can compensate for magnitude distortion " Minimum energy-delay property! Generalized Linear Phase Systems " Useful for design of causal FIR filters 77 Penn ESE 531 Spring 2017 Khanna 78 13
Admin! HW 6 " Out now " Due Tuesday 3/28 " No MATLAB problem! Midterm returned on Thursday " Exam and solutions will be posted 79 14