CHAPTER 7. The Discrete Fourier Transform 436

Similar documents
Discrete Fourier Transform

Fast Fourier Transform Discrete-time windowing Discrete Fourier Transform Relationship to DTFT Relationship to DTFS Zero padding

Fast Fourier Transform Discrete-time windowing Discrete Fourier Transform Relationship to DTFT Relationship to DTFS Zero padding

LAB 2: DTFT, DFT, and DFT Spectral Analysis Summer 2011

Digital Signal Processing: Signal Transforms

Lecture 10. Digital Signal Processing. Chapter 7. Discrete Fourier transform DFT. Mikael Swartling Nedelko Grbic Bengt Mandersson. rev.

Overview of Discrete-Time Fourier Transform Topics Handy Equations Handy Limits Orthogonality Defined orthogonal

Lecture 19: Discrete Fourier Series

ECE 301 Fall 2010 Division 2 Homework 10 Solutions. { 1, if 2n t < 2n + 1, for any integer n, x(t) = 0, if 2n 1 t < 2n, for any integer n.

Discrete Fourier Transform

x[n] = x a (nt ) x a (t)e jωt dt while the discrete time signal x[n] has the discrete-time Fourier transform x[n]e jωn

Digital Signal Processing Lab 3: Discrete Fourier Transform

EEO 401 Digital Signal Processing Prof. Mark Fowler

Multimedia Signals and Systems - Audio and Video. Signal, Image, Video Processing Review-Introduction, MP3 and MPEG2

Discrete-time Fourier transform (DTFT) representation of DT aperiodic signals Section The (DT) Fourier transform (or spectrum) of x[n] is

NAME: 11 December 2013 Digital Signal Processing I Final Exam Fall Cover Sheet

VII. Discrete Fourier Transform (DFT) Chapter-8. A. Modulo Arithmetic. (n) N is n modulo N, n is an integer variable.

X. Chen More on Sampling

Review of Discrete-Time System

! Circular Convolution. " Linear convolution with circular convolution. ! Discrete Fourier Transform. " Linear convolution through circular

EEL3135: Homework #3 Solutions

Chap 4. Sampling of Continuous-Time Signals

ELEN 4810 Midterm Exam

DSP Laboratory (EELE 4110) Lab#5 DTFS & DTFT

Overview of Sampling Topics

/ (2π) X(e jω ) dω. 4. An 8 point sequence is given by x(n) = {2,2,2,2,1,1,1,1}. Compute 8 point DFT of x(n) by

EE 224 Signals and Systems I Review 1/10

Discrete Fourier transform (DFT)

Discrete-Time Fourier Transform (DTFT)

Digital Signal Processing Lecture 10 - Discrete Fourier Transform

ECE-314 Fall 2012 Review Questions for Midterm Examination II

Core Concepts Review. Orthogonality of Complex Sinusoids Consider two (possibly non-harmonic) complex sinusoids

Fall 2011, EE123 Digital Signal Processing

Frequency-domain representation of discrete-time signals

Chapter 4 Discrete Fourier Transform (DFT) And Signal Spectrum

Chapter 6: Applications of Fourier Representation Houshou Chen

Random signals II. ÚPGM FIT VUT Brno,

Interchange of Filtering and Downsampling/Upsampling

Solutions - Homework # 3

Signals & Systems. Lecture 4 Fourier Series Properties & Discrete-Time Fourier Series. Alp Ertürk

Digital Signal Processing. Midterm 2 Solutions

LAB # 5 HANDOUT. »»» The N-point DFT is converted into two DFTs each of N/2 points. N = -W N Then, the following formulas must be used. = k=0,...

Good Luck. EE 637 Final May 4, Spring Name: Instructions: This is a 120 minute exam containing five problems.

Signals and Systems Spring 2004 Lecture #9

Discrete-time Signals and Systems in

FFT Octave Codes (1B) Young Won Lim 7/6/17

Summary of lecture 1. E x = E x =T. X T (e i!t ) which motivates us to define the energy spectrum Φ xx (!) = jx (i!)j 2 Z 1 Z =T. 2 d!

J. McNames Portland State University ECE 223 DT Fourier Series Ver

Digital Signal Processing. Midterm 1 Solution

Chapter 5. Fourier Analysis for Discrete-Time Signals and Systems Chapter

VU Signal and Image Processing. Torsten Möller + Hrvoje Bogunović + Raphael Sahann

DISCRETE FOURIER TRANSFORM

Discrete Time Signals and Systems Time-frequency Analysis. Gloria Menegaz

DFT Octave Codes (0B) Young Won Lim 4/15/17

! Spectral Analysis with DFT. ! Windowing. ! Effect of zero-padding. ! Time-dependent Fourier transform. " Aka short-time Fourier transform

Sinc Functions. Continuous-Time Rectangular Pulse

Final Exam of ECE301, Section 3 (CRN ) 8 10am, Wednesday, December 13, 2017, Hiler Thtr.

J. McNames Portland State University ECE 223 Sampling Ver

Introduction to DFT. Deployment of Telecommunication Infrastructures. Azadeh Faridi DTIC UPF, Spring 2009

ENT 315 Medical Signal Processing CHAPTER 2 DISCRETE FOURIER TRANSFORM. Dr. Lim Chee Chin

Homework 6 Solutions

Windowing Overview Relevance Main lobe/side lobe tradeoff Popular windows Examples

The Fourier Transform (and more )

LABORATORY 1 DISCRETE-TIME SIGNALS

HW13 Solutions. Pr (a) The DTFT of c[n] is. C(e jω ) = 0.5 m e jωm e jω + 1

INTRODUCTION TO THE DFS AND THE DFT

EDISP (NWL2) (English) Digital Signal Processing Transform, FT, DFT. March 11, 2015

Complex symmetry Signals and Systems Fall 2015

Homework 3 Solutions

Signals and Systems. Lecture 14 DR TANIA STATHAKI READER (ASSOCIATE PROFESSOR) IN SIGNAL PROCESSING IMPERIAL COLLEGE LONDON

Digital Signal Processing Chapter 10. Fourier Analysis of Discrete- Time Signals and Systems CHI. CES Engineering. Prof. Yasser Mostafa Kadah

7.16 Discrete Fourier Transform

Discrete-Time Fourier Transform

Review of Fundamentals of Digital Signal Processing

8 The Discrete Fourier Transform (DFT)

TTT4120 Digital Signal Processing Suggested Solutions for Problem Set 2

Module 3. Convolution. Aim

ECE 301: Signals and Systems Homework Assignment #7

Signals and Systems Profs. Byron Yu and Pulkit Grover Fall Midterm 2 Solutions

Final Exam ECE301 Signals and Systems Friday, May 3, Cover Sheet

Solutions to Problems in Chapter 4

DIGITAL SIGNAL PROCESSING LABORATORY

! Introduction. ! Discrete Time Signals & Systems. ! Z-Transform. ! Inverse Z-Transform. ! Sampling of Continuous Time Signals

ECGR4124 Digital Signal Processing Midterm Spring 2010

Digital Signal Processing

Music 270a: Complex Exponentials and Spectrum Representation

University of Connecticut Lecture Notes for ME5507 Fall 2014 Engineering Analysis I Part III: Fourier Analysis

Discrete-Time Signals and Systems

MAHALAKSHMI ENGINEERING COLLEGE-TRICHY

Review: Continuous Fourier Transform

Fourier Analysis of Signals Using the DFT

Fourier analysis of discrete-time signals. (Lathi Chapt. 10 and these slides)

Digital Signal Processing Module 6 Discrete Fourier Transform (DFT)

sinc function T=1 sec T=2 sec angle(f(w)) angle(f(w))

CCNY. BME I5100: Biomedical Signal Processing. Linear systems, Fourier and z-transform

ELEN E4810: Digital Signal Processing Topic 11: Continuous Signals. 1. Sampling and Reconstruction 2. Quantization

Department of Electrical and Computer Engineering Digital Speech Processing Homework No. 6 Solutions

BME 50500: Image and Signal Processing in Biomedicine. Lecture 2: Discrete Fourier Transform CCNY

ESE 531: Digital Signal Processing

Assignment 4 Solutions Continuous-Time Fourier Transform

Transcription:

CHAPTER 7. The Discrete Fourier Transform 36 hfa figconfg( P7a, long ); subplot() stem(k,abs(ck), filled );hold on stem(k,abs(ck_approx), filled, color, red ); xlabel( k, fontsize,lfs) title( Magnitude Response, fontsize,tfs) legend( c_k, \tilde{c}_k, location, best ) subplot() stem(k,angle(ck), filled );hold on stem(k,angle(ck_approx), filled, color, red ); xlabel( k, fontsize,lfs) title( Phase Response, fontsize,tfs) legend( c_k, \tilde{c}_k, location, best ) 3. (a) Solution: The DTFT of (.9) n u[n] is: The DTFT of x[n] is: (b) See plot below. (c) See plot below. (d) See plot below. MATLAB script:.9e jω X ( e jω) ( j) d dω (.9e jω (.9e jω ).9e jω % P73: umerical DFT approximation of DTFT close all; clc w linspace(,,)*pi; X.9*exp(-j*w)./(-.9*exp(-j*w)).^; ; % Part (b) % 5; % Part (c) % ; % Part (d) n :-; )

CHAPTER 7. The Discrete Fourier Transform 37 Magnitude Response Phase Response 8 6 X(e jω ) X (e jω ) X(e jω ) X (e jω ).5.5.5.5 FIGURE 7.6: Magnitude and phase responses of X ( e jω) and X ( e jω ) when. Magnitude Response Phase Response 8 6 X(e jω ) X (e jω ) X(e jω ) X (e jω ).5.5.5.5 FIGURE 7.7: Magnitude and phase responses of X ( e jω) and X ( e jω ) when 5. xn n.*.9.^n; X fft(xn); wk /*(:-); %% Plot: hfa figconfg( P73a, long ); subplot() plot(w/pi,abs(x));hold on stem(wk,abs(x), filled, color, red ); xlabel( \omega/\pi, fontsize,lfs) title( Magnitude Response, fontsize,tfs) legend( X(e^{j\omega}), X_(e^{j\omega}), location, best ) subplot() plot(w/pi,angle(x));hold on

CHAPTER 7. The Discrete Fourier Transform 38 Magnitude Response Phase Response 8 6 X(e jω ) X (e jω ) X(e jω ) X (e jω ).5.5.5.5 FIGURE 7.8: Magnitude and phase responses of X ( e jω) and X ( e jω ) when. stem(wk,angle(x), filled, color, red ); xlabel( \omega/\pi, fontsize,lfs) title( Phase Response, fontsize,tfs) legend( X(e^{j\omega}), X_(e^{j\omega}), location, best ). (a) Solution: The DFT matrix is: W W W...... W W ( )( ) Thekth column of W is w k [W k...w k] T. Thei,jth element of W is: ( ) W i,j ( W T W ) i,j wt i w j {, i+j, i+j Hence, we proved that W.... J........

CHAPTER 7. The Discrete Fourier Transform 8. (a) See plot below. (b) See plot below. (c) See plot below. MATLAB script: % P78: Regenerate Figure~7.5 and Example 7.3 close all; clc 6; a.9; % Part (a) % 8; a.8; % Part (b) % 6; a.8; % Part (c) wk *pi/*(:-); Xk./(-a*exp(-j*wk)); xn real(ifft(xk)); w linspace(,,)*pi; X fft(xn,length(w)); X_ref./(-a*exp(-j*w)); n :-; xn_ref a.^n; %% Plot: hfa figconfg( P78a, small ); plot(w/pi,abs(x_ref), color, black ); hold on plot(w/pi,abs(x)) stem(wk/pi,abs(xk), filled ); ylim([ max(abs(x))]) xlabel( \omega/\pi, fontsize,lfs) ylabel( Magnitude, fontsize,lfs) title( Magnitude of Spectra, fontsize,tfs) hfb figconfg( P78b, small ); plot(n,xn,. ); hold on plot(n,xn_ref,., color, black ) ylim([.*max(xn)]) xlim([ -]) xlabel( Time index (n), fontsize,lfs) ylabel( Amplitude, fontsize,lfs) title( Signal Amplitudes, fontsize,tfs) legend( x[n], \tilde{x}[n], location, northeast )

CHAPTER 7. The Discrete Fourier Transform 3 Magnitude of Spectra 8 Magnitude 6.5.5 (a) Signal Amplitudes Amplitude..8.6. x[n] \tilde{x}[n]. 5 5 Time index (n) (b) FIGURE 7.9: (a) Magnitude response of the DTFT signal. (b) Time sequence and reconstructed time sequence.

CHAPTER 7. The Discrete Fourier Transform 5 Magnitude of Spectra Magnitude 3.5.5 (a) Signal Amplitudes Amplitude..8.6. x[n] \tilde{x}[n]. 3 5 6 7 Time index (n) (b) FIGURE 7.: (a) Magnitude response of the DTFT signal and (b) time sequence and reconstructed time sequence for a.8 and 8.

CHAPTER 7. The Discrete Fourier Transform 5 5 Magnitude of Spectra Magnitude 3.5.5 (a) Signal Amplitudes Amplitude.8.6. x[n] \tilde{x}[n]. 3 5 6 Time index (n) (b) FIGURE 7.: (a) Magnitude response of the DTFT signal and (b) time sequence and reconstructed time sequence for a.8 and 6.

CHAPTER 7. The Discrete Fourier Transform 9 3. (a) Proof: If k is even and is even, the correspondent DFT is: X[k] / / x[n]e jπ nk / x[n]e jπ nk + n/ (x[n]e jπ nk +x[n+ ]e jπ (n+ )k ) (x[n] x[n])e jπ nk (b) Proof: If m,k l, the correspondent DFT is: x[n]e jπ nk X[l] + x[n]e jπ n(l) x[n]e jπ n(l) + 3 x[n]e jπ n(l) x[n]e jπ n(l) + x[n]e jπ n(l) n n n 3 x[n]e jπ n(l) + x[n+ ]e jπ (n+ )(l) 3 + x[n]e jπ n(l) + n 3 n (x[n]+x[n+ ) ] e jπ n(l) + x[n+ ]e jπ (n+ )(l) 3 n. (a) Solution: Solving the circular convolution using hand calculation: 3 6 5 7 (x[n]+x[n+ ] ) e jπ n(l)

CHAPTER 7. The Discrete Fourier Transform 5 (b) See script below. (c) See script below. MATLAB script: % P7: Circular convolution close all; clc xn :5; xn [ - -]; %% Part (b): xn circonv(xn,[xn ] ); %% Part (c): max(length(xn),length(xn)); Xk fft(xn,); Xk fft(xn,); Xk Xk.*Xk; xn_dft ifft(xk); 5. (a) Proof: X [K] can be obtained by frequency sampling of X 3 [k], hence in the time domain, according the aliasing equation, we have x [n] l x 3 [n+l] (b) Proof: When L, there is no time aliasing, we conclude x [n] x 3 [n], for n L When max(, ) < L, since L +, we conclude that x [n] x 3 [n]+x 3 [n+], for n Hence, we proved the equation (??). (c) MATLAB script: % P75: Verify formula in Problem 75 close all; clc xn :; xn :-:;

CHAPTER 7. The Discrete Fourier Transform 55 Linearity. w c (t) (a x c (t)+a x c (t)) a w c (t)x c (t)+a w c (t)x c (t) a x c (t)+a x c (t) Time-varying. In general, w c (t τ)x c (t τ) w c (t)x c (t). Proof: Hence, x c (t τ) x c (t). (b) Proof: If t T, and n L, we have x[n] w[n]x[n] (7.6) x c (nt) w c (nt)x c (nt) x c (nt) x[n] x[n] w[n]x[n] x[n] If t > T, and > L, we have The CTFT of ˆx c (t) is: ˆX c (jω) x c (nt) x[n] ˆX c (jω) X c (jθ)w c (j(ω θ))dθ (7.7) π π π ˆx c (t)e jωt dt ( w c (t)e jωt π ( X c (jθ) w c (t)x c (t)e jωt dt ) X c (jθ)e jθt dθ dt w c (t)e )dθ j(ω θ)t X c (jθ)w c (j(ω θ))dθ. Proof: The CTFT of x c (at) is: Scaling Property: x c (at)e jωt dt a x c (at) CTFT a X c ( ) jω a x c (at)e jω a at dat (7.7)

CHAPTER 7. The Discrete Fourier Transform 56 If a >, we have a If a <, we have a x c (at)e jω a at dat a x c (at)e jω a at dat a Hence, we proved the scaling property. 3. Proof: x c (t)x c (t)dt Suppose a is a real number, define function p(a) as where A p(a) x c (t)e jω a t dt a X c x c (t)e jω a t dt a X c ( ) jω a ( ) jω a x c (t) dt x c (t) dt (7.79) (a x c (t)+x c (t)) dt Aa +Ba+C x c(t)dt, B x c (t)x c (t)dt, C Since we have B AC, that isb AC, ( x c (t)x c (t)dt) x c(t)dt x c(t)dt. Proof: The CTFT of generic window is: x c(t)dt. W(jΩ) aw R (jω)+bw R (j(ω π)/t )+bw R (j(ω+π)/t ) (7.89) The ICTFT is: π [aw R (jω)+bw R (j(ω π)/t )+bw R (j(ω+π)/t )]e jωt dω aw R (t)+be jπ t T wr (t)+be jπ )] [ a+bcos ( π T t w R (t) T t wr (t)

CHAPTER 7. The Discrete Fourier Transform 66 6 Magnitude Response Phase Response 5 X(e jω ) X (e jω ) 3 X(e jω ) X (e jω ).5.5.5.5 FIGURE 7.: Magnitude and phase responses of X ( e jω) and X ( e jω ) when. 3. tba title( Magnitude Response, fontsize,tfs) legend( X(e^{j\omega}), X_(e^{j\omega}), location, best ) subplot() plot(w/pi,angle(x));hold on stem(wk,angle(x), filled, color, red ); xlabel( \omega/\pi, fontsize,lfs) title( Phase Response, fontsize,tfs) legend( X(e^{j\omega}), X_(e^{j\omega}), location, best ) 3. (a) See plot below. (b) See plot below. (c) See plot below. (d) See plot below. MATLAB script: % P73: Compute and plot DFT and IDFT close all; clc %% Part (a): 8; n :-; xn zeros(size(n)); xn() ; %% Part (b):

CHAPTER 7. The Discrete Fourier Transform 67 Magnitude Response 3 Phase Response.8.6.. 3 5 6 7 k (a) 3 3 5 6 7 k Time Sequence.8.6.. 6 6 8 Time index (n) (b) FIGURE 7.: -point (a) DFT and (b) IDFT of x[n] δ[n], 8 in the range ( ) n ( ). % ; % n :-; % xn n; %% Part (c): % 3; % n :-; % xn cos(6*pi*n/5); %% Part (d): % 3; % n :-; % xn cos(.*pi*n); Xk fft(xn);

CHAPTER 7. The Discrete Fourier Transform 68 5 Magnitude Response 3 Phase Response 3 6 8 k (a) 3 6 8 k Time Sequence 8 6 5 5 5 Time index (n) (b) FIGURE 7.3: -point (a) DFT and (b) IDFT of x[n] n, in the range ( ) n ( ). ind abs(xk) < e-; Xk(ind) ; xn_ref ifft(xk); nn -(-):*-; xn_plot xn_ref(mod(nn,)+); %% Plot: hfa figconfg( P73a, long ); subplot() stem(n,abs(xk), filled ); xlim([ -]) xlabel( k, fontsize,lfs) title( Magnitude Response, fontsize,tfs) subplot() stem(n,angle(xk), filled );

CHAPTER 7. The Discrete Fourier Transform 69 5 Magnitude Response 3 Phase Response 5 5 5 5 k (a) 3 5 5 5 k Time Sequence.5.5 3 5 Time index (n) (b) FIGURE 7.: -point (a) DFT and (b) IDFT of x[n] cos(6πn/5), 3 in the range ( ) n ( ). xlim([ -]) ylim([-pi pi]) xlabel( k, fontsize,lfs) title( Phase Response, fontsize,tfs) hfb figconfg( P73b, long ); stem(nn,xn_plot, filled ) xlim([nn() nn(end)]) xlabel( Time index (n), fontsize,lfs) title( Time Sequence, fontsize,tfs)

CHAPTER 7. The Discrete Fourier Transform 7 5 Magnitude Response 3 Phase Response 5 5 5 5 k (a) 3 5 5 5 k Time Sequence.5.5 3 5 Time index (n) (b) FIGURE 7.5: -point (a) DFT and (b) IDFT of x[n] cos(.πn), 3 in the range ( ) n ( ). 3. (a) Solution: The DFS of x[n] and x 3 [n] can be written as: X[k] X[ k ] X 3 [k] X 3 [ k 3 ] 3 x[n]e jπ n k π j x[n]e 3 n k 3

CHAPTER 7. The Discrete Fourier Transform 7 We have 3 π j X 3 [k] x[n]e 3 n k 3 π j x[n]e 3 n k 3 + n π j x[n]e 3 n k 3 + + ( 3 X[k/3] (b) MATLAB script: π j x[n+]e 3 (n+) k 3 x[n]e jπ n k ) % P73: Matlab verification close all; clc xn [ 3 3 3]; Xk fft(xn(:)); X3k fft(xn); 3 π j x[n]e 3 n k 3 + n π j x[n+]e 3 (n+) k 3 ( ) + e jπ 3 k 3 + e jπ 3 k 3 π j x[n]e 3 n k 3

CHAPTER 7. The Discrete Fourier Transform 79 (c) By applying the correlation property, the DFT of x 3 [n] is: X 3 [k] X[k]X [k] X[k] (d) By applying the windowing property, the DFT of x [n] is: X [k] 9 X[k] 9 X[k] (e) By applying the frequency-shifting property, the DFT of x 5 [n] is: 39. (a) Proof: X[] X 5 [k] X[ k + 9 ] x[n]w n which is real-valued since x[n] is real-valued. (b) Proof: If k, since x[] is real, we have If k, we have X[ k ] X [k] X[] X [] x[n] x[n]w n k x[n]w nk ( x[n]w n( k) x[n]w nk Hence, we proved X[ k ] X [k] for every k. (c) Proof: ) X[/] x[n]w n x[n]e jnπ x[n] cos(nπ) which is real-valued since x[n] and cos(nπ) are both real-valued.