chapter3 Frequency Spectrum of Signal Fourier Transform Analysis of networks <Principles of Circuit Analysis) 53-1 Frequeney Spectrum of periodie signal: Fourier Chapter 3: Fourier Transform Analysis 3-2 Frequency Spectrum of non-periodic signals ourier transform 2009-10-22 53-3 Basic theorem of frequency analysis 83-4 Outputs of constant-parameter linear circuits Content for todays lecture Focuses a From Fourier series to fourier transform D Fourier Transform Analysis of non-periodic signals a Fourier Transform(definition, Dirichlet condition) spectral density, spectral distribution, speetral bandwidth a Physical implications of Fourier Transform o Time domain characteristics of signals Corresponding relation Frequency domain (spectral density, spectral distribution, spectral bandwidth) characteristies u Fourier Transform and Laplace Transform u Learning the basie properties from frequency domain analysis 口 u Frequeney domain characteristics of basic signals ral density of basic signals 33-2 Physical implications of Fourier Transform +p $ 3-2: Fourier Transform-definition, Dirichlet conditions *in f(t)=∑c ∫nf(t)elud Theorem description If a non-periodic signal f(t) meets Dirichlet conditions in the interval (-oo, to limTcn=liminf F(o)=f(t)e dt Cf(t)e iut dt=F(o)# image Abbreviated as:FIf()]=F(co) F(o is the mapping of f(t)in g domain, which is called Definition spectral density= 2* times the line the Fourier Transform of f(t) amplitude per unit frequency bandwidth Tc,=2 C,/wo Inverse transfom: f(t)=2F(o)-e wdo Definition: F(@)is the spectral density of fo) in frequency Abbreviated as:FF(o)=f(t) domain
北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 第 ?讲: 复习 北京大学 wwhu 北京大学 《Principles of Circuit Analysis》 Chapter 3: Fourier Transform Analysis Lecture 2 2009-10-22 Interest Focus Persistence Originality 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 第三章:信号的频谱 §3-1 Frequency Spectrum of periodic signal: Fourier Series §3-2 Frequency Spectrum of non-periodic signals: Fourier Transform §3-3 Basic theorem of frequency analysis §3-4 Outputs of constant-parameter linear circuits chapter3 Frequency Spectrum of Signals (Fourier Transform Analysis of networks) chapter3 Frequency Spectrum of Signals (Fourier Transform Analysis of networks) 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Content for today’s lecture From Fourier Series to Fourier Transform Fourier Transform (definition, Dirichlet condition) Physical implications of Fourier Transform (spectral density, spectral distribution, spectral bandwidth) Fourier Transform and Laplace Transform Basic property (theoremes) Spectral density of basic signals 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Focuses Fourier Transform Æ Analysis of non-periodic signals spectral density, spectral distribution, spectral bandwidth Time domain characteristics of signals Åcorresponding relationÆFrequency domain characteristics Learning the basic properties from frequency domain analysis Frequency domain characteristics of basic signals 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu Definition: F( domain. 北京大学ω) is the spectral density of f(t) in frequency wwhu 北京大学 Definition :spectral density= 2π times the line amplitude per unit frequency bandwidth §3-2 Physical implications of Fourier Transform *** Tcn 2 cn/ω0 = π⋅ ∑ ∞ = −∞ = n jn ω t n e 0 f(t) c ∫ + = T/2 -T/2 - jn ω t n f(t) e dt T 1 c 0 f(t) e dt F(ω) limTc lim f(t) e dt - -jωt T/2 -T/2 -jnω t T n T 0 = ⋅ = = ⋅ ∫ ∫ +∞ ∞ + →∞ →∞ image function 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 §3-2:Fourier Transform – definition, Dirichlet conditions Theorem description: If a non-periodic signal f(t) meets Dirichlet conditions in the interval (-∞,+∞), then: ∫ +∞ ∞ = ⋅ - -jωt F(ω) f(t) e dt F [f(t)] = F(ω) F(ω) e dω 2 1 f(t) jωt = ⋅ ∫ +∞ π −∞ is the mapping of f(t) in domain, which is called the Fourier Transform of f(t). F(ω) ω Inverse transform: Abbreviated as: F [F(ω)] = f(t) -1 *** Abbreviated as:
3-2: Fourier Transform- definition. Dirichlet conditions 33-2: Fourier Transform and Laplace Transform *ak O Dirichlet conditions for non-periodic signals: Fourier Transform: F(s)=f(t)e*dt Laplace Transform: F(o)=f(t)e udt f(x)must have a finite number of extrema in any given nterval F(s)in=F(o) f(x)must have a finite number of discontinuities in any Laplace Transform can be transformed to Fourier given interval Transform if let 0=0 in s=0+J0 f(x) must be absolutely integrable over a period f(x)must be bounded From the view of convergence: F(s)=.f(t)e ot,e dt → Sometimes, a signal may have a Fourier If f(t)is not bounded, which makes If(t)I dt does not exist. Transform even if it is not bounded f(t)e°| dt exists. Usually, we use Fourier Transform to analysis the frequency characteristics of signal and use Laplace Transform to solve 53.2: Fourier Tra frequeney characteristies of three basic signal frequeney characteristies of three basie signal Qunit-pulse signal 5(0) flat Junit-pulse signal: 8(0)+1 F(o F(o)=f(t)e u dt=1 F(ao=f(t) 8(t) The network response of a signal whose frequency characteristics is flat 目 PUnit constant (DC) signal:1…2a can describe the frequency response of this network. Another representation of unit-pulse signal 8(0): 20a2(01-1t2 (t)=F(1)=ed 53-2: Fourier Transform-frequeney characteristies of three hasic signal §3-2: Fourier T rm-frequency characteristics of three basie signals a Complex exponential signal aUnit symbol signal sgn(0)*Fo)= en(r) ejoot *2nd(o-w) F(o=f(t)e dt el]= 1/2 Unit constant(DC)signal: 1 ++2m8() 厂e F(o )=2n(-a)=1 glim jo-a e-/ f 2 Jo 2 Jojo
北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 ◆Dirichlet conditions for non-periodic signals: Sometimes, a signal may have a Fourier Transform even if it is not bounded. §3-2:Fourier Transform – definition, Dirichlet conditions * • f(x) must have a finite number of extrema in any given interval . • f(x) must have a finite number of discontinuities in any given interval . • f(x) must be absolutely integrable over a period. • f(x) must be bounded . 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 F(s) F(ω) s jω = = Laplace Transform can be transformed to Fourier Transform if let in σ =0 s = σ + jω §3-2:Fourier Transform and Laplace Transform ∫ +∞ ∞ = ⋅ - -jωt F(ω) f(t) e dt ∫ +∞ ∞ = ⋅ - -st Fourier Transform: F(s) f(t) e dt Laplace Transform: *** Usually, we use Fourier Transform to analysis the frequency characteristics of signal and use Laplace Transform to solve the response of linear circuits. From the view of convergence: If f(t) is not bounded, which makes does not exist. We can always find a which makes exists. 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 §3-2: Fourier Transform — frequency characteristics of three basic signals unit-pulse signal δ(t) F(ω) f(t) e dt 1 - -jωt = ⋅ = ∫ +∞ ∞ ω F(ω) 1 +∞ ∞ = ∫ π -1 jωt - 1 δ(t)=F (1) e dω 2 Another representation of unit-pulse signal δ(t): The network response of a signal whose frequency characteristics is flat can describe the frequency response of this network. *** flat Frequency-domain H(j H(jωω)) X(jω) Y(jω) h(t) h(t) x(t) y(t) input (stimulation) output (response) Time-domain y(t) = F{ x(t), h(t) } Y(jω) = H(jω) · X(jω) input (stimulation) output (response) 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 unit-pulse signal : δ(t) Unit constant (DC) signal: 1 F(ω) f(t) e dt 1 - -jωt = ⋅ = ∫ +∞ ∞ ω F(ω) 1 ∫ +∞ ∞ = - jωt e dω 2 1 δ(t) π *** +∞ ∞ = ∫ π jωt - 1 δ(ω) e dt 2 ω F(ω) +∞ (2 ) π ∞ = =⋅ ∫ π π -jωt - 2 δ(ω) 2 δ(-ω) 1 e dt 1 2πδ(ω) §3-2: Fourier Transform — frequency characteristics of three basic signals 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Complex exponential signal ejωot Complex exponential signal ejωot ω F(ω) ω0 2πδ(ω 0 -ω ) [ ]= jω t F e 0 ∫ +∞ ∞ ⋅ - jω t -jωt e e dt 0 *** +∞ ∞ = ∫ π jωt - 1 δ(ω) e dt 2 ω F(ω) +∞ (2 ) π ∞ = =⋅ ∫ π π -jωt - 2 δ(ω) 2 δ(-ω) 1 e dt ■Unit constant (DC) signal: 1 2πδ(ω) §3-2: Fourier Transform — frequency characteristics of three basic signals 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Unit symbol signal sgn(t) Unit symbol signal sgn(t) t sgn( )t 1/2 -1/2 * +∞ ∞ = ⋅ ∫ -jωt - F(ω) f(t) e dt ω ω ω ω ω ω ω ω ω ω ω ω j j j e j a e j a e e dt e e dt F e dt e dt j a t a j a t a at j t a at j t a j t j t 1 1 2 1 1 2 1 | 1 2 1 | 1 2 1 2 1 2 1 2 1 ) 2 1 ( ) ( 0 ( ) 0 0 ( ) 0 0 0 0 0 0 0 lim lim lim lim = ⋅ + ⋅ = ⋅ − − ⋅ + − = − = − ⋅ + ⋅ = − ⋅ + ⋅ − − ∞ → − ∞ → − ∞ − → ∞ − → − ∞ − −∞ ∫ ∫ ∫ ∫ ω ω j F 1 ( ) = §3-2: Fourier Transform — frequency characteristics of three basic signals
JUnit step signal u(t) DUnit step signal u(t) method 1 method 2: using the f.t. of known signals First: Fe au(t)= d+10 u(tsgn(t)+1/2 Let: a=0 gF(t)]-F(ol∠(a) so:F(a1/j心+x8() IF(oI /2 T/2 33-3 Basic property of Fourier Transform 33-3 Basic property of Fourier Transform niqueness:f(t)+>F(o one-to-one F(ω)=f(t)edt The uniqueness of F T tells the internal relation between the he f characteristic of a signal. That is to sav. once the time-domain Cf(t).cosctdt-jLf(t).sinotdt characteristic of a signal is fixed, its frequency-domain characteristie is also fixed F.T. is a method which can tell us the frequency-domain If f(t-f(-t), which means f(t)is an even function on t. characteristic of the signal using mathematics, Then f(o)is a real even function on o. LInearity(Superposition) f(t)+F(o); f,(t)+F(o If f(t-f(-th, which means f(t) is a odd function on L. Then F(o) is a imaginary odd function on a →af1(t)+βf2(t)分aF1()+β(0 33-3 Basic property of Fourier Transform- Symmetry 33-3 Basic property of Fourier Transform 1: Find out the F.t. of 1/27 the F.t. of the unit impulse signal 5(t) f(t)=)F(o) do F(o)=8(t). it dt=1 let it be: F((t)]=1 If f(t)f(-th, which means f(t) is an even function on L. That is:8(t)=i0"1 1.ejt do f(t)=f(t)=F(o).e do 8(u) we can get: So the F.t. of 1/nt is o(a (1nx)- the waveforms of F(o)and f(t) can be replace verification After twice F.T., the real even signal f(t)return F厂edt=(-)-(a)
北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 ω |F(ω)| ω ϕ(ω) π/2 -π/2 Unit step signal u(t) method 1: Unit step signal u(t) method 1: [ ] α jω 1 e u(t) -αt + First:F = Let: a = 0 [ ] = =|F(ω|∠ jω 1 F u(t) ϕ(ω) §3-2: Fourier Transform — frequency characteristics of three basic signals* 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 ωwwhu 北京大学 |F(ω)| ω ϕ(ω) π/2 -π/2 Unit step signal u(t) method 2: using the F.T. of known signals: u(t)=sgn(t)+1/2 so: F(ω)=1/jω+πδ(ω) ? Unit step signal u(t) method 2: using the F.T. of known signals: u(t)=sgn(t)+1/2 so: F(ω)=1/jω+πδ(ω) ? §3-2: Fourier Transform — frequency characteristics of three basic signals* 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 §3-3 Basic property of Fourier Transform uniqueness uniqueness:: one-to-one one-to-one Linearity (Superposition) Linearity (Superposition) f (t) F (ω); 1 ↔ 1 f2(t)↔F2(ω) αf1(t)+βf2(t)↔ αF1(ω)+βF2(ω) The uniqueness of F.T. tells the internal relation between the time-domain characteristic and the frequency-domain characteristic of a signal. That is to say, once the time-domain characteristic of a signal is fixed, its frequency-domain characteristic is also fixed. F.T. is a method which can tell us the frequency-domain characteristic of the signal using mathematics. f (t) F ( ↔ ω) 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Symmetry Symmetry If f(t)=f(-t), which means f(t) is an even function on t. Then F(ω) is a real even function on ω. ∫ ∫ ∫ +∞ ∞ +∞ ∞ +∞ ∞ = ⋅ − ⋅ ⋅ = ⋅ - - - -jωt f(t) cosωtdt j f(t) sinωtdt F(ω) f(t) e dt §3-3 Basic property of Fourier Transform *** If f(t)=-f(-t), which means f(t) is a odd function on t. Then F(ω) is a imaginary odd function on ω. 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Symmetry Symmetry ∫ +∞ ∞ = ⋅ - -jωt F(ω) f(t) e dt The two equations is same in form except for the factor (1/2π) Æ the waveforms of F(ω) and f(t) can be replaced. ∫ +∞ ∞ = ⋅ - jωt F(ω) e dω 2 1 f(t) π ∫ +∞ ∞ = = ⋅ - -jωt F(ω) e dω 2 1 f(t) f(-t) π After twice F.T., the real even signal f(t) return to its original waveform. *** comparison That is to say: §3-3 Basic property of Fourier Transform -- Symmetry If f(t)=f(-t), which means f(t) is an even function on t. 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 e.g.1: Find out the F.T. of 1/2π using the F.T. of the unit impulse signal . δ(t) F ∫ +∞ ∞ = ⋅ - jωt 1 e dω 2 1 δ(t) π That is: let it be: [ ] δ(t) = 1 +∞ ∞ = ⋅ = ∫ -jωt - F(ω) δ(t) e dt 1 So the F.T. of 1/2π is If we change the variables, we can get: ∫ +∞ ∞ = ⋅ - jωt 1 e dt 2 1 δ(ω) π δ(ω) e dt δ( ω) δ(ω) 2 1 2 1 - -jωt = = − = ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ∫ +∞ F π π ∞ verification : §3-3 Basic property of Fourier Transform
33-3 Basic property of Fourier Transform 2 Tea break/ F(o Homework 3-891921 f(t=arSa(tr/2)/2T F(o 53-3 Basic property of Fourier Transform 3-3 Basic property of Fourier Bandwidth x pulse uTime scaling e. g: let a>1 domain is equivalent to stretching FCo= AtSa( 0T/2) if f(t)<>F(o in frequeney domain. hen f(at)*,-F(o/a) F(o)=AT/a- Sa( 00T/2a) FIf(at)]-[f(at).e-istdt fa)r2a丌-2xAT =ALf(at). e dat a -F(o/a) Compression of signals in time domain is equivalent to stretching in frequency domain. 33-3 Basic property of Fourier Transform 33-3 Basic property of Fourier Transform Time shifting FIf(t)]=F f(t)+F(ω) Then: FI(t)e-]-F(-。 Then;f(t-t0)←F(o)e- Proof: FIf()e u]=[f(t).elu'e" dt =f(t)e-(a"o dt=F(o-w) The meaning of time shifting: When a signal get through a linear system which shifts the f(t)e-dF(ω-u)f(t) ea eF(+。) alter. But a phase(sift is introduced in the phase Then: f(t)cost+=[F(o+wo)+F(o-0o)] Sum frequency and Difference frequency Both are called: beat frequenc
北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 f(t) t F(ω) ω τ a F(ω) ω τ a aτ F(ω) = aτSa(ωτ/2) f(t) t π τ 2 a f(t)= aτSa(tτ/2)/2π e.g.2: §3-3 Basic property of Fourier Transform 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Tea break! Tea break! Homework: Æ3-8,9,19,21 Homework: Æ3-8,9,19,21 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Time scaling Time scaling if f(t)↔F(ω) F(ω/a) a 1 then f(at)↔ F[ ] f(at) = ∫ +∞ ∞ ⋅ - -jωt f(at) e dt F(ω/a) a 1= ∫ +∞ ∞ = ⋅ - -jωt f(at) e dat a 1 Proof: *** Compression of signals in time domain is equivalent to stretching in frequency domain. Compression of signals in time domain is equivalent to stretching in frequency domain. §3-3 Basic property of Fourier Transform 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 §3-3 Basic property of Fourier Transform e.g.: let a>1 f(t) t F( ω) ω τ A A τ F( ω) = A τSa( ω τ/2) τ/a f(at) t A Compression of signals in time domain is equivalent to stretching in frequency domain. Compression of signals in time domain is equivalent to stretching in frequency domain. F( ω) ω F( ω) = A τ/a ⋅ Sa( ω τ/2a) A τ/a π τ π τ 2 2 ⋅ = π τ τ π 2 2a a⋅ = Bandwidth × pulse width = constant Bandwidth × pulse width = constant *** 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 If f(t)↔F(ω) -jωt0 f(t - t0)↔F(ω)e The meaning of time shifting: When a signal get through a linear system which shifts the signal with time τ, the magnitude of its F.T. does not alter. But a phase shift is introduced in the phase spectrum . ωτ Then: Time shifting: §3-3 Basic property of Fourier Transform *** 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 Sum frequency wwhuandDifference frequency 北京大学 Both are called: beat frequency If: F [f(t)] = F(ω) Frequency shifting: [f(t)e ] F(ω-ω0) jω0t Then:F = Proof: [ ] ∫ +∞ ∞ = ⋅ - jω t jω t - jωt f(t)e f(t) e e dt 0 0 F f(t) e dt F( ω - ω 0 ) - - j( ω -ω 0 )t = ⋅ = ∫ +∞ ∞ Then: [F(ω ω ) F(ω ω )] 2 1 f(t)cosω t0 ↔ + 0 + − 0 f(t)e F(ω ω0) jω0t ↔ − f(t)e F(ω ω0) -jω0t ↔ + §3-3 Basic property of Fourier Transform ***
33-3 Basic property of Fourier Transform 33-3 Basic property of Fourier Transform Application of Frequency shifting-amplitude modulation (AM) Amplitude modulation (AM): s f(t) f(t) F(ω+0)+F(-ω0 Fo COS f(t)cost f(t)coso t mt△NA f(t)cos he signal for modulation is AM signals Mixer(non-linear) the envelope of the AM signaL SSB=LSB or USB estigial side band modulation(VSB) 33-3 Basic property of Fourier Transform A simple frequency division multiplexing (FDM Modulation and demodulation of signals Acos(uot+φ) f(t).cos(oot+p) coso. t coso. t The modulation which changes the amplitude of signal BPF LPF t cosopt oscillation cosopt f(t t hannel LPF f(t) f(t)coswot f(t)cosopt f(t) f (t) A wDM optical fiber communication system -ROF system chapter3 Frequency Spectrum of Signals Fourier Transform Analysis of networks) 53-1 Frequency Spectrum of periodic signal: Fourier s 3-2 Frequency Spectrum of non-periodic signal Fourier Transform 33-3 Basic theorem of frequency analysis 83-4 Outputs of constant-parameter linear circuits The law of causation in cireuits
北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 [ ] F(ω ω ) F(ω-ω ) 2 1 f(t)cosω t0 ↔ + 0 + 0 Application of Frequency shifting -- amplitude modulation (AM): f(t) t F(ω) ω cosω t0 f(t) f(t)cosω t0 ω ω0 ω0 - t f(t)cosω t0 carrier carrier Signal for modulation Signal for modulation AM signals AM signals Mixer (non-linear) Mixer (non-linear) §3-3 Basic property of Fourier Transform 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Amplitude modulation (AM): f(t) t F(ω) t ω ω ω0 ω0 - f(t)cos ω t0 The signal for modulation is the envelope of the AM signal. LSB USB DSB=LSB+USB SSB=LSB or USB vestigial sideband modulation (VSB) §3-3 Basic property of Fourier Transform 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 = A⋅cos(ω0t +φ) sinusoidal signal f(t)cosω t0 channel channel cosω t0 LPF LPF cosω t0 f(t) f(t)cos ω t f(t) 0 2 The modulation which changes the amplitude of signals is called amplitude modulation, which is AM for short. The modulation which changes the amplitude of signals is called amplitude modulation, which is AM for short. f(t)⋅cos(ω0t +φ) Modulation and demodulation of signals: ω -ω0 ω0 ω -2ω0 2ω0 ω ω Introducing Nyquist's theorem local oscillation modulation demodulation §3-3 Basic property of Fourier Transform 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 A simple frequency division multiplexing communication system (FDM) cosω t2 f2(t) cosω t1 f1(t) cosω tn fn(t) ∑ channel channel cosω t2 cosω t1 cosω tn f2(t) f1(t) fn(t) LPF LPF LPF LPF LPF LPF BPF BPF BPF BPF BPF BPF 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 A WDM optical fiber communication system -ROF system Control Station UpLink DownLink Back bone Network Optical mm-wave WDM Sources M U X D E M U X λu1 λu2 ...λuN BS1 : EDFA λu1 λu2 ...λuN User Terminal Data Down Data Up BS2 BSn Mm-wave Wireless Link Photo Detector : : : : : λd1 λd2 ...λdN 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 The non-distortion condition Linear distortion and nonlinear distortion The law of causation in circuits The non-distortion condition Linear distortion and nonlinear distortion The law of causation in circuits Qualitatively。。。 Qualitatively。。。 第三章:信号的频谱 §3-1 Frequency Spectrum of periodic signal: Fourier Series §3-2 Frequency Spectrum of non-periodic signal: Fourier Transform §3-3 Basic theorem of frequency analysis §3-4 Outputs of constant-parameter linear circuits chapter3 Frequency Spectrum of Signals (Fourier Transform Analysis of networks) chapter3 Frequency Spectrum of Signals (Fourier Transform Analysis of networks)
s3-4: The non-distortion condition 534: The waveform distortion of signals after cireuits: low-pass cireuits Time f(t) y(t) y(t)=A.f(t-D) In s-field: H(s)1 src f(t)CT y(t) transfer HG)-1 R Y(o Ae function H(o)|=|A What may the response be if the put f(t) is a square pulse which (u) φ(o 丌-rABW, tIH(o)I,IF(o)I RC=58 S3-4: The law of causation in circuits 33-4: The law of causation in cireuits pl* The law of causation Ideal Low-pass Filter: The output of any physically realizable circuits Lis after the inputin the time domain. Time f(t) y(t) y(t)=A.f(t-t) H(o output? We must have T>0
北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 §3-4:The non-distortion condition f(t) y(t) τ Time domain H(ω) F(ω) Y(ω) Frequency domain y(t)= A⋅ f(t - τ) Y(ω) A e F(ω) -jω = ⋅ ⋅ τ -jωτ H(ω) = A⋅e ω H(ω) ω φ(ω) |H(ω)|=|A| ⎩ ⎨ ⎧ − = A 0 A 0 φ(ω) π ωτ ωτ Non-distortion condition Linear distortion: the distortion caused by linear circuits Amplitude ≠ contant Phase ≠ linear line Nonlinear distortion: the distortion caused by nonlinear circuits beat frequency *** { 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 §3-4: The waveform distortion of signals after circuits: low-pass circuits f(t) y(t) + + - - R C 1 sRC 1 H(s) + In s-field: = transfer function 1 jωRC 1 H(jω) + = 2 1 0 H(ω) ωc = 1/RC ω 1 t f(t) ←a→ 1 F(ω) a 2π ω What may the response be if the input f(t) is a square pulse which width is a? BWH = ωc = π a BW 2 / f 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 |H(ω)|,|F(ω)| ω |H(ω)|,|F(ω)| ω |H(ω)|,|F(ω)| ω t y(t), f(t) 1 t y(t), f(t) 1 t y(t), f(t) 1 *** BWH =BWf BWH >BWf BWH 0 * 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 H(ω) ω Ideal Low-pass Filter: input t δ(t) *** ω F(ω) 1 output? §3-4:The law of causation in circuits
33-3 Basic property of Fourier Transform 3-4: The law of causation in cireuits 2 F(o=at Sa(or/2) Ideal Low-pass Filter input F() o(t) 几 f(t)=aTSa(tr/2)/2T F() So the ideal low-pass filter is not a physically realizable cireuit Fourier analysis, Laplace analysis chapter3 Frequency Spectrum of signior Fourier Transform Analysis of networks) 日 Fourier analysis Fourier Series: spectrum analysis of periodie signals s3-1 Frequency Spectrum of periodic signal: Fourier snab)T f(t) dt A(t)- 2c,e hmx s 3-2 Frequency Spectrum of non-periodic signal i- Fourier Transform: spectrum analysis of non-periodic sd n. Fourier series Fourier transform F(o)=Lf(t)-e dt H f(t).2.C nnm时 periodic signal req frequency analysis of am吗电四由 日 Laplace analysis: f(t=f(t+T f(t),T∞ 9 Laplace transform: Analysis of the network responses and dynamic elements) F(s)=Cf(tle"dt f(t)= F(s)e tds ontinuous spectrum Cf(t)e-dt
北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 f(t) t F(ω) ω τ a F(ω) ω τ a aτ F(ω) = aτSa(ωτ/2) f(t) t π τ 2 a f(t)= aτSa(tτ/2)/2π e.g.2: §3-3 Basic property of Fourier Transform Review 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 H(ω) ω Ideal Low-pass Filter: So the ideal low-pass filter is not a physically realizable circuit. output input t δ(t) h(t) t t<0- §3-4:The law of causation in circuits *** 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 第三章:信号的频谱 §3-1 Frequency Spectrum of periodic signal: Fourier Series §3-2 Frequency Spectrum of non-periodic signal: Fourier Transform §3-3 Basic theorem of frequency analysis §3-4 Outputs of constant-parameter linear circuits chapter3 Frequency Spectrum of Signals (Fourier Transform Analysis of networks) chapter3 Frequency Spectrum of Signals (Fourier Transform Analysis of networks) Fourier series Fourier transform frequency analysis of periodic signals f(t)=f(t+T) 0 0 nω ω discrete spectrum f(t), TÆ∞ ω dω continuous spectrum Summary frequency analysis of any signals 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 北京大学 wwhu 北京大学 wwhu 北京大学 wwhu 北京大学 Fourier analysis, Laplace analysis Fourier analysis: Æ Fourier Series: spectrum analysis of periodic signals Laplace analysis: Æ Laplace transform: Analysis of the network responses and characteristics of any stimulated signal (commonly used when there are dynamic elements) *** ( ) () ∫ ∞ −∞ − F s = f t e dt st ( ) ( ) ∫ + ∞ − ∞ = σ j σ j st f t F s e ds 2πj 1 ( ) ∫ ∞ − − = 0 f t e dt st ∫ +∞ ∞ = ⋅ - -jωt F(ω) f(t) e dt F( ω) e d ω 2 1 f(t) jωt = ⋅ ∫ +∞ π − ∞ ∑ ∞ =−∞ = n jnω t ne 0 f(t) c ∫ + = t T t -jnω t n 0 0 0 f(t)e 0 dt T 1 c (nω ) Æ Fourier Transform: spectrum analysis of non-periodic signals Summary