Discrete Time Signals and Systems

发布时间:2023年12月25日

Discrete Time Signals and Systems

文章目录

Signal classification

  • Periodic and non-periodic
  • Odd and even signals
  • energy signal and power signal

basic signal

  • Impulse function
  • unit step function
  • Ramp function

Operation on signal

?

Periodic and Aperiodic Discrete-Time Sinusoids
x ( n ) = A c o s [ 2 π f 0 n ] = x ( n + N ) = A c o s [ 2 π f 0 ( n + N ) ] = A c o s [ 2 π f 0 n + 2 π f 0 N ] x(n)=Acos[2 \pi f_0 n]=x(n+N)=Acos[2 \pi f_0 (n+N)]=Acos[2 \pi f_0 n+2 \pi f_0 N] x(n)=Acos[2πf0?n]=x(n+N)=Acos[2πf0?(n+N)]=Acos[2πf0?n+2πf0?N]
2 π f 0 N = 2 π k 2\pi f_0 N =2 \pi k 2πf0?N=2πk just f 0 = k N f_0 = \frac{k}{N} f0?=Nk?

n is integer: Periodic

n is not integer: Aperiodic

Periodic judgment of composite signals

  1. Find N for each signal

    if N 1 N 2 = r a t i o n a l ? n u m b e r \frac{N_1}{N_2} = rational \ number N2?N1??=rational?number it is Periodic

  2. Find the lowest common multiple($ LCM(N_1,N_2)$)

    Periodic is L C M ( N 1 , N 2 ) LCM(N_1,N_2) LCM(N1?,N2?)

Odd and even signals

odd: x 0 ( t ) = 1 2 [ x ( t ) ? x ( ? t ) ] x_0(t)=\frac{1}{2}[x(t)-x(-t)] x0?(t)=21?[x(t)?x(?t)]

even: x e ( t ) = 1 2 [ x ( t ) + x ( ? t ) ] x_e(t) = \frac{1}{2}[x(t)+x(-t)] xe?(t)=21?[x(t)+x(?t)]

x ( t ) = x 0 ( t ) + x e ( t ) x(t) = x_0(t) + x_e(t) x(t)=x0?(t)+xe?(t)

energy signal and power signal

energy: E = ∑ n = ? ∞ ∞ ∣ x [ n ] ∣ 2 E=\sum_{n=-\infty}^{\infty}|x[n]|^2 E=n=??x[n]2

power:

Periodic: P ∞ = lim ? N → ∞ 1 2 N + 1 ∑ n = ? ∞ + ∞ ∣ x [ n ] ∣ 2 P_\infty=\lim_{N\to\infty}\frac1{2N+1}\sum_{n=-\infty}^{+\infty}|x[n]|^2 P?=limN?2N+11?n=?+?x[n]2

Aperiodic: P x = 1 N ∑ n = 0 N ? 1 ∣ x [ n ] ∣ 2 P_x=\frac1{N}\sum_{n=0}^{N-1}|x[n]|^2 Px?=N1?n=0N?1?x[n]2

energy signal:energy is finite,power is zero

power signal:energy is infinite,power is finite

?

find the energy and power for

  • Impulse function
  • unit step function
  • Ramp function

?

  1. Time Shifting(left is +;right is -)
  2. Time-scale
  3. Time Reversal

?

System of discrete signal

  • Linear systems and nonlinear systems

  • Causal and Acausal Systems

  • Time-varying and time-invariant systems

  • static system and dynamic system

  • Stable and unstable systems

  • convolution

  • convolution sum

  • circular convolution

  • Stability of linear time-invariant systems

Linear systems and nonlinear systems

  • Linear systems satisfy uniformity and superposition
  • A system that satisfies uniformity and superposition is a linear system

image-20231222222522112

Must have: x ( n ) = y ( n ) = 0 x(n) = y(n) = 0 x(n)=y(n)=0

Four steps to solve problems:

  1. x1 to y1 = F(x1)
  2. x2 to y2 = F(x2)
  3. y3 = ay1 + by2
  4. F(ax1 +bx2) = y4

if y4 = y3 ,the system is linear

?

Causal and non-causal Systems

casual system: The output depends only on present and past signals

Acausal Systems:The output depends on at least one future input

eg:Y(n) = x(-n) is non-casual (at n = -1)

even and odd is non-casual

ps: anti-casual system: output only upon “only future input” for all time

?

Time-varying and time-invariant systems

Time-varying system(TVS):

y(n) = F[x(n)] = x(n)cos(2 w π w \pi wπn)

y(n,k) = F[x(n-k)] = x(n-k)cos(2 w π w \pi wπn)

y(n-k) ≠ \ne = y(n,k)

y(n) to y(n,k):only change n for x(n)

time-invariant systems :

y(n,k) = y(n-k)

eg:y(n-k) = sin (x(n-k))

y change n for all n

?

static system and dynamic system

static system (memory-less system):

output only depends on now input for all time

dynamic system:

output depends on past and/or future inputs

?

Stable and unstable systems

Stable system: BIBO

bounded input to bounded output

unstable systems:

bounded input to unbounded output

eg: y ( n ) = y 2 ( n ? 1 ) + 2 δ ( n ) y(n)=y^2(n-1)+2\delta (n) y(n)=y2(n?1)+2δ(n)

?

convolution

  • time-invariant systems

y ( n ) = ∑ k = ? ∞ ∞ x ( k ) h ( n ? k ) y(n)=\sum_{k=-\infty}^{\infty}x(k)h(n-k) y(n)=k=??x(k)h(n?k)

y ( n ) = x ( n ) ? h ( n ) y(n)=x(n)*h(n) y(n)=x(n)?h(n)

  1. time reversal
  2. shifting of h(-k) to h(n-k)
  3. multiply x(k)h(n-k)
  4. Sum
  • linear convolution
  • circular convolution

matrix method:

x(n) = {1,2,3,4} h(n)={0,2,2,2}

length of x(n):4 samples;k

length of h(n):4 samples;m

length of y(n): = k + m - 1 = 4 + 4 - 1 = 7 samples

image-20231223180642723

?

circular convolution

x(n) = {2,1,2} h(n) = {1,2,3}

y ( n ) = ∑ k = ? ∞ ∞ x ( k ) h ( n ? k ) y(n)=\sum_{k=-\infty}^{\infty}x(k)h(n-k) y(n)=k=??x(k)h(n?k)

image-20231223181124148

?

Stability of linear time-invariant systems

S = ∑ k = ? ∞ ∞ ∣ h ( n ) ∣ < ∞ S=\sum_{k=-\infty}^{\infty}|h(n)| < \infty S=k=??h(n)<

it is stability

eg: h ( n ) = ( 0.8 ) n u ( n + 2 ) h(n) = (0.8)^n u(n+2) h(n)=(0.8)nu(n+2)

?

Fourier Series

  • Fourier Series basic concepts
  • Discrete time fouriew series (DTFS)

Fourier Series basic concepts

Condition for existence of Fourier Series

  1. Finite Number of maxima and minima over the time period
  2. Finite number of discontinuities over the time period T
  3. Signal should be absolutely integrable over the range of time period T

period function:

f ( x ) = a 0 + ∑ n = 1 ∞ ( a n c o s ( n w 0 t ) + b n s i n ( n w 0 t ) f(x)=a_0+\sum_{n=1}^\infty(a_ncos(nw_0t)+b_n sin(nw_0t) f(x)=a0?+n=1?(an?cos(nw0?t)+bn?sin(nw0?t)

a 0 = 1 T 1 ∫ 0 T 1 f ( t ) d t a_{0}=\frac{1}{T_{1}}\int_{0}^{T_1}f(t)dt a0?=T1?1?0T1??f(t)dt

a n = 2 T 1 ∫ 0 T 1 f ( t ) cos ? n w 1 t d t a_{n}=\frac{2}{T_{1}}\int_{0}^{T_{1}}f(t)\cos nw_1tdt an?=T1?2?0T1??f(t)cosnw1?tdt

b n = 2 T 1 ∫ 0 T 1 f ( t ) sin ? n w 1 t d t b_{n}=\frac{2}{T_{1}}\int_{0}^{T_{1}}f(t)\sin nw_1tdt bn?=T1?2?0T1??f(t)sinnw1?tdt

eg:image-20231223183235229

?

Discrete time fouriew series (DTFS)

x n = ∑ k = 0 N ? 1 c ( k ) e J k 2 π N n x_n=\sum_{k=0}^{N-1}c(k)e^{Jk\frac{2\pi }{N}n} xn?=k=0N?1?c(k)eJkN2π?n

c ( k ) = 1 N ∑ n = 0 N ? 1 x ( n ) e ? J k 2 π N n c(k)=\frac{1}{N}\sum_{n=0}^{N-1}x(n)e^{-Jk\frac{2\pi}{N}n} c(k)=N1?n=0N?1?x(n)e?JkN2π?n

?

CTFS

x ( w ) = ∫ ? ∞ + ∞ x ( t ) e ? J w t d t x(w)=\int_{-\infty}^{+\infty}x(t)e^{-Jwt}dt x(w)=?+?x(t)e?Jwtdt

Inverse transformation: x ( t ) = 1 2 π ∫ ? ∞ + ∞ x ( w ) e j w t d w x(t)=\frac{1}{2\pi}\int_{-\infty}^{+\infty}x(w)e^{jwt}dw x(t)=2π1??+?x(w)ejwtdw

?

Continuous time Fourier transform (CTFT)

  • basic concept
  • fourier fransform
  • magnitude specturm
  • phase spectrum

properties of CTFT

  • linearity
  • time shifting
  • frequency shifting
  • time scaling
  • time differenciation
  • frequency differenciation
  • convolution

Parseval’s Theorem

  1. Any signal is built up addition of elementary signals which are at different frequencies.

  2. CTFT is used to transform the signal from time domain to frequency domain.

  3. With the help of CTFT,plot the amplitude and phase spectrum.

  4. CTFT can be efxpressed as:

    x ( w ) = ∫ ? ∞ + ∞ x ( t ) e ? J w t d t x(w)=\int_{-\infty}^{+\infty}x(t)e^{-Jwt}dt x(w)=?+?x(t)e?Jwtdt

    Inverse transformation: x ( t ) = 1 2 π ∫ ? ∞ + ∞ x ( w ) e j w t d w x(t)=\frac{1}{2\pi}\int_{-\infty}^{+\infty}x(w)e^{jwt}dw x(t)=2π1??+?x(w)ejwtdw

?

Dirichlet’s conditions

  1. f(t) should be absolutely integrable.
  2. f(t) should have a finite number of maxima and minima over any finite interval.
  3. f(t) should have finite number of discontinues over any finite interval

These conditions are sufficient but not necessary Condition.

If the conditions are not met, there is not necessarily no Fourier transform.

?

lecture(77)

  • fourier fransform
  • magnitude specturm
  • phase spectrum

?

lecture(78)

Inverse transformation for function

important

?

properties of CTFT

linearity

F [ f ( t ) ] = F ( ω ) , F [ g ( t ) ] = G ( ω ) \mathscr{F}[f(t)]=F(\omega),\quad\mathscr{F}[g(t)]=G(\omega) F[f(t)]=F(ω),F[g(t)]=G(ω)

F [ α f ( t ) + β g ( t ) ] = α F ( ω ) + β G ( ω ) F ? 1 [ α F ( ω ) + β G ( ω ) ] = α f ( t ) + β g ( t ) \begin{aligned}\mathscr{F}[\alpha f(t)+\beta g(t)]&=\alpha F(\omega)+\beta G(\omega)\\\mathscr{F}^{-1}[\alpha F(\omega)+\beta G(\omega)]&=\alpha f(t)+\beta g(t)\end{aligned} F[αf(t)+βg(t)]F?1[αF(ω)+βG(ω)]?=αF(ω)+βG(ω)=αf(t)+βg(t)?

time shifting

F [ f ( t ) ] = F ( ω ) \mathcal{F}[f(t)]=F(\omega) F[f(t)]=F(ω)

F [ f ( t ? t 0 ) ] = e ? j ω t 0 F ( ω ) F ? 1 [ F ( ω ? ω 0 ) ] = e j ω 0 t f ( t ) \begin{aligned}\mathscr{F}[f(t-t_0)]&=e^{-j\omega t_0}F(\omega)\\\mathscr{F}^{-1}[F(\omega-\omega_0)]&=e^{j\omega_0t}f(t)\end{aligned} F[f(t?t0?)]F?1[F(ω?ω0?)]?=e?t0?F(ω)=ejω0?tf(t)?

time scaling

F [ f ( t ) ] = F ( ω ) \mathcal{F}[f(t)]=F(\omega) F[f(t)]=F(ω)

F [ f ( a t ) ] = 1 ∣ a ∣ F ( ω a ) \mathscr{F}[f(at)]=\frac{1}{|a|}F\left(\frac{\omega}{a}\right) F[f(at)]=a1?F(aω?)

frequency shifting

F [ f ( t ) ] = F ( ω ) \mathcal{F}[f(t)]=F(\omega) F[f(t)]=F(ω)

F [ f ( t ) e j w 0 t ] = F ( ω ? ω 0 ) F ? 1 [ f ( t ) e ? j w 0 t ] = F ( ω + ω 0 ) \begin{aligned}\mathscr{F}[f(t)e^{jw_0t}]&=F(\omega - \omega_0)\\\mathscr{F}^{-1}[f(t)e^{-jw_0t}]&=F(\omega + \omega_0)\end{aligned} F[f(t)ejw0?t]F?1[f(t)e?jw0?t]?=F(ω?ω0?)=F(ω+ω0?)?

time differenciation

F [ f ( t ) ] = F ( ω ) \mathcal{F}[f(t)]=F(\omega) F[f(t)]=F(ω)

F [ d n f ( t ) d t n ] = ( j ω ) n F ( ω ) F ? 1 [ d n F ( ω ) d ω n ] = ( ? j t ) n f ( t ) \begin{gathered} \mathscr{F}\left[\frac{d^{n}f(t)}{dt^{n}}\right]=(j\omega)^{n}F(\omega) \\ \mathscr{F}^{-1}\left[\frac{d^{n}F(\omega)}{d\omega^{n}}\right]=(-jt)^{n}f(t) \end{gathered} F[dtndnf(t)?]=()nF(ω)F?1[dωndnF(ω)?]=(?jt)nf(t)?

frequency differenciation

F [ f ( t ) ] = F ( ω ) \mathcal{F}[f(t)]=F(\omega) F[f(t)]=F(ω)

F [ ( ? j t ) n f ( t ) ] = d n F ( w ) d w n \mathscr{F} [(-jt)^{n}f(t)]= \frac{d^{n}F(w)}{dw^{n}} F[(?jt)nf(t)]=dwndnF(w)?

F [ t f ( t ) ] = j d F ( w ) d w \mathscr{F} [tf(t)]= j\frac{dF(w)}{dw} F[tf(t)]=jdwdF(w)?

convolution

F [ f ( t ) ] = F ( ω ) , F [ g ( t ) ] = G ( ω ) \mathscr{F}[f(t)]=F(\omega),\quad\mathscr{F}[g(t)]=G(\omega) F[f(t)]=F(ω),F[g(t)]=G(ω)

F [ f 1 ( t ) ? f 2 ( t ) ] = F 1 ( ω ) ? F 2 ( ω ) \mathscr{F}[f_1(t) *f_2(t)]=F_1(\omega)·F_2(\omega) F[f1?(t)?f2?(t)]=F1?(ω)?F2?(ω)

F [ f 1 ( t ) ? f 2 ( t ) ] = 1 2 π F 1 ( ω ) ? F 2 ( ω ) \mathscr{F}[f_1(t) ·f_2(t)]=\frac{1}{2\pi}F_1(\omega)*F_2(\omega) F[f1?(t)?f2?(t)]=2π1?F1?(ω)?F2?(ω)

?

Parseval’s Theorem

if x(t) CTFT to x ( w ) x(w) x(w) or x ( f ) x(f) x(f)

E = ∫ ? ∞ ∞ ∣ x ( t ) ∣ 2 d t = ∫ ? ∞ ∞ ∣ x ( f ) ∣ 2 d f E=\int_{-\infty}^{\infty}|x(t)|^{2}dt=\int_{-\infty}^{\infty}|x(f)|^{2}df E=??x(t)2dt=??x(f)2df

?

Fourier transform of Periodic signal

  • spectrum of e j ω 0 t e^{j\omega_0t } ejω0?t and e ? j ω 0 t e^{- j\omega_0t } e?jω0?t
  • spectrum of s i n w t sinwt sinwt and c o s w t coswt coswt
  • Spectrum of a general periodic signal

spectrum of e j ω 0 t e^{j\omega_0t } ejω0?t and e ? j ω 0 t e^{- j\omega_0t } e?jω0?t

F [ e j ω 0 t ] = 2 π δ ( ω ? ω 0 ) F [ e ? j ω 0 t ] = 2 π δ ( ω + ω 0 ) \begin{aligned}\mathscr{F}[e^{j\omega_0t}]&=2\pi\delta(\omega-\omega_0)\\\mathscr{F}[e^{-j\omega_0t}]&=2\pi\delta(\omega+\omega_0)\end{aligned} F[ejω0?t]F[e?jω0?t]?=2πδ(ω?ω0?)=2πδ(ω+ω0?)?

?

spectrum of s i n w 0 t sinw_0t sinw0?t and c o s w 0 t cosw_0t cosw0?t

F [ cos ? ( ω 0 t ) ] = F [ e j ω 0 t + e ? j ω 0 t 2 ] = π [ δ ( ω ? ω 0 ) + δ ( ω + ω 0 ) ] F [ sin ? ( ω 0 t ) ] = F [ e j ω 0 t ? e ? j ω 0 t 2 j ] = j π [ δ ( ω + ω 0 ) ? δ ( ω ? ω 0 ) ] \begin{gathered} \mathscr{F}[\cos(\omega_{0}t)]=\mathscr{F}[\frac{e^{j\omega_{0}t}+e^{-j\omega_{0}t}}{2}]=\pi[\delta(\omega-\omega_{0})+\delta(\omega+\omega_{0})] \\ \mathscr{F}[\sin(\omega_{0}t)]=\mathscr{F}[\frac{e^{j\omega_{0}t}-e^{-j\omega_{0}t}}{2j}]=j\pi[\delta(\omega+\omega_{0})-\delta(\omega-\omega_{0})] \end{gathered} F[cos(ω0?t)]=F[2ejω0?t+e?jω0?t?]=π[δ(ω?ω0?)+δ(ω+ω0?)]F[sin(ω0?t)]=F[2jejω0?t?e?jω0?t?]=[δ(ω+ω0?)?δ(ω?ω0?)]?

?

Spectrum of a general periodic signal

f ( t ) = ∑ n = ? ∞ ∞ F ( n w 0 ) e j n w 0 t f(t)=\sum_{n=-\infty}^{\infty}F(nw_0)e^{jnw_0 t} f(t)=n=??F(nw0?)ejnw0?t

F ( w ) = F [ f ( t ) ] = 2 π ∑ n = ? ∞ ∞ F ( n w 0 ) δ ( ω ? n ω 0 ) F(w) = \mathscr{F}[f(t)] = 2\pi \sum_{n=-\infty}^{\infty}F(nw_0)\delta(\omega-n\omega_{0}) F(w)=F[f(t)]=2πn=??F(nw0?)δ(ω?nω0?)

F ( n w 1 ) = 1 T ∫ 0 T f ( t ) e ? j n w 0 t d t F(nw_1) = \frac{1}{T}\int_{0}^{T} f(t)e^{-jnw_0t}dt F(nw1?)=T1?0T?f(t)e?jnw0?tdt

?

Discrete time fouriew series (DTFT)

  • basic properties
  • DTFT for common signals

basic properties

DTFT: X ( e j ω ) = ∑ n = ? ∞ + ∞ x [ n ] e ? j ω n X\left(e^{j\omega}\right)=\sum_{n=-\infty}^{+\infty}x[n]e^{-j\omega n} X(e)=n=?+?x[n]e?jωn

Inverse transform of DTFT: x [ n ] = 1 2 π ∫ 2 π X ( e j ω ) e j ω n d ω x\left[n\right]=\frac1{2\pi}\int_{2\pi}X(e^{j\omega})e^{j\omega n}d\omega x[n]=2π1?2π?X(e)ejωndω

DTFT for common signals

x ( n ) = a n u ( n ) ; ∣ a ∣ < 1 x(n) = a^n u(n);|a|<1 x(n)=anu(n);a<1

DTFT?[X(n)] = ∑ n = ? ∞ ∞ x ( n ) e ? j w n = ∑ n = 0 + ∞ a n e ? j w n = ∑ n = 0 + ∞ ( a ? e ? j w ) n = 1 1 ? a e ? j w = 1 1 ? a ( c o s w ? j s i n w ) = 1 ( 1 ? a c o s w ) + j a s i n w \begin{aligned}\text{DTFT [X(n)]}&=\sum_{n=-\infty}^{\infty}x({n})e^{-jwn}\\&=\sum_{n=0}^{+\infty}a^ne^{-jwn}\\&=\sum_{n=0}^{+\infty}(a\cdot e^{-jw})^n\\&=\frac1{1-ae^{-jw}} \\&=\frac{1}{1-a(cosw-jsinw)} \\&=\frac{1}{(1-acosw)+jasinw}\end{aligned} DTFT?[X(n)]?=n=??x(n)e?jwn=n=0+?ane?jwn=n=0+?(a?e?jw)n=1?ae?jw1?=1?a(cosw?jsinw)1?=(1?acosw)+jasinw1??

magnitude:

∣ x ( e j w ) ∣ = 1 ( 1 ? a c o s w ) 2 + ( a s i n w ) 2 = 1 1 + a 2 ? 2 a c o s w |x(e^{jw})| = \frac{1}{\sqrt{(1-acosw)^2}+(asinw)^2 }=\frac{1}{\sqrt{1+a^2-2acosw} } x(ejw)=(1?acosw)2 ?+(asinw)21?=1+a2?2acosw ?1?

phase:

∠ x ( e j w ) = ? a r c t a n a s i n w 1 ? a c o s w \angle x(e^{jw}) = -arctan\frac{asinw}{1-acosw} x(ejw)=?arctan1?acoswasinw?

eg:x(n) = {-4,3,-2,3,-4}

find:(1). x ( e j 0 ) x(e^{j0}) x(ej0) (2). x ( e j w ) x(e^{jw}) x(ejw) (3). ∫ π ? π x ( e j w ) d w \int_{\pi }^{-\pi} x(e^ {jw} )dw π?π?x(ejw)dw

?

properties

linear

a x 1 [ n ] + b x 2 [ n ] ? a X 1 ( e j ω ) + b X 2 ( e j ω ) ax_1\left[n\right]+bx_2\left[n\right]\leftrightarrow aX_1(e^{j\omega})+bX_2(e^{j\omega}) ax1?[n]+bx2?[n]?aX1?(e)+bX2?(e)

time shifting

x ( n ) ? x ( e j w ) x(n) \leftrightarrow x(e^{jw}) x(n)?x(ejw)

x ( n ? n 0 ) ? x ( e j w ) ? e ? j w n 0 x( n - n_0) \leftrightarrow x(e^{jw})·e^{-jwn_0} x(n?n0?)?x(ejw)?e?jwn0?

frequency shifting

x ( n ) ? x ( e j w ) x(n) \leftrightarrow x(e^{jw}) x(n)?x(ejw)

x ( n ) ? e ? j w n 0 ? x ( e j ( w ? w 0 ) ) x(n)·e^{-jwn_0} \leftrightarrow x(e^{j(w-w_0)}) x(n)?e?jwn0??x(ej(w?w0?))

Frequency domain differential

x ( n ) ? x ( e j w ) x(n) \leftrightarrow x(e^{jw}) x(n)?x(ejw)

n x ( n ) ? j d x ( e j w ) d w n x(n) \leftrightarrow j \frac{dx(e^{jw})}{dw} nx(n)?jdwdx(ejw)?

?

eg:

1. y ( n ) ? A y ( n ? 1 ) = x ( n ) ; ∣ A ∣ < 1 y(n) - Ay(n-1) = x(n);|A|<1 y(n)?Ay(n?1)=x(n);A<1

2. y ( n ) ? 3 4 y ( n ? 1 ) + 1 8 y ( n ? 2 ) = 2 x ( n ) y(n) - \frac{3}{4}y(n-1)+\frac{1}{8}y(n-2)=2x(n) y(n)?43?y(n?1)+81?y(n?2)=2x(n)

3. x ( e j w ) = e ? j w x(e^{jw})=e^{-jw} x(ejw)=e?jw for ? π ≤ w ≤ π -\pi \le w \le \pi ?πwπ

4.find convolution x ( n ) = 1 2 n u ( n ) x(n) = \frac{1}{2}^n u(n) x(n)=21?nu(n) and x ( n ) = 1 3 n u ( n ) x(n) = \frac{1}{3}^n u(n) x(n)=31?nu(n)

?

Discrete Fourier Transform(DFT)

DFT is equally spaced sampling of DTFT

DFT: x ( n ) —— > x ( k ) x(n) ——> x(k) x(n)——>x(k)

x ( k ) = ∑ n = 0 N ? 1 x ( n ) e ? j 2 π k N n ; k = 0 , 1 , 2 , 3 , . . . , N ? 1 x(k) = \sum_{n=0}^{N-1}x(n)e^{-j\frac{2\pi k}{N}n};k=0,1,2,3,...,N-1 x(k)=n=0N?1?x(n)e?jN2πk?n;k=0,1,2,3,...,N?1

making W N = e ? j 2 π N W_N = e^{-j\frac{2\pi}{N}} WN?=e?jN2π?

x ( k ) = ∑ n = 0 N ? 1 x ( n ) W N k n x(k) = \sum_{n=0}^{N-1}x(n)W_N^{kn} x(k)=n=0N?1?x(n)WNkn?

?

Calculation of DFT

eg: x ( k ) = ∑ n = 0 4 ? 1 x ( n ) e ? j 2 π k 4 n x(k) = \sum_{n=0}^{4-1}x(n)e^{-j\frac{2\pi k}{4}n} x(k)=n=04?1?x(n)e?j42πk?n

1. x ( k ) = ∑ n = 0 N ? 1 x ( n ) e ? j 2 π k N n ; k = 0 , 1 , 2 , 3 , . . . , N ? 1 x(k) = \sum_{n=0}^{N-1}x(n)e^{-j\frac{2\pi k}{N}n};k=0,1,2,3,...,N-1 x(k)=n=0N?1?x(n)e?jN2πk?n;k=0,1,2,3,...,N?1

2.matrix method

image-20231223221845915

image-20231224153828648
Rotation matrix of inverse DFT transform(IDFT)

image-20231224153845233

?

Properties of DFT

1.linear

X 1 ( k ) = D F T [ x 1 ( n ) ] X_1(k)=DFT[x_1(n)] X1?(k)=DFT[x1?(n)]

X 2 ( k ) = D F T [ x 2 ( n ) ] X_2(k)=DFT[x_2(n)] X2?(k)=DFT[x2?(n)]

D F T [ a x 1 ( n ) + b x 2 ( n ) ] = a X 1 ( k ) + b X 2 ( k ) DFT[ax_1(n)+bx_2(n)]=aX_1(k)+bX_2(k) DFT[ax1?(n)+bx2?(n)]=aX1?(k)+bX2?(k)

2.circular convolution

DFT: x [ n ] ? h [ n ] ? X [ k ] ? H [ k ] x[n]\circledast h[n]\Leftrightarrow X[k]·H[k] x[n]?h[n]?X[k]?H[k]

DFT: x [ n ] ? h [ n ] ? X [ k ] ? H [ k ] x[n]· h[n]\Leftrightarrow X[k]\circledast H[k] x[n]?h[n]?X[k]?H[k]

3.periodicity

D F T [ x ( n ) ] = x ( k ) DFT[x(n)] = x(k) DFT[x(n)]=x(k)

x ( n + N ) = x ( n ) x(n+N) = x(n) x(n+N)=x(n);for all n

x ( k + N ) = x ( k ) x(k+N) = x(k) x(k+N)=x(k);for all k

4.cyclic frequency shift

D F T [ x ( n ) e j 2 π N l n ] = x ( ( k ? l ) ) N DFT[x(n)e^{j\frac{2\pi}{N}ln}]=x((k-l))_N DFT[x(n)ejN2π?ln]=x((k?l))N?

(( )) is cyclic frequency

?

DTF calculates linear convolution and circular convolution

linear convilution

[ y ( n ) = x ( n ) ? h ( n ) ] [y(n) = x(n)*h(n)] [y(n)=x(n)?h(n)]

then the length of x(n) is L,and h(n) is M

the length of linear convolution: N = L + M -1

  1. filling the length of x(n) and h(n) to N (add 0)
  2. calculates DTF[x(n)] and DTF[h(n)]
  3. then$ y(k) = x(k)·h(k)$
  4. calculate I D F T [ Y ( k ) ] = y ( n ) IDFT[Y(k)] = y(n) IDFT[Y(k)]=y(n)

?

cyclical convolution

[ y ( n ) = x ( n ) ? h ( n ) ] [y(n) = x(n)\circledast h(n)] [y(n)=x(n)?h(n)]

then the length of x(n) is L,and h(n) is M(add 0)

the length of cyclical convolution is :N = max(L,M)

  1. filling the length of x(n) and h(n) to N
  2. calculates DTF[x(n)] and DTF[h(n)]
  3. then y ( k ) = x ( k ) ? h ( k ) y(k) = x(k)·h(k) y(k)=x(k)?h(k)
  4. calculate I D F T [ Y ( k ) ] = y ( n ) IDFT[Y(k)] = y(n) IDFT[Y(k)]=y(n)

?

another way to calculate linear convolution

  • Overlap-Add and Overlap-Save
Overlap-Add

then the length of x k ( n ) x_k(n) xk?(n) ( x k ( n ) x_k(n) xk?(n) is not x ( n ) x(n) x(n)) is L,and h(n) is M

the length of convolution is :N = L + M - 1 (the size of N is depend on you)

eg: x ( n ) = { 3 , ? 1 , 0 , 1 , 3 , 2 , 0 , 1 , 2 , 1 } x(n) = \{ 3,-1,0,1,3,2,0,1,2,1\} x(n)={3,?1,0,1,3,2,0,1,2,1} h ( n ) = { 1 , 1 , 1 } h(n) = \{ 1,1,1\} h(n)={1,1,1}

1.we take N = 6,then 6= L + 3 - 1;so L = 4.

x 1 ( n ) = { 0 , 0 , 3 , ? 1 , 0 , 1 } x_1(n) = \{ 0,0,3,-1,0,1 \} x1?(n)={0,0,3,?1,0,1}

x 2 ( n ) = { 0 , 1 , 3 , 2 , 0 , 1 } x_2(n) = \{ 0,1,3,2 ,0,1\} x2?(n)={0,1,3,2,0,1}

x 3 ( n ) = { 0 , 1 , 2 , 1 , 0 , 0 } x_3(n) = \{ 0,1,2,1,0,0 \} x3?(n)={0,1,2,1,0,0}

h ( n ) = { 1 , 1 , 1 , 0 , 0 , 0 } h(n) = \{1,1,1,0,0,0\} h(n)={1,1,1,0,0,0}

2.calculate

x 1 ( n ) ? h ( n ) = { 1 , 1 , 3 , 2 , 2 , 0 } x_1(n)\circledast h(n) =\{1,1,3,2,2,0\} x1?(n)?h(n)={1,1,3,2,2,0}

x 2 ( n ) ? h ( n ) = { 1 , 2 , 4 , 6 , 5 , 3 } x_2(n)\circledast h(n)=\{1,2,4,6,5,3\} x2?(n)?h(n)={1,2,4,6,5,3}

x 3 ( n ) ? h ( n ) = { 0 , 1 , 3 , 4 , 3 , 1 } x_3(n)\circledast h(n) =\{0,1,3,4,3,1\} x3?(n)?h(n)={0,1,3,4,3,1}

3.Remove first (M-1) points ,Concatenate all results

the result is { 3 , 2 , 2 , 0 , 4 , 6 , 5 , 3 , 3 , 4 , 3 , 1 } \{3,2,2,0,4,6,5,3,3,4,3,1\} {3,2,2,0,4,6,5,3,3,4,3,1}

?

Overlap-Save

then the length of x k ( n ) x_k(n) xk?(n) ( x k ( n ) x_k(n) xk?(n) is not x ( n ) x(n) x(n)) is L,and h(n) is M

the length of convolution is :N = L + M - 1 (the size of N is depend on you)

eg: x ( n ) = { 3 , ? 1 , 0 , 1 , 3 , 2 , 0 , 1 , 2 , 1 } x(n) = \{ 3,-1,0,1,3,2,0,1,2,1\} x(n)={3,?1,0,1,3,2,0,1,2,1} h ( n ) = { 1 , 1 , 1 } h(n) = \{ 1,1,1\} h(n)={1,1,1}

1.we take N = 5,then 5 = L + 3 - 1;so L = 3.

x 1 ( n ) = { 0 , 0 , 3 , ? 1 , 0 } x_1(n) = \{ 0,0,3,-1,0 \} x1?(n)={0,0,3,?1,0}

x 2 ( n ) = { ? 1 , 0 , 1 , 3 , 2 } x_2(n) = \{ -1,0,1,3,2 \} x2?(n)={?1,0,1,3,2}

x 3 ( n ) = { 3 , 2 , 0 , 1 , 2 } x_3(n) = \{ 3,2,0,1,2 \} x3?(n)={3,2,0,1,2}

x 4 ( n ) = { 1 , 2 , 1 , 0 , 0 } x_4(n) = \{ 1,2,1,0,0 \} x4?(n)={1,2,1,0,0}

h ( n ) = { 1 , 1 , 1 , 0 , 0 } h(n) = \{1,1,1,0,0\} h(n)={1,1,1,0,0}

2.calculate

x 1 ( n ) ? h ( n ) = { ? 1 , 0 , 3 , 2 , 2 } x_1(n)\circledast h(n) =\{-1,0,3,2,2\} x1?(n)?h(n)={?1,0,3,2,2}

x 2 ( n ) ? h ( n ) = { 4 , 1 , 0 , 4 , 6 } x_2(n)\circledast h(n)=\{4,1,0,4,6\} x2?(n)?h(n)={4,1,0,4,6}

x 3 ( n ) ? h ( n ) = { 6 , 7 , 5 , 3 , 3 } x_3(n)\circledast h(n) =\{6,7,5,3,3\} x3?(n)?h(n)={6,7,5,3,3}

x 4 ( n ) ? h ( n ) = { 1 , 3 , 4 , 3 , 1 } x_4(n)\circledast h(n) =\{1,3,4,3,1\} x4?(n)?h(n)={1,3,4,3,1}

3.Remove first (M-1) points ,Concatenate all results

the result is { 3 , 2 , 2 , 0 , 4 , 6 , 5 , 3 , 3 , 4 , 3 , 1 } \{3,2,2,0,4,6,5,3,3,4,3,1\} {3,2,2,0,4,6,5,3,3,4,3,1}

?

?

z-transformation

  • Basic concepts of z-transformation
  • ROC
  • Properties of z-transformation
  • Inverse transformation of z-transform
  • stable

Basic concepts of z-transformation

  • Bilateral Z Transform

  • Unilateral Z Transform

Bilateral Z Transform:

X ( Z ) = Z { x [ n ] } = ∑ n = ? ∞ + ∞ x [ n ] Z ? n , Z ∈ R x X(Z)=Z\left\{x[n]\right\}=\sum_{n=-\infty}^{+\infty}x\left[n\right]Z^{-n}\text{,}Z\in Rx X(Z)=Z{x[n]}=n=?+?x[n]Z?n,ZRx

Unilateral Z Transform:

X ( Z ) = Z { x [ n ] } = ∑ n = 0 + ∞ x [ n ] Z ? n , Z ∈ R x X(Z)=Z\left\{x[n]\right\}=\sum_{n=0}^{+\infty}x\left[n\right]Z^{-n},Z\in Rx X(Z)=Z{x[n]}=n=0+?x[n]Z?n,ZRx

?

region of convergence (ROC)

making z = r ? e j w z = r·e^{jw} z=r?ejw

x ( z ) = ∑ n = ? ∞ ∞ x ( n ) z ? n = ∑ n = ? ∞ ∞ x ( n ) r ? n e ? j w n x(z) = \sum_{n=-\infty}^{\infty}x(n)z^{-n} \\=\sum_{n=-\infty}^{\infty}x(n)r^{-n}e^{-jwn} x(z)=n=??x(n)z?n=n=??x(n)r?ne?jwn

if r = 1 r = 1 r=1,then x ( z ) = D T F T x(z) = DTFT x(z)=DTFT

eg: x ( n ) = { 1 , 2 , 3 , 4 , 0 , 1 } x(n) = \{1,2,3,4,0,1 \} x(n)={1,2,3,4,0,1}

x ( z ) = ∑ n = ? ∞ ∞ x ( n ) z ? n = ∑ n = 0 5 x ( n ) z ? n = x ( 0 ) z ? 0 + x ( 1 ) z ? 1 + x ( 2 ) z ? 2 + x ( 3 ) z ? 3 + x ( 4 ) z ? 4 + x ( 5 ) z ? 5 = 1 + 2 z ? 1 + 3 z ? 2 + 4 z ? 3 + z ? 5 x(z) = \sum_{n=-\infty}^{\infty}x(n)z^{-n}=\sum_{n=0}^{5}x(n)z^{-n}=x(0)z^{-0}+x(1)z^{-1}+x(2)z^{-2}+x(3)z^{-3}+x(4)z^{-4}+x(5)z^{-5}\\=1+2z^{-1}+3z^{-2}+4z^{-3}+z^{-5} x(z)=n=??x(n)z?n=n=05?x(n)z?n=x(0)z?0+x(1)z?1+x(2)z?2+x(3)z?3+x(4)z?4+x(5)z?5=1+2z?1+3z?2+4z?3+z?5

ROC:exist entire z-plane except z = 0 z = 0 z=0

using plot:

eg: x ( n ) = { a n ?? ; n ? ≥ 0 0 ?? ; n ? < 0 x(n)=\left\{\begin{matrix} a^n\;;n\ \ge 0 \\ 0\;;n\ < 0 \end{matrix}\right. x(n)={an;n?00;n?<0?

x ( z ) = ∑ n = 0 + ∞ ( a z ? 1 ) n = 1 1 ? a z ? 1 = z z ? a x(z) = \sum_{n=0}^{+\infty}(az^{-1})^{n}\\=\frac{1}{1-az^{-1}}=\frac{z}{z-a} x(z)=n=0+?(az?1)n=1?az?11?=z?az?

needing that : ∣ a z ? 1 ∣ < 1 |az^{-1}|<1 az?1<1 just ROC: ∣ z ∣ > ∣ a ∣ |z|> |a| z>a

eg : x ( n ) = ? a n u ( ? n ? 1 ) x(n) = -a^nu(-n-1) x(n)=?anu(?n?1)

ROC: ∣ a ? 1 ∣ < 1 |a^{-1}|<1 a?1<1 is just ∣ z ∣ < ∣ a ∣ |z| < |a| z<a

?

Properties of z-transformation

  • linear
  • time shifting
  • scaling
  • differential
  • convolution
  • initial value theorem
  • terminal value theorem
linear

x 1 ( n ) ? x 1 ( z ) ; R O C 1 x_1(n) \longrightarrow x_1(z);ROC_1 x1?(n)?x1?(z);ROC1?

x 2 ( n ) ? x 2 ( z ) ; R O C 2 x_2(n) \longrightarrow x_2(z);ROC_2 x2?(n)?x2?(z);ROC2?

a x 1 ( n ) + b x 2 ( n ) ? a x 1 ( z ) + b x 2 ( z ) ; R O C : [ R O C 1 ∩ R O C 2 ] ax_1(n) + bx_2(n) \longrightarrow ax_1(z)+bx_2(z);\\ROC:[ROC_1\cap ROC_2] ax1?(n)+bx2?(n)?ax1?(z)+bx2?(z);ROC:[ROC1?ROC2?]

time shifting

x ( n ) ? x ( z ) x(n) \longrightarrow x(z) x(n)?x(z)

x ( n ? n 0 ) ? x ( z ) z ? n 0 x(n - n_0) \longrightarrow x(z)z^{-n_0} x(n?n0?)?x(z)z?n0?

scaling

x ( n ) ? x ( z ) ; R O C : ?? ∣ z ∣ > 1 x(n) \longrightarrow x(z);ROC: \;|z|>1 x(n)?x(z);ROC:z>1

a n x ( n ) ? x ( z a ) a^nx(n) \longrightarrow x(\frac{z}{a}) anx(n)?x(az?)

differential

x ( n ) ? x ( z ) x(n) \longrightarrow x(z) x(n)?x(z)

n x ( n ) ? ? z d x ( z ) d z nx(n) \longrightarrow -z \frac{dx(z)}{dz} nx(n)??zdzdx(z)?

convolution

x 1 ( n ) ? x 2 ( n ) ? x 1 ( z ) ? x 2 ( z ) x_1(n) * x_2(n) \longrightarrow x_1(z)·x_2(z) x1?(n)?x2?(n)?x1?(z)?x2?(z)

initial value theorem

x ( 0 ) = lim ? n → 0 x ( n ) = lim ? z → + ∞ x ( z ) x(0) = \lim_{n \to 0}x(n) = \lim_{z \to +\infty}x(z) x(0)=limn0?x(n)=limz+?x(z)

terminal value theorem

x ( + ∞ ) = lim ? n → ∞ x ( n ) = lim ? z → 1 ( 1 ? z ? 1 ) x ( z ) x(+\infty) = \lim_{n \to \infty}x(n) = \lim_{z \to 1}(1-z^{-1})x(z) x(+)=limn?x(n)=limz1?(1?z?1)x(z)

?

Inverse transformation of z-transform

  • long division method
  • partial fraction expansion method
  • Residue method
long division method
image-20231224203722689
partial fraction expansion method

x ( z ) = z ( z ? 1 2 ) ( z + 1 2 ) ( z + 1 4 ) x(z) = \frac{z(z-\frac{1}{2})}{(z+\frac{1}{2})(z+\frac{1}{4})} x(z)=(z+21?)(z+41?)z(z?21?)?

x ( z ) z = z ? 1 2 ( z + 1 2 ) ( z + 1 4 ) = 4 z + 1 2 ? 3 z + 1 4 \frac{x(z)}{z} = \frac{z-\frac{1}{2}}{(z+\frac{1}{2})(z+\frac{1}{4})}=\frac{4}{z+\frac{1}{2}}-\frac{3}{z+\frac{1}{4}} zx(z)?=(z+21?)(z+41?)z?21??=z+21?4??z+41?3?

x ( z ) = 4 z z + 1 2 ? 3 z z + 1 4 x(z) = \frac{4z}{z+\frac{1}{2}} - \frac{3z}{z+\frac{1}{4}} x(z)=z+21?4z??z+41?3z?

x ( n ) = 4 ( ? 1 2 ) n u ( n ) ? 3 ( ? 1 4 ) n u ( n ) x(n) = 4(-\frac{1}{2})^{n}u(n) -3 (-\frac{1}{4})^{n}u(n) x(n)=4(?21?)nu(n)?3(?41?)nu(n)

Residue method

x ( z ) = 1 ( z ? 2 ) ( z ? 3 ) x(z) = \frac{1}{(z-2)(z-3)} x(z)=(z?2)(z?3)1?

1. z n ? 1 x ( z ) = z n ? 1 ( z ? 2 ) ( z ? 3 ) z^{n-1}x(z) = \frac{z^{n-1}}{(z-2)(z-3)} zn?1x(z)=(z?2)(z?3)zn?1?

2. R 1 = ( z ? 2 ) z n ? 1 ( z ? 2 ) ( z ? 3 ) ∣ z = 2 = ? 2 n ? 1 R_1 =(z-2)\frac{z^{n-1}}{(z-2)(z-3)}|_{z=2} = -2^{n-1} R1?=(z?2)(z?2)(z?3)zn?1?z=2?=?2n?1

3. R 2 = ( z ? 3 ) z n ? 1 ( z ? 2 ) ( z ? 3 ) ∣ z = 3 = 3 n ? 1 R_2 =(z-3)\frac{z^{n-1}}{(z-2)(z-3)}|_{z=3} = 3^{n-1} R2?=(z?3)(z?2)(z?3)zn?1?z=3?=3n?1

4. x ( n ) = ? 2 n ? 1 + 3 n ? 1 x(n) = -2^{n-1} + 3^{n-1} x(n)=?2n?1+3n?1

?

stability

S = ∑ k = ? ∞ ∞ ∣ h ( n ) ∣ < + ∞ S=\sum_{k=-\infty}^{\infty}|h(n)| < +\infty S=k=??h(n)<+

文章来源:https://blog.csdn.net/u011146203/article/details/135209254
本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。