Discrete Time Signals and Systems

Discrete Time Signals and Systems

文章目录

  • Discrete Time Signals and Systems
    • Signal classification
    • basic signal
    • Operation on signal
    • System of discrete signal
      • Linear systems and nonlinear systems
      • Causal and non-causal Systems
      • Time-varying and time-invariant systems
      • static system and dynamic system
      • Stable and unstable systems
      • convolution
      • circular convolution
      • Stability of linear time-invariant systems
    • Fourier Series
      • Fourier Series basic concepts
      • Discrete time fouriew series (DTFS)
      • CTFS
    • Continuous time Fourier transform (CTFT)
      • Dirichlet's conditions
      • lecture(77)
      • lecture(78)
      • properties of CTFT
        • linearity
        • time shifting
        • time scaling
      • frequency shifting
      • time differenciation
      • frequency differenciation
      • convolution
      • Parseval's Theorem
    • Fourier transform of Periodic signal
    • Discrete time fouriew series (DTFT)
      • basic properties
      • DTFT for common signals
      • properties
        • linear
        • time shifting
        • frequency shifting
        • Frequency domain differential
    • Discrete Fourier Transform(DFT)
      • Calculation of DFT
        • Rotation matrix of inverse DFT transform(IDFT)
      • Properties of DFT
        • 1.linear
        • 2.circular convolution
        • 3.periodicity
        • 4.cyclic frequency shift
      • DTF calculates linear convolution and circular convolution
        • linear convilution
        • cyclical convolution
      • another way to calculate linear convolution
        • Overlap-Add
        • Overlap-Save
    • z-transformation
      • Basic concepts of z-transformation
      • region of convergence (ROC)
      • Properties of z-transformation
        • linear
        • time shifting
        • scaling
        • differential
        • convolution
        • initial value theorem
        • terminal value theorem
      • Inverse transformation of z-transform
        • long division method
        • partial fraction expansion method
        • Residue method
      • stability

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[2πf0n]=x(n+N)=Acos[2πf0(n+N)]=Acos[2πf0n+2πf0N]
2 π f 0 N = 2 π k 2\pi f_0 N =2 \pi k 2πf0N=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 N2N1=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=limN2N+11n=+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=N1n=0N1x[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(n1)+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(nk)

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(nk)

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(ancos(nw0t)+bnsin(nw0t)

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=T110T1f(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=T120T1f(t)cosnw1tdt

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=T120T1f(t)sinnw1tdt

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=0N1c(k)eJkN2πn

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)=N1n=0N1x(n)eJkN2πn

 

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)eJwtdt

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)eJwtdt

    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)]F1[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(tt0)]F1[F(ωω0)]=et0F(ω)=ejω0tf(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)]=a1F(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)ejw0t]F1[f(t)ejw0t]=F(ωω0)=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)F1[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)]=2π1F1(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 } ejω0t and e − j ω 0 t e^{- j\omega_0t } ejω0t
  • 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 } ejω0t and e − j ω 0 t e^{- j\omega_0t } ejω0t

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[ejω0t]F[ejω0t]=2πδ(ωω0)=2πδ(ω+ω0)

 

spectrum of s i n w 0 t sinw_0t sinw0t and c o s w 0 t cosw_0t cosw0t

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(ω0t)]=F[2ejω0t+ejω0t]=π[δ(ωω0)+δ(ω+ω0)]F[sin(ω0t)]=F[2jejω0tejω0t]=[δ(ω+ω0)δ(ωω0)]

 

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)ejnw0t

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(ωnω0)

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)=T10Tf(t)ejnw0tdt

 

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]ejωn

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]=2π12πX(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)ejwn=n=0+anejwn=n=0+(aejw)n=1aejw1=1a(coswjsinw)1=(1acosw)+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)=(1acosw)2 +(asinw)21=1+a22acosw 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)=arctan1acoswasinw

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 ππx(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(nn0)x(ejw)ejwn0

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)ejwn0x(ej(ww0))

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(n1)=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)43y(n1)+81y(n2)=2x(n)

3. x ( e j w ) = e − j w x(e^{jw})=e^{-jw} x(ejw)=ejw 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)=21nu(n) and x ( n ) = 1 3 n u ( n ) x(n) = \frac{1}{3}^n u(n) x(n)=31nu(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=0N1x(n)ejN2πkn;k=0,1,2,3,...,N1

making W N = e − j 2 π N W_N = e^{-j\frac{2\pi}{N}} WN=ejN2π

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=0N1x(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=041x(n)ej42πkn

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=0N1x(n)ejN2πkn;k=0,1,2,3,...,N1

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)ejN2πln]=x((kl))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]Zn,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]Zn,ZRx

 

region of convergence (ROC)

making z = r ⋅ e j w z = r·e^{jw} z=rejw

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)zn=n=x(n)rnejwn

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)zn=n=05x(n)zn=x(0)z0+x(1)z1+x(2)z2+x(3)z3+x(4)z4+x(5)z5=1+2z1+3z2+4z3+z5

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+(az1)n=1az11=zaz

needing that : ∣ a z − 1 ∣ < 1 |az^{-1}|<1 az1<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(n1)

ROC: ∣ a − 1 ∣ < 1 |a^{-1}|<1 a1<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:[ROC1ROC2]

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(nn0)x(z)zn0

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)=limn0x(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(+)=limnx(n)=limz1(1z1)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(z21)

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)z21=z+214z+413

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+214zz+413z

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)=(z2)(z3)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)} zn1x(z)=(z2)(z3)zn1

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=(z2)(z2)(z3)zn1z=2=2n1

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=(z3)(z2)(z3)zn1z=3=3n1

4. x ( n ) = − 2 n − 1 + 3 n − 1 x(n) = -2^{n-1} + 3^{n-1} x(n)=2n1+3n1

 

stability

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

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。如若转载,请注明出处:http://www.mfbz.cn/a/271186.html

如若内容造成侵权/违法违规/事实不符,请联系我们进行投诉反馈qq邮箱809451989@qq.com,一经查实,立即删除!

相关文章

助力打造清洁环境,基于美团最新YOLOv6-4.0开发构建公共场景下垃圾堆放垃圾桶溢出检测识别系统

公共社区环境生活垃圾基本上是我们每个人每天几乎都无法避免的一个问题&#xff0c;公共环境下垃圾投放点都会有固定的值班时间&#xff0c;但是考虑到实际扔垃圾的无规律性&#xff0c;往往会出现在无人值守的时段内垃圾堆放垃圾桶溢出等问题&#xff0c;有些容易扩散的垃圾比…

使用travelbook架设自己的实时位置共享服务

travelbook 是一款开源的安卓APP&#xff0c;它能以低功耗提供实时位置共享&#xff0c;它包含功能如下&#xff1a; 好友之间分享实时位置&#xff1b;记录行程轨迹&#xff1b;标记收藏地点&#xff1b; 这款软件的主要解决的问题包括&#xff1a; 场景1&#xff1a;查看老…

【开源】基于Vue+SpringBoot的新能源电池回收系统

目录 一、摘要1.1 项目介绍1.2 项目录屏 二、功能模块2.1 用户档案模块2.2 电池品类模块2.3 回收机构模块2.4 电池订单模块2.5 客服咨询模块 三、系统设计3.1 用例设计3.2 业务流程设计3.3 E-R 图设计 四、系统展示五、核心代码5.1 增改电池类型5.2 查询电池品类5.3 查询电池回…

安防视频云平台/可视化监控云平台EasyCVR如何快速定位占用大量存储空间的文件?

安防视频监控/视频集中存储/云存储/磁盘阵列EasyCVR平台可拓展性强、视频能力灵活、部署轻快&#xff0c;可支持的主流标准协议有国标GB28181、RTSP/Onvif、RTMP等&#xff0c;以及支持厂家私有协议与SDK接入&#xff0c;包括海康Ehome、海大宇等设备的SDK等。平台既具备传统安…

【Vue2+3入门到实战】(4)Vue基础之指令修饰符 、v-bind对样式增强的操作、v-model应用于其他表单元素 详细示例

目录 一、今日学习目标1.指令补充 二、指令修饰符1.什么是指令修饰符&#xff1f;2.按键修饰符3.v-model修饰符4.事件修饰符 三、v-bind对样式控制的增强-操作class1.语法&#xff1a;2.对象语法3.数组语法4.代码练习 四、京东秒杀-tab栏切换导航高亮1.需求&#xff1a;2.准备代…

小白的实验室服务器深度学习环境配置指南

安装nvidia 本文在ubuntu server 22.04上实验成功&#xff0c;其他版本仅供参考 注意&#xff0c;本文仅适用于ubuntu server&#xff0c;不需要图形界面&#xff0c;没有对图形界面进行特殊考虑和验证&#xff01;依赖图形操作界面的读者慎用 查看是否安装了gcc gcc -v若没…

如何快速删除pdf周围的空白

问题&#xff1a;写论文往往需要pdf格式的图片&#xff0c;但pdf往往四周存在大量空白需要手动截图很麻烦 解决&#xff1a; 打开命令行输入&#xff1a;pdfcrop 图片名.pdf

【Mysql】InnoDB统计数据的收集(十三)

我们前边在计算查询成本的时候会用到一些统计数据&#xff0c;比如通过 SHOW TABLE STATUS 可以看到关于表的统计数据&#xff0c;通过 SHOW INDEX 可以看到关于某个索引的统计数据&#xff0c;那么这些统计数据是怎么来的呢&#xff1f;本章节将分享 InnoDB 存储引擎的统计数据…

深圳锐科达SIP矿用电话模块SV-2801VP

深圳锐科达SIP矿用电话模块SV-2801VP 一、简介 SV-2800VP系列模块是我司设计研发的一款用于井下的矿用IP音频传输模块&#xff0c;可用此模块打造一套低延迟、高效率、高灵活和多扩展的IP矿用广播对讲系统&#xff0c;亦可对传统煤矿电话系统加装此模块&#xff0c;进行智能化…

在Vue3中使用vue-qrcode库实现二维码生成

本文主要介绍在Vue3中使用qrcode库实现二维码生成的方法。 目录 一、基础用法实现vue-qrcode库的参数介绍 在Vue3中实现二维码生成需要使用第三方库来处理生成二维码的逻辑。常用的库有 qrcode和 vue-qrcode。 一、基础用法实现 在Vue3中使用vue-qrcode库实现二维码生成的方…

记录一次云主机故障排查

云上某云主机&#xff0c;在安全组test-a中&#xff0c;同一安全组下还有另外两台主机。 从本地可以ping 通另外两台主机的公网地址。但是不能ping通这个主机的公网地址。 与是重启主机&#xff0c;发现问题依然存在。依然是不能ping 通&#xff0c;主机上部署的业务也不能访…

LabVIEW在齿轮箱故障诊断中的应用

LabVIEW在齿轮箱故障诊断中的应用 在现代机械工业中&#xff0c;齿轮箱作为重要的传动设备&#xff0c;其性能稳定性对整体机械系统的运行至关重要。故障的及时诊断和处理不仅保障了设备的稳定运行&#xff0c;还减少了维护成本。利用LabVIEW强大数据处理和仿真能力&#xff0…

新版IDEA中Git的使用(二)

说明&#xff1a;前面介绍了在新版IDEA中Git的基本操作&#xff0c;本文介绍关于分支合并、拉取等操作&#xff1b; 例如&#xff0c;现在有一个项目&#xff0c;分支如下&#xff1a; main&#xff1a;主分支&#xff1b; dev&#xff1a;开发分支&#xff1b; test&#x…

Springboot整合MVC进阶篇

一、概述 1.1SpringBoot整合SpringMVC配置 SpringBoot对SpringMVC的配置主要包括以下几个方面&#xff1a; 自动配置&#xff1a;SpringBoot会自动配置一个嵌入式的Servlet容器&#xff08;如Tomcat&#xff09;&#xff0c;并为我们提供默认的SpringMVC配置。这样我们无需手动…

【Java、Python】获取电脑当前网络IP进行位置获取(附源码)

我相信看到这篇博客的时候心里肯定是想解决自己的一个问题的&#xff0c;而这篇博客我就以简单快速的方式解决这些烦恼&#xff01; 一、获取当前IP 在Java中自带了一些自己的流对象来获取当前的IP地址&#xff0c;不多说我们直接上代码。 //获取当前网络ip地址 ipAddress Ine…

在k8s中将gitlab-runner的运行pod调度到指定节点

本篇和前面的 基于helm的方式在k8s集群中部署gitlab 具有很强的关联性&#xff0c;因此如果有不明白的地方可以查看往期分享&#xff1a; 基于helm的方式在k8s集群中部署gitlab - 部署基于helm的方式在k8s集群中部署gitlab - 备份恢复基于helm的方式在k8s集群中部署gitlab - 升…

论文阅读——X-Decoder

Generalized Decoding for Pixel, Image, and Language Towards a Generalized Multi-Modal Foundation Model 1、概述 X-Decoder没有为视觉和VL任务开发统一的接口&#xff0c;而是建立了一个通用的解码范式&#xff0c;该范式可以通过采用共同的&#xff08;例如语义&#…

实战:朴素贝叶斯文本分类器搭建与性能评估

&#x1f497;&#x1f497;&#x1f497;欢迎来到我的博客&#xff0c;你将找到有关如何使用技术解决问题的文章&#xff0c;也会找到某个技术的学习路线。无论你是何种职业&#xff0c;我都希望我的博客对你有所帮助。最后不要忘记订阅我的博客以获取最新文章&#xff0c;也欢…

生成超清分辨率视频,南洋理工开源Upscale-A-Video

大模型在生成高质量图像方面表现出色,但在生成视频任务中&#xff0c;经常会面临视频不连贯、图像模糊、掉帧等问题。 这主要是因为生成式抽样过程中的随机性,会在视频序列中引入无法预测的帧跳动。同时现有方法仅考虑了局部视频片段的时空一致性,无法保证整个长视频的整体连贯…

基于电商场景的高并发RocketMQ实战-Broker写入读取流程性能优化总结、Broker基于Pull模式的主从复制原理

&#x1f308;&#x1f308;&#x1f308;&#x1f308;&#x1f308;&#x1f308;&#x1f308;&#x1f308; 【11来了】文章导读地址&#xff1a;点击查看文章导读&#xff01; &#x1f341;&#x1f341;&#x1f341;&#x1f341;&#x1f341;&#x1f341;&#x1f3…
最新文章