MIT res 6.007 signals & systems video note

14:24 2014-2-8 Saturday


mit res 6.007 signals & systems, video 1


15:47 2014-2-8
continue lec 2, sinusoidal signal


16:06 2014-2-8
continuous sinusoid: time shift <-> phase change


discrete sinusoid: time shift -> phase change, but...


the reverse is not true!


it's not necessarily true


16:09 2014-2-8
periodic?


continuous sinusoid always periodic,


discrete-time ~ is not necessarily periodic


16:17 2014-2-8
real exponential,


complex exponential


16:36 2014-2-8
90 degree out of phase


16:45 2014-2-8
///
12:54 2014-2-9
signals & systems, lec 3


12:54 2014-2-9
unit step & unit impulse


12:54 2014-2-9
memoryless




15:43 2014-2-9
start signals & systems, lec 4


convolution


15:44 2014-2-9
decomposition:


1. unit impulse <-> convolution


2. complex exponential <-> fourier analysis


15:52 2014-2-9
convolution sum


16:00 2014-2-9
convolution sum, convolution integral


16:23 2014-2-9
property of LTI systems


17:27 2014-2-9
commutative property


17:27 2014-2-9
for LTI system, stable means:


stable -> h[n] absolutely summable(absolutely integrable)


18:09 2014-2-9
linear system -> ZIZO


18:09 2014-2-9
initially at rest


18:16 2014-2-9
system properties within the playground of LTI:


memory?


causal?


stability?


18:23 2014-2-9
many system properties can be determined from "impulse response"


in the LTI playground


18:23 2014-2-9
LCCDE + initially at rest => LTI system


18:38 2014-2-9
operational definition(generalized function)


18:55 2014-2-9
singulatity functions


18:57 2014-2-9
start signals & systems, lec 6


systems described by LCCDE


19:04 2014-2-9
LCCDE == Linear Constant-Coefficent Differential Equations


L == Linear combination


19:14 2014-2-9
homogeneous solution


19:29 2014-2-9
initial rest:


output remains zero when t < to, until input acts when t == t0


because LTI => ZIZO


19:32 2014-2-9
for LCCDE:


initially rest at t <= to => causal + LTI


auxilliary condition == 0 at t == t0 => linear


19:32 2014-2-9
special solution -> linear combination == homogeneous solution


paticular solution


X == Xp + Xh


19:51 2014-2-9
for LCCDE, 


initial rest <=> causal, LTI


19:57 2014-2-9
causal, LTI <=> initial rest // for LCCDE


20:42 2014-2-9
start signals & systems, lec 7


continuous-time Fourier Series


20:47 2014-2-9
decomposition:


impulse -> convolution


complex exponential -> fourier series


22:01 2014-2-9
buiding block: complex exponential


22:02 2014-2-9
eigenfunction property


22:03 2014-2-9
imaginary exponential


22:04 2014-2-9
complex constant: H(w)


22:09 2014-2-9
complex exponential is the "eigenfunction" of LTI system


22:09 2014-2-9
H(w) // frequency response is the "eigenvalue" of the LTI system


22:10 2014-2-9
Fourier analysis:


1. Fourier series:    periodic signals


2. Fourier transform: aperiodic signals


22:11 2014-2-9
harmonically related


22:31 2014-2-9
complex exponential form of "Fourier series"


22:42 2014-2-9
analysis equation, synthesis equation


22:42 2014-2-9
Fourier series coefficient


22:50 2014-2-9
trignometric form of Fourier series


22:52 2014-2-9
convergence of the Fourier series


* square integrable


* Direchlet condition // more stricter condition


23:39 2014-2-9
energy in the error


23:39 2014-2-9
start signals & systems, lec 8


continuous-time Fourier Transform


23:55 2014-2-9
linear combination of complex exponentials


23:55 2014-2-9
periodically replicating an aperiodic signal


23:56 2014-2-9
derive "Fourier transform" from "Fourier series"


using periodic replication!


0:14 2014-2-10
convergence condition


0:28 2014-2-10
x(t) is real => X(w) is complex


0:39 2014-2-10
Bode plot


0:46 2014-2-10
since we now know the relationship between Fourier series


& Fourier transform, we can also get Fourier series coefficient


from Fourier transform spectrum:


by sampling the envelope!


0:47 2014-2-10
What is the Fourier transform of periodic signal?


// ordinarily, periodic signal only has Fourier series defined!


1:51 2014-2-10
go to bed...


/
14:09 2014-2-10 Monday
signals & systems, lec 9


Fourier transform's property


14:10 2014-2-10
x(t) => X(w) conjugate symmetric


14:10 2014-2-10
frequency response


14:56 2014-2-10
convolution property:


x(t) * h(t) <=> X(w) H(w) 


// for LTI only, interpreted with 


//eigenfunction


14:56 2014-2-10
for LCCDE, 


causal, LTI <=> initial rest


15:25 2014-2-10
partial fraction expansion


15:41 2014-2-10
inverse Fourier transform 


15:42 2014-2-10
start signals & systems, lec 10


Discrete-Time Fourier Series // DFS


15:49 2014-2-10
DTFS is the same as CTFS:


decompose signal into linear combination of eigenfunctions!


// complex exponential


16:00 2014-2-10
eigenfunction property


16:00 2014-2-10
fundamental period, fundamental frequency


16:23 2014-2-10
condition of convergence


16:53 2014-2-10
periodically replicating the aperiodic signal


17:18 2014-2-10
impulse train


18:42 2014-2-10
start signals & systems, lec 11


DTFT


19:20 2014-2-10
the property of DTFT


19:20 2014-2-10
frequency response  // eigenvalue


complex exponential // eigenfunction


19:35 2014-2-10
in the discrete-time case(DTFT), the frequency response


is periodic


19:41 2014-2-10
for LCCDE, initial rest <=> causal + LTI


19:47 2014-2-10
LPF == Low Pass Filter, HPF == High Pass Filter


19:56 2014-2-10
convolution property,


modulation property


20:29 2014-2-10
periodic convolution  // convolving periodic functions!


20:39 2014-2-10
duality


21:11 2014-2-10
convolution property -> filtering


modulation property -> modulation


21:17 2014-2-10
signals & systems, lec 12


filtering


21:18 2014-2-10
x[n] * h[n] <=> X(w) H(w)  // convolution property


21:29 2014-2-10
impulse response of the ideal LPF


21:34 2014-2-10
frequency response <=> impulse response


21:35 2014-2-10
Sa(x) = sin(x) / x is noncausal!


21:37 2014-2-10
step response of an ideal LPF


21:42 2014-2-10
nonideal filter


21:43 2014-2-10
implementation of a nonideal LPF


21:49 2014-2-10
DE == Differential Equation


21:51 2014-2-10
Bode plot


21:57 2014-2-10
cascade several filter of the same type, what happens?


21:57 2014-2-10
equalization filter


22:07 2014-2-10
graphic equalizer


22:18 2014-2-10
Moving Average filter // nonrecursive filter


22:18 2014-2-10
MA == Moving Average


22:19 2014-2-10
equal-ripple characteristic


22:24 2014-2-10
moving average filters // nonrecursive filters


22:28 2014-2-10
recursive filters // has feedback


22:28 2014-2-10
Moving Average filter + recursive filter


22:37 2014-2-10
start signals & systems, lec 13


continuous-time modulation


22:40 2014-2-10
impulse train


22:49 2014-2-10
FM == Frequency Modulation


22:51 2014-2-10
AM == Amplitude Modulation


22:54 2014-2-10
sinusoidal amplitude modulation


22:54 2014-2-10
PM == Phase Modulation


22:55 2014-2-10
modulating signal * carrier signal == modulated output


22:56 2014-2-10
complex exponential carrier


23:00 2014-2-10
input signal == modulating signal


23:01 2014-2-10
demodulation


23:04 2014-2-10
cosine modulator, sine modulator


23:13 2014-2-10
modulation // shifting that spectrum


23:26 2014-2-10
complex exponential modulation


23:32 2014-2-10
sinusoidal modulation


23:32 2014-2-10
modulator/demodulator


23:41 2014-2-10
shift the signal into another frequency band,...


23:44 2014-2-10
broadband signal


23:46 2014-2-10
multiplexing


23:49 2014-2-10
FDM == Frequency Division Multiplexing


23:49 2014-2-10
demultiplexing + demodulation


23:49 2014-2-10
synchronous modulation


23:51 2014-2-10
must keep synchronization between transmitter & receiver


23:51 2014-2-10
synchronization in terms of phase


23:53 2014-2-10
phase difference between the modulator & demodulator


23:56 2014-2-10
asynchronous demodulation:


diode + RC LPF // tracks the envelope


23:58 2014-2-10
envelope detector


0:06 2014-2-11
asynchronous demodulation: radio


synchronous demodulation:  satellite


0:10 2014-2-11
SSB == Single-Side Band


0:14 2014-2-11
HPF == High-Pass Filter


0:16 2014-2-11
DSB == Double Side Band


0:17 2014-2-11
go to bed!


//
13:01 2014-2-11 Tuesday


start lec 14, demonstration of amplitude modulation


13:01 2014-2-11
envelope detector


13:59 2014-2-11
injecting some additional carrier


14:00 2014-2-11
over-modulated


14:16 2014-2-11
superhyterodyne receiver


14:31 2014-2-11
IF stage


14:31 2014-2-11
IF == Intermediate Frequency


14:31 2014-2-11
start lec 15, discrete-time modulation


14:38 2014-2-11
synchronous demodulation


asynchronous demodulation  // envelope detector


14:53 2014-2-11
PAM == Pulse Amptitude Modulation


14:53 2014-2-11
periodic convolution  // DTFT


14:59 2014-2-11
complex exponential carrier,


sinusoidal carrier


15:05 2014-2-11
discrete-time modulation // analysis using DTFT


15:13 2014-2-11
synchronization between the modulator & demodulator


// frequency synchronization, phase synchronization


15:18 2014-2-11
synchronization is much important for continuous systems


than discrete-time systems


15:20 2014-2-11
pulse train


15:34 2014-2-11
modulating signal // original signal


carrier signal


modulated signal


15:46 2014-2-11
FDM, TDM == Time Division Multiplexing


16:03 2014-2-11
impulse train


16:21 2014-2-11
sampling


16:26 2014-2-11
sampling theorem


16:43 2014-2-11
equally spaced sample


16:44 2014-2-11
start lec 16, sampling


17:01 2014-2-11
sampling theorem


17:05 2014-2-11
sampling is a kind of modulation with impulse train


17:05 2014-2-11
discrete-time processing of continuous-time signals


17:49 2014-2-11
what is a "phase reversal"?


21:39 2014-2-11
start signals & systems, lec 17,


interpolation


21:39 2014-2-11
sampling process, reconstruction process


21:41 2014-2-11
reconstructed signal


21:49 2014-2-11
band-limited interpolation


21:55 2014-2-11
ZOH == Zero Order Hold   // rectangular impulse response


21:56 2014-2-11
FOH == First Order Hold  // triangular impulse response


22:02 2014-2-11
ZOH                  // rectangular


FOH                  // triangular 


// different impulse response


22:07 2014-2-11
interpolating filter


22:10 2014-2-11
undersample


22:38 2014-2-11
relabling process


22:39 2014-2-11
frequency normalization  // corresponding time normalization


22:43 2014-2-11
sampling period: T


22:53 2014-2-11
discrete-time processing of continuous-time signals


22:53 2014-2-11
sequence


22:54 2014-2-11
C/D == Continuous to Discrete converter


22:59 2014-2-11
relabling process: 


conversion of impulse train to sequence


23:00 2014-2-11
frequency normalization


23:03 2014-2-11
anti-aliasing filter


23:07 2014-2-11
desample through a LPF


23:18 2014-2-11
reconstructing LPF


23:22 2014-2-11
digital filter


discrete-time filter


23:26 2014-2-11
desampling filter


0:13 2014-2-12
effective filter cutoff frequency depends on


sampling frequency!  // digital filter


//
11:11 2014-2-12 Wednesday


lec 19, sampling of a sequence


11:11 2014-2-12
sampling rate conversion


12:29 2014-2-12
decimation/interpolation


12:30 2014-2-12
filter impulse response


12:52 2014-2-12
sampled sequence -> decimated sequence


13:09 2014-2-12
sampling process


decimation process


13:10 2014-2-12
discrete-time sampling == decimation


13:20 2014-2-12
down-sampling


13:32 2014-2-12
decimated sequence -> sampled sequence


13:33 2014-2-12
decimation/interpolation


13:39 2014-2-12
sampling rate conversion


13:40 2014-2-12
assumed sampling period


13:45 2014-2-12
use a combination of upsampling & downsampling


// decimation & interpolation


13:48 2014-2-12
sampling in the frequency domain


14:02 2014-2-12
modulation theorem


sampling theorem


14:14 2014-2-12
time window


14:19 2014-2-12
Laplace transform & z-transform are both


generalization of the Fourier transform


14:22 2014-2-12
start lec 20, Laplace transform


14:23 2014-2-12
the generalization of Fourier transform:


1. Laplace transform


2. z-transform


14:32 2014-2-12
LTI system's


eigenfunction: complex exponential 


eigenvalue:    frequency response  H(w)


14:37 2014-2-12
excitation -> response


14:37 2014-2-12
Lapalace transform is the Fourier transform of


an exponential weighted x(t)


15:00 2014-2-12
analysis equation <-> synthesis equation


15:06 2014-2-12
ROC == Region Of Convergence


15:13 2014-2-12
zeros, poles


15:52 2014-2-12
convergence <-> divergence(blow up)


15:52 2014-2-12
the ROC of Laplace transform must a single strip, 


not multiple strips


16:03 2014-2-12
decaying exponential


16:07 2014-2-12
absolutely integrable


16:07 2014-2-12
gowing exponential


decaying exponential


16:25 2014-2-12
right-sided time function


16:33 2014-2-12
left-sided signal, right-sided signal


16:33 2014-2-12
partial fraction expansion


16:39 2014-2-12
start lec 21, continuous-time 2nd order system


16:58 2014-2-12
for LTI(LSI) system:


BIBO stability <=> absolute integrability of the impulse response


17:18 2014-2-12
stable, causal <=> all poles must lie in the left half s-plane


17:20 2014-2-12
1st order differential equations


17:21 2014-2-12
LCCDE == Linear Constant-Coefficient Differential Equations


17:21 2014-2-12
differential equation


17:27 2014-2-12
2nd order differential equation


17:30 2014-2-12
LCCDE => algebric expression


17:31 2014-2-12
pole vector


17:50 2014-2-12
the frequency response has a resonance


17:50 2014-2-12
resonance


17:51 2014-2-12
speech synthesizer


18:08 2014-2-12
resonator


18:12 2014-2-12
analog speech synthesizer,


digital speech synthesizer


18:18 2014-2-12
start lec 22, z-transform


19:51 2014-2-12
CTFT -> Laplace


DTFT -> z-transform


20:38 2014-2-12
2nd order difference equation


20:38 2014-2-12
unit circle


20:41 2014-2-12
start lec 23, mapping continuous-time filters


to discrete-time filters


20:52 2014-2-12
for LTI,


stable <=> absolutely summable


DTFT{x[n]} converge <=> absolutely summable


20:59 2014-2-12
pole-zero pattern


21:12 2014-2-12
continuous-time filters =>


discrete-time filters   // mapping


21:17 2014-2-12
H(s)  differential equation


H(z)  difference equation


21:18 2014-2-12
discrete-time filters to simulate continuous-time filters


21:22 2014-2-12
mapping differential equations to difference equation


21:22 2014-2-12
backward difference


21:23 2014-2-12
jw-axis => unit circle


21:26 2014-2-12
replacing derivatives to backward differences


21:27 2014-2-12
method 2:


the shape of the impulse response is preserved


"impulse invariance"


21:35 2014-2-12
1st order poles,


2nd order poles


21:37 2014-2-12
partial fraction expansion


21:37 2014-2-12
continuous-time filter => discrete-time filter


21:39 2014-2-12
"impulse invariance"


21:50 2014-2-12
resonant character


21:53 2014-2-12
aliasing


21:54 2014-2-12
forward difference, backward difference


21:57 2014-2-12
bilinear transformation


21:57 2014-2-12
start lec 24, butterworth filter


21:58 2014-2-12


21:27 2014-2-12
method 2:


the shape of the impulse response is preserved


"impulse invariance"


21:35 2014-2-12
1st order poles,


2nd order poles


21:37 2014-2-12
partial fraction expansion


21:37 2014-2-12
continuous-time filter => discrete-time filter


21:39 2014-2-12
"impulse invariance"


21:50 2014-2-12
resonant character


21:53 2014-2-12
aliasing


21:54 2014-2-12
forward difference, backward difference


21:57 2014-2-12
bilinear transformation


21:57 2014-2-12
start lec 24, butterworth filter


21:58 2014-2-12