MIT OpenCourseWare

http://ocw.mit.edu

MAS.160 / MAS.510 / MAS.511 Signals, Systems and Information for Media Technology
Fall 2007

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Period of a Discrete Sinusoid

y [n ] = sin(2 501 n) T=50 samples y [ n ] = sin(2 503 n)

y [ n ] = y[n + 50] y [ n ] = y[n + T] T=?? samples [integer] 50/3 � integer
sin(0) = sin(2 )

irrational frequency
1

0.8

0.6

0.4 y ( t ) = sin(2 2 t)
T = 12 sec

y=sin(2*pi*sqrt(2)/25*n)
0.2

-0.2
continuous function
-0.4
TextEnd periodic
-0.6

-0.8

-1

0 0.5 1 1.5 2 2.5 3 3.5 4

time (sec)

y [ n ] = sin(2 503 n) 2 503 n = 2k Ts=1/25 sec

y [ n ] = sin(2 2
n) 2 252 n = 2k
y [ n ] = y[n + T] T=?? samples n 50 samples
=
Ratio of 25

k 3 cycle integers y [ n ] = y[n + T] T=?? samples n 25 2

sin(0) = sin(2k) =
k=1,2… rational number sin(0) = sin(2k) k=1,2… k 2
periodic
Equiv. discrete sinusoid not periodic irrational number
T=n=50 samples, k=3 cycles
Period of Sum of Sinusoids
1

0.5

-0.5

-1

0 1 2 3 4 5
6

T1=0.2 seconds, T2=.75 seconds

2

-1

-2
0 1 2 3 4 5 6

y ( t ) = y (t + T )
time (sec)

Tsum=3 seconds

Least common multiple Complex Conversions

seconds to complete seconds to complete cartesian polar polar cartesian

j� a tan ( b a )
cycles cycles s=a+jb s= a +b e2 2 s=rej� s = r cos� + jr sin �
T1=1/5 seconds
T2=3/4 seconds
Complex Arithmetic
1/5s, 2/5s, 3/5s … 3/4s, 6/4s, …
Addition cartesian (a1 + jb1 ) + (a2 + jb2 ) = ( a1 + a2 ) + j (b1 + b2 )
4/20s, 8/20s, 12/20s,
15/20s. 30/20s,
16/20s, 20/20s, 24/20s, Subtraction cartesian (a1 + jb1 ) � (a2 + jb2 ) = ( a1 � a2 ) + j (b1 � b2 )
45/20s, 60/20s
28/20s, 32/20s, 36/20s,
40/20s, 44/20s, 48/20s, 4 cycles Multiplication polar r1e j� 1 r2e j� 2 = r1r2e j (� 1 +� 2 )
52/20s, 56/20s, 60/20s 1/5*k=3/4*l r1e j� 1 r1 j (� 1 �� 2 )
rational number Division polar = e
15 cycles k/l=15/4 r2e j� 2 r2

Tsum=15*T1=15/5=3 seconds Tsum=4*T2=3/4*4=3 seconds Powers polar (re ) j� n

= r n e jn�

Tsum=3 seconds
Roots polar z n = s = re j� k = 1,2Kn �1
z = s1/ n = r1/ n e j (� / n + 2 k / n )
Complex Conversions Representations of Sinusoids
cartesian polar polar cartesian Acos(2kf 0 t + � k ) Re{ Ae j 2 ft +� } Ae j� � ( e j 2 ft +e � j 2 ft
2 )
( )
j 3
= Re{ Ae j� e j 2 ft }
j� a tan ( )
3 + j4 = 32 + 4 2 e = 5e j� 0.927 = 2cos 3 + j2sin 3 = 1+ j 3
4
2e e j 2 ft + e � j 2 ft
=X�
3

Complex Arithmetic
= Re{ Xe j 2 ft }

Addition cartesian (1+ j2) + ( 3 + j4 ) = (4 + j6) Sum multiple cosines same frequency
n n n
�A cos(2�ft + � k ) = � Re{ Ak e 2 �ft +� k } = � Re{ Ak e� k e 2 �ft }
(1+ j2) � ( 3 + j4 ) = (�2 � j2)
k
Subtraction cartesian
k=1 k=1 k=1
�n
Multiplication polar 5e
j 3
6e
j 4
= 5 6e ( 3 4 ) = 30e 12
j + j 7 = � Re{ Ak e� k }e 2 �ft
� k=1

10e
j� 2
÷ 5e
j� 4
= ( 105 )e ( 2 4 ) = 2e 4
j� � j�
Ex. 3cos(2 40t + 2 ) �1cos(2 40t � 6 ) + 2 cos(2 40t + 3 )
Division polar
� j �j �
Re� 3e 2 e j2 40t �1e 6 e j2 40t + 2e 3 e j2 40t �
� �

(3e ) = 33 e ( 4 ) = 27e ( 4 )
3
polar j 4 j 3 j 3 �� j � �

Powers �j
Re� 3e 2 �1e 6 + 2e 3 e j2 40t

� � �

z 3 = 64 = 64e j 0 4 Re{5.234 e j1.545e2 40t }

Roots polar
4e j ( 2 / 3)
z = 641/ 3 e j ( 0 / 3+2 k / 3) = 4e j ( 2 k / 3) 4e j ( 4 / 3) 5.234 cos(2 40t + 1.545)

multiply cosines of different frequency

Composite signals (waveform synthesis)
� �� �
x(t) = A0 + �
Ak cos(2�kf 0 t + � k ) = X 0 + Re��
X k e j 2 �kf 0 t �
A1 cos(�1t ) � A2 cos(� 2 t + � ) k=1 � k=1 �

� e j�1 t + e� j�1 t � e j (� 2 t +� ) + e� j (� 2 t +� ) decompose a periodic signal x(t) into a sum of a series of

A1 A2
� 2
� 2
sinusoids - the Fourier series.

4
e e(
A1 A2 j�1 t j (� 2 t +� )
+ e j�1 t e� j (� 2 t +� ) + e� j�1 t e j (� 2 t +� ) + e� j�1 t e� j (� 2 t +� ) )
Note: The sum of periodic functions is periodic.
4
e (
A1 A2 j (�1 t +� 2 t +� ) � j (� 2 t��1 t +� )
+e + e j (� 2 t��1 t +� ) + e� j (�1 t +� 2 t +� ) )
ex.
A1 A2
2
(
cos((�1 + � 2 ) t + � ) + cos((� 2 � �1 ) t + � ) ) �� �8
X k = � 2k 2
k odd
�� 0 k even

f 0 = 25Hz
Composite signals (waveform synthesis) Composite signals (waveform synthesis)

� � � � � �� �
x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re��
X k e j 2 �kf 0 t � x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re��
X k e j 2 �kf 0 t �
k=1 � k=1 � k=1 � k=1 �
�� �8 �� 8 j �� 8
k odd e k odd e j
X k = � 2k 2 X k = � 2k 2 X k = � 2k 2
k odd
�� 0 k even �� 0 k even �� 0 k even
k=1 k=3
x(t) = 0.8105cos(2 25t + ) x(t) = 0.8105cos(2 25t + ) + 0.0901cos(2 75t + )
1 1

1
0.8 0.8 1

0.8
0.6 0.8
0.6

0.6 0.6
0.4 0.4

0.4 0.4
0.2
0.2

0
= 0
0.2 = 0.2

0 0
-0.2
-0.2

-0.2 -0.2
-0.4
-0.4

-0.4 -0.4
-0.6
-0.6
-0.6

-0.6
-0.8
-0.8

-0.8
-0.8
-1
-1 0 0.02 0.04 0.06 0.08 0.1 0.12
0 0.02 0.04 0.06 0.08 0.1 0.12 -1
-1 0 0.02 0.04 0.06 0.08 0.1 0.12
0 0.02 0.04 0.06 0.08 0.1 0.12

Composite signals (waveform synthesis) 0.8

0.6

� �� � spectrum
x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re��
X k e j 2 �kf 0 t �
0.4

0.2

k=1 � k=1 � 0.45

0

�� 8
-0.2

e j k odd 0.4053e� j 0.4053e j

-0.4

X k = � 2k 2 0.4
-0.6

�� 0 k even -0.8

0.35 -1
0 0.02 0.04 0.06 0.08 0.1 0.12

k=5
0.3

x(t) = 0.8105cos(2 25t + ) + 0.0901cos(2 75t + ) + 0.0324 cos(2 125t + ) X 0.25

1 1
0.2
0.8 0.8

0.6 0.6
0.15
0.4 0.4

= 0.2
0.1
0.0450e� j 0.0450e j
0.2

0.0162e� j
-0.2

-0.2 0.05
-0.4
0.0162e j
-0.4

-0.6 0
-0.6 -300 -200 -100 0 100 200 f
300
-0.8 -125 -75 -25 25 75 125
-0.8

x(t) = 0.8105cos(2 25t + ) + 0.0901cos(2 75t + ) + 0.0324 cos(2 125t + ) + ...

-1
0 0.02 0.04 0.06 0.08 0.1 0.12
-1
0 0.02 0.04 0.06 0.08 0.1 0.12
Fourier Series
Fourier Series
x(t) = t 0 � t < T0
j� k
For a given signal, how do we find X k = Ak e � �� �
for each k ? x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re��
X k e j 2 �kf 0 t �
k=1 � k=1 �

Fourier Analysis T0
1 0.04

�
x(t)dt

0.04

� �� � X 0 =

x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re��

X k e j 2 �kf 0 t � T0 0 t 0.02
k=1 � k=1 �
T0 2 T0
where 1 1 t 1 T0 2 T0
� tdt =
0
= = =
0
0 0.01 0.02 0.03 0.04
T0 0 t 0.04
1 T0 T0 2 T0 2 2
X0 =
T0
� x(t)dt f 0:fundamental frequency
0 0

0 T0 = 1/ f 0 Mathematica:
T0
2 athena%add math
� x(t)e
� j 2 kt
Xk = T0
dt athena%math
T0 0 In[1]:=1/T*Integrate[t,{t,0,T}] T0
2
� x(t)e
� j 2 kt
Out[1]:=T/2 Xk = T0
dt
T0 0

Fourier Series Fourier Series

x(t) = t 0 � t < T0
� �� �
x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re�� X k e j 2 �kf 0 t �
k=1 � k=1 � Mathematica:
T0
2
T0
�
te
� j 2 kt

X0 = In[2]:= 2/T*Integrate[t*Exp[-I*2*Pi*k*t/T],{t,0,T}]
Xk = T0
dt
2 T0 0
T0
2
� te
� j 2 kt
Xk = T0
dt
T0 0 (2 I) k Pi
T0 ( j2k + 1) �2 jk T -((-1 + E - (2 I) k Pi) T)
Xk = � 20 2 e�2 jk = (e� j 2 )
e k
2
k 2 2
2 k Out[2]= -----------------------------------­
T0 ( j2k + 1)
(2 I) k Pi 2 2

Xk =
T0
� 2 2 = 1k 2E k Pi
2 2k 2 2 k
=1 e�2 jk = 1
T 2k T T In[3]:= Simplify[%,Element[k,Integers]]
X k = j 0 2 2 + 20 2 � 20 2 e� jk = �1k
2 k 2 k 2 k
T T j IT
Xk = j 0 = 0 e 2 Out[3]= ----
k k Xk = j
T0 T0 j 2
= e
k Pi k k
Fourier Series Fourier Series
x(t) = t 0 � t < T0 x(t) = t 0 � t < T0
� �� � � �� �
x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re�� X k e j 2 �kf 0 t � x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re�� X k e j 2 �kf 0 t �
k=1 � k=1 � k=1 � k=1 �

T0 T0

X0 = X0 =
2 f 0 :fundamental frequency 2
T0 j 2 T0 = 1/ f 0
T0 j
2
Xk = e Xk = e
k k

T0 � T0 T0 � T0
x(t) = + � cos(2�kf 0 t + �
) x(t) = + � cos(2�kf 0 t + �
)
2 k=1 �k 2
2 k=1 �k 2

T0 T0 T T0 T0 T
x(t) = + cos(2f 0 t +
) + 20 cos(2 2 f 0 t + 2 ) + K x(t) = + cos(2f 0 t +
) + 20 cos(2 2 f 0 t + 2 ) + K
2 2
2 2

0.04 0.04 0.04 0.04

f 0 = 25Hz
y t
0.02 T0 = 1/ f 0 = 0.04 y t
0.02 Defined between 0<t<0.04
t t
Periodic with period 0.04
0
7 terms 0
0 0
0 0.01 0.02 0.03 0.04 0.04 0.02 0 0.02 0.04

0
t 0.04 0.04 t 0.04

Fourier Series:Square Wave Fourier Series:Square Wave

T0
T
� 1 0 � t < T0 2
2
1.1 1
2 2 0
� 1e �
� j 2 kt � j 2 kt
Xk = T0
dt + �1e T0 dt
x(t) = � T0 T0 T0
��1 T0 2 � t < T0 z t 0 0 2

1.1 1 In[2]:=2/T*Integrate[Exp[-I*2*Pi*k*t/T],{t,0,T/2}]+
0 0.01 0.02 0.03 0.04

0 t
0.04 2/T*Integrate[- Exp[-I*2*Pi*k*t/T],{t,T/2,T}]
-I k Pi I k Pi
T0
-I (1 - E ) I (-1 + E )
T0
1 2
1 Out[2]= ----------------- + ---------------­
X0 =
T0
� 1dt + T � �1dt T0
k Pi (2 I) k Pi
0 0 2
E k Pi

In[3]:= Simplify[%,Element[k,Integers]] � j (�1�1) j (�2)

2 2

( )
2
� j �1+ (�1)
k
Xk = =�
In[1]:=1/T*Integrate[1,{t,0,T/2}]+ 1/T*Integrate[-1,{t,T/2,T}] Xk = k k
k2
Out[1]:=0 k j4

-I (-1 + (-1) ) =�
Out[7]= ---------------- �� 4 k

�j k odd
X0 = 0 k Pi X k = � k
�� 0 � j (�1+ 1)
2
k even
Xk =
k
� 1 0 � t < T0 2
x(t) = �
��1 T0 2 � t < T0

X0 = 0

�� 4 �� 4 � j 2
�j e k odd
X k = � k
k odd X k = � k
�� 0 �� 0 k even
k even

� �� �
x(t) = A0 + � Ak cos(2�kf 0 t + � k ) = X 0 + Re�� X k e j 2 �kf 0 t �
k=1 � k=1 �
1.2

x(t) = cos(2f 0 t �
4


2 )+ 4
3 cos(2 3 f 0 t �
2 ) +K 0.5

y t
0
z t

0.5

1.2
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04
0 t 0.04