2.

6 Linear Combinations of Harmonically Related Complex Exponentials

Previously, it was said that a signal x(t) is periodic if there exists some positive value T such that:
x  t   x  t  T  for all t. (2.6)

The smallest number of T which satisfies (2.6) is by definition, the fundamental period, and the fundamental
frequency is defined as 0  2 / T .
2 basic periodic signals were introduced last week, the sinusoidal signal
x  t   cos 0t  (2.7)

and the periodic complex exponential

x  t   e j0t . (2.8)

Both signals have a fundamental frequency, and fundamental period of 0 , and T  2 / 0 respectively.
Associated with the signal in eq. (2.8) is the set of harmonically related complex exponentials
k t   e jk0t  e jk  2 /T t , k  0, 1, 2,... (2.9)

Each of these signals has a fundamental frequency that is a multiple of 0 , and therefore, each is periodic with
period T (although for k  2 , the fundamental period of k  t  is a fraction of T). Thus, a linear combination of
harmonically related complex exponentials of the form
 
x t   
k 
ak e jk0t  a e
k 
k
jk  2 /T t
(2.10)

is also periodic with period T. In eq.(2.10), the term for k  0 is constant. The terms when k  1 , k  1 and are
called the fundamental components or the first harmonic components, and have periods of T (or equivalently,
fundamental frequencies of 0 ). The terms when k  2 and k  2 , are called the second harmonic components,
and have fundamental periods of T/2 (or equivalently, a frequencies of 2 0 ). Generally speaking, the components
corresponding to k   N and k   N are called the Nth harmonic components.

19
 cos(2 t )

cos(4 t )

cos(2 t ) cos(4 t )

Figure 2.16. cos(2 t ) , cos(4 t ) and its sum cos(2 t ) cos(4 t ) .

The representation of a periodic signal in the form of eq.(2.10) is referred to as the Fourier series representation.
Before developing the properties of this representation, let us consider an example.

Example 2.7
Consider a periodic signal x  t  , with fundamental frequency 2 , that is expressed in the form of eq.(2.10), as
3
a0  1,
x t   a e k
jk 2 t
1
, (2.11)
k 3 a1  a1  ,
where 2
1
a0  2, a2  a2  ,
3
1 1
a1  a1  , a3  a3  .
2 4
Rewriting eq.(2.11) and collectingaeach of 1 harmonic components which have the same fundamental frequency,
the
2  a2  ,
we obtain 3

 2 4
 3
 e
4
e  
1 j 2 t1  j 2 t 1 j 4 t  j 4 t 1 j 6 t  j 6 t
x  t   a23 ae3  . e  e e .  (2.12)

Equivalently, using Euler's relation, we can write x(t) in the form

2 1
x  t   2  cos 2 t  cos 4 t  cos 6 t . (2.13)
3 2
In Figure 2.17, we illustrate graphically how the signal x(t) is built up from its harmonic components.

20
 Figure 2.17. Construction of the signal x(t) in Example 2.7 as a linear combination of harmonically related sinusoidal signals.

Example 2.8

Let
 
x  t   1  sin 0t  2cos 0t  cos  20t   .
 4
which has fundamental frequency 0 . As with the previous example, we can again expand x(t) directly in terms of
complex exponentials, so that
x t   1 
1  j0t
e   1 e j  20t  /4   e  j  20t  /4   .
 e  j0t   e j0t  ea0j0t 1,
2j   2 
 1  1
Collecting terms, we obtain a1  1    1  j ,
 1  j  /4  j 2 t  1  j  /4   j 2 t
2j 2
   
  e   e 0   e  
1 1
x  t   1  1  j t
e
0
 1  e
 j t
0
e
0
.
2  1 
a  1    1   j , 
 2j  2j 12
1  
Thus, the Fourier series coefficients for this example are  2j 2
a0  1,
1 j  /4  2
 1  1
a2  e  1  j  ,
a1  1    1  j , 2 4
 2 j  2
a2  e   
1  j  /4 2
1  j  ,
 1  1 2 4
a1  1    1  j ,
 2 j  2 ak  0, k  2.
1 j  /4  2
a2  e  1  j  ,
2 4

a2  e   
1  j  /4 2
1  21
j,
2 4
ak  0, k  2.
In Figure 2.18, we show a bar graph of the magnitude and phase of ak as a function of k. Alternatively, one can plot
the ak as a function of frequency.

Figure 2.18 Plots of the magnitude and phase of the Fourier coefficients of the signal considered in Example 2.8.

We are leading towards something much more general. Actually, any periodic function x(t) of period T can be
expressed as an infinite sum of complex exponentials. This summation is called Fourier series. Fourier series can
be written in several forms. One such form is the exponential Fourier series:

x(t )  ae
n 
n
jn0t

where
T /2
1
ak 
T 
T / 2
x(t )e jk0t dt k  0, 1, 2,... (2.14)

Proof: If

x(t )  ae
n 
n
j 0 nt
,

we can multiply both sides by e jk0t to obtain

 
x(t )e  jk0t  
n 
an e j0 ( n  k )t  ae
n 
n
j 0 ( nk )t
 ak .
nk

Integrate both sides from t  T / 2 to t  T / 2 ,

T /2 T /2  T /2

 x(t )e  jk0t dt   ae 

 0 ( n k )t
n
j
dt  ak dt
T /2 T /2 n  T /2
nk
T /2

1
  an e j0 ( n  k )t  Tak
n  j0 (n  k )
nk T /2

1
  an e j ( n  k )  e  j ( n  k )   Tak
n  j0 (n  k ) 
nk

1
 a  1  (1)  Tak
j0 (n  k )
n
n 
nk

 Tak

22
If x(t) has period T, the term
a1e j0t  a1e j 2 t /T
is referred to as the fundamental frequency and the other terms are known as the harmonics of x(t). Note that the
integration limits in calculating the Fourier coefficients ak doesn’t have to be from –T/2 to T/2: it can be from 0 to T,
T to 2T, -3.5T to -2.5T etc as long as it covers one period of x(t)
Perhaps one of the major applications of the concept of Fourier Series is on sound analysis. In fact, for a given pitch
(or frequency), different instruments sound differently and we are able to distinguish whether it is a tone played by a
piano, guitar, violin, saxophone. Why is that? The following are sample waveforms of a sound with the same pitch
played by different instruments.

Figure 2.19 Waveform of a note played by F) a flute; G) a clarinet; H) an oboe and I) a saxophone.

Graphically, we can tell that even though the pitch (frequency) of the note is the same, the waveform of sound
played by different instruments look differently. Therefore, one would expect that our ear can distinguish which
instrument the note is played by based on the waveform differences. Mathematically, for a periodic waveform

x (t )  ae
n 
n
jn0t

with the same fundamental pitch (frequency) 0 , the values of the Fourier Series coefficients
 a2 , a1 , a0 , a1 , a2 ,  are generally different for each instrument. Therefore, the values of the Fourier Series
coefficients  a2 , a1 , a0 , a1 , a2 ,  relative to each other is what distinguish one instrument from another. In
music, such a feature is called the timbre, or quality of a musical tone.
Example 2.9
The periodic square wave, sketched in Figure 2.20 and defined over one period as
1, t  T1
x t    , (2.15)
 0, T1  t T /2
is a signal that we will encounter a number of times throughout this book. This signal is periodic with fundamental
period T and fundamental frequency 0  2 / T .

Figure 2.20 Periodic square wave with period T.

To determine the Fourier series coefficients for x(t), we can use (2.14). Because of the symmetry of x(t) about t  0 ,
it is convenient to choose T / 2  t  T / 2 as the interval over which the integration is performed, although any

23
interval of length T is equally valid and thus will lead to the same result. Using these limits of integration and
substituting from eq.(2.15), we have first, for k  0 ,
T1 2T1

1
a0  dt  .
T T1 T
As mentioned previously, a0 is interpreted to be the average value of x(t), which in this case equals the fraction of
each period during which x  t   1 . For k  0 , we obtain
T1
T1

1 1
ak  e jk0t dt   e jk0t ,
T T1 jk0T T1
which we may re-write as
2  e jk0T1  e jk0T1 
ak   .
k0T  2j 
Noting that the term in brackets is sin k0T1 , we can express the coefficients ak as
2sin  k0T1  sin  2k T1 / T  2T1 sin  2k T1 / T  2T1
ak     sinc  2kT1 / T  , k  0, (2.16)
k0T k T 2k T1 / T T
where we have used the fact that 0T  2 .
Figure 2.21 is a bar graph of the Fourier series coefficients for this example. In particular, the coefficients are plotted
for a fixed value of T1 and several values of T. For this specific example, the Fourier coefficients are real, and
consequently, they can be depicted graphically with only a single graph. More generally, of course, the Fourier
coefficients ak are complex, so the two graphs corresponding to magnitude ak and phase ak would be required.
The plot ak is called spectral amplitude and represents the amplitude of nth harmonic and the plot ak is called
the spectral phase. Graphic representation of ak along with ak is called complex frequency spectrum. If ak is
purely real or purely imaginary, we can disregard the phase spectrum. The exponential Fourier series find extensive
applications in communication theory. It is equivalent to resolving the function in terms of harmonically related
frequency components of a fundamental frequency.

Lets set T  4T1 i.e. x(t) is a square wave that is unity for half the period and zero for half the period. In this case,
0T1   / 2 , T1 / T  1/ 4 and from eq.(2.16),
1  k  1 sin  k / 2  sin  k / 2 
ak  sinc     , k  0, (2.17)
2  2  2 k / 2 k
while
1
a0  . (2.18)
2
From eq. (2.17), ak  0 for k even and nonzero. Also sin  k / 2  alternates between +1 and -1 for successive odd
values of k. Therefore,
1 1 1
a1  a1  , a3  a3   , a5  a5 
 3 5
The plot of Fourier Series coefficients ak as a function of k or  is known as the spectrum of the signal x  t  .

ak

Figure 2.21. Plots of the Fourier series coefficients ak for the periodic square wave with T  4T1 . The coefficients are regularly spaced
samples of the envelope  2 sin T1  / T at   k0 .

24