Complex Signals
Complex Signals
Complex Signals
where
q q
a(t) = x2R (t) + x2I (t) and a(n) = x2R (n) + x2I (n)
xI (t) xI (n)
θ(t) = arctan and θ(n) = arctan . (2.4)
xR (t) xR (n)
xR (t) = a(t) cos(θ(t)) and xR (n) = a(n) cos(θ(n))
xI (t) = a(t) sin(θ(t)) and xI (n) = a(n) sin(θ(n))
The magnitudes a(t) and a(n) are also known as envelopes of x(t) and x(n),
respectively.
45
46 CHAPTER 2. COMPLEX SIGNALS
Im
aejθ
xI a
θ
xR Re
Figure 2.1: Illustration of the relationship between the real and imaginary parts
of the complex number x and its magnitude and phase.
Example 1 on page 59 shows how to use these identities to plot the magnitude
of the spectral density function.
Useful results
Table 2.1 shows some useful identities.
r θ rejθ
1 0 ej0 = 1
1 ±π e±jπ = −1
1 ±nπ e±jnπ = −1 n odd integer
1 ±2π e±j2π = 1 n
1 ±2nπ e±j2nπ = 1 n integer
1 ±π/2 e±jπ/2 = ±j
1 ±nπ/2 e±jnπ/2 = ±j n = 1, 5, 6, 13, ...
1 ±nπ/2 e±jnπ/2 = ∓j n = 3, 7, 11, 15, ...
2.1.2 Phasors
The word phasor is often used by mathematicians to mean any complex number.
In engineering, it is frequently used to denote a complex exponential function of
constant modulus and linear phase, that is a function of pure harmonic behavior.
Here is an example of such a phasor:
which has a constant modulus A and a linearly varying phase. It is not un-
common that the modulus and phase are plotted separately. Different ways to
depict phasors are illustrated in Fig. 2.2.
2.1. INTRODUCTION TO COMPLEX SIGNALS 47
Re{x(t)} Im{x(t)}
A A
a(t) θ(t)
A 2π
A
Im
Re
1/f0
t
A
1/
f
0
where x∗ (t) is the complex conjugate of x(t), and (↔) denotes a Fourier trans-
form pair. Let the complex signal x(t) be expressed in the form:
where x1 (t) and x2 (t) are real signals. Let their spectra be X1 (f ) and X2 (f ),
respectively, i.e. x1 (t) ↔ X1 (f ) and x2 (t) ↔ X2 (f ). The real part of x(t) can
be expressed as1 :
1
x1 (t) = (x(t) + x∗ (t)) . (2.9)
2
Using the linear property of the Fourier transform, we get:
1
G1 (f ) = (G(f ) + G∗ (−f )) (2.10)
2
1 1
g2 (t) = −j (g(t) − g ∗ (t)) ↔ G2 (f ) = −j (G(f ) − G∗ (−f )) . (2.11)
2 2
If one uses the indexes R and I to denote the real and imaginary parts of a
signal, the following simple relations are obtained:
GR (f ) = G1R (f ) − G2I (f )
(2.12)
GI (f ) = G1I (f ) + G2R (f ).
1. Linearity
ax1 (t) + bx2 (t) ↔ aX1 (f ) + bX2 (f ) (2.13)
2. Symmetry
X(t) ↔ x(−f ) (2.14)
3. Scaling
1 f
x(kt) ↔ X (2.15)
|k| k
4. Time reversal
x(−t) ↔ X(−f ) (2.16)
6. Frequency shift
dp x(t) dp X(f )
↔ (j2πf )p X(f ), (−j2πt)p x(t) ↔ , p is a real number
dtp df p
(2.19)
8. Convolution
Parseval’s theorem
Z∞ Z
x1 (t)x∗2 (t)dt = ∞X1 (f )X2∗ (f )df (2.21)
−∞ −∞
Table 2.2: Properties of the Fourier transform for complex analog signals
.
50 CHAPTER 2. COMPLEX SIGNALS
1. Linearity
5. Time shift
7. Differentiation
dp X(f )
(j2πn∆T )p x(n) ↔ (2.26)
df p
8. Convolution
Parseval
fZs /2
∞
X 1
x1 (n)x∗2 (n) = X1 (f )X2∗ (f )df (2.28)
−∞
fs
−fs /2
Table 2.3: Properties of the Fourier transform for complex digital signals.
2.3. ANALYTIC SIGNALS 51
If the filter has a complex impulse response, then the operation corresponds to
4 real filtering operations as shown in Fig. 2.3
An example of an often-used filter with complex impulse response is the filter
given by:
(
1 jm 2π
e Nn 0≤n≤N −1
hm (n) = N (2.30)
0 otherwise
is a function of the parameter m. Figure 2.4 illustrates both the impulse response
and the transfer function of the filter.
0.06 0.06
0.04 0.04
0.02 0.02
h2R(n)
h2I(n)
0 0
−0.02 −0.02
−0.04 −0.04
−0.06 −0.06
0 5 10 15 0 5 10 15
n n
N = 16
1
0.8
|H2(f)|, N=16
0.6
0.4
0.2
0
−0.5 −0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3 0.4 0.5
f/fs
1
arg H2(f)
−1
−2
−3
−0.5 −0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3 0.4 0.5
f/fs
Figure 2.4: Impulse response and transfer function of a complex filter used to
carry out the Discrete-time Fourier transform.
2.3. ANALYTIC SIGNALS 53
f = 0.
54 CHAPTER 2. COMPLEX SIGNALS
0.02 0.02
0.015 0.015
|G0(f)|
|G(f)|
0.01 0.01
0.005 fg 0.005
0 0
−0.5 0 0.5 −0.5 0 f0 0.5
f f
difficulties. Many of them can, however, be circumvented if one applies the fact, that a given
digital signal corresponds to an equivalent analog signal.
2.4. INSTANTANEOUS AMPLITUDE AND FREQUENCY 55
1. Linearity
H{ax1 (t) + bx2 (t)} = ax̃1 (t) + bx̃2 (t), a and b constants (2.50)
2. Time shift
H{x(t + t0 )} = x̃(t + t0 ) (2.51)
3. Applying two times the Hilbert transform gives the original signal
5. Even/odd property
x(t) even ⇔ x̃(t) odd
(2.54)
x(t) odd ⇔ x̃(t) even
6. Conservation of energy
Z ∞ Z ∞
2
x (t)dt = x̃2 (t)dt (2.55)
−∞ −∞
7. Orthogonality Z ∞
x(t)x̃(t) = 0 (2.56)
−∞
8. Modulation
H{x(t) cos(2πf0 t)} = x(t) sin(2πf0 t) (2.57)
if (
X(f ) |f | ≤ F, f0 > F
X(f ) =
0 otherwise
9. Convolution
1
1
0.5
0.8
g0(t)
az(t)
0 0.6
0.4
−0.5
0.2
−1
0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
t t
3
1
2
0.5
1
g0H(t)
ψz(t)
0 0
−1
−0.5
−2
−1
−3
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
t t
1
1
0.5
0.8
g0(t)
az(t)
0 0.6
0.4
−0.5
0.2
−1
0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
t t
3
1
2
0.5
1
g0H(t)
ψz(t)
0 0
−1
−0.5
−2
−1
−3
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
t t
Figure 2.7: Instantaneous amplitude and phase for an analytic signal with a
point of discontinuity.
1
1
0.5
0.8
gR(t)
az(t)
0 0.6
0.4
−0.5
0.2
−1
0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
t t
3
1
2
0.5
1
ψz(t)
gI(t)
0 0
−1
−0.5
−2
−1
−3
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
t t
(2 r)−0.5
0.5
gR(t)
0
−0.5
−1
−3 −2 −1 0 1 2 3
t
r−0.5
0.5
gI(t)
−0.5
−1
−3 −2 −1 0 1 2 3
t
0.5
0
gI(t)
−0.5
−1
0.5
−0.5 3
2
1
−1 0
gR(t) −1
−2
−3
t
The signal is illustrated in Fig. 2.9. The spectrum of the signal is formally
given by:
Z ∞ Z ∞ r
jπrt2 −j2πf t −jπf 2 /r jπr(t−f /r)t2 j −jπf 2 /r
XF M (f ) e e dt = e e dt = e .
−∞ −∞ r
(2.66)
The amplitude spectrum of the signal is constant (frequency independent), and
its phase spectrum varies quadratically with frequency. Notice the similarity
between the signal and its spectrum.
2.4. INSTANTANEOUS AMPLITUDE AND FREQUENCY 59
θ0
fi = = rt. (2.67)
2π
where a(t) is appropriately chosen real signal, does. The spectrum X(f ) of this
signal can be found from:
Z ∞
−jπf 2 /r 2
X(f ) = e a(t)ejπr(f /r−t) dt. (2.69)
−∞
if a(f ) is continuous. Under these conditions, |X(f )| has almost the same shape
os |a(t)| (appropriately scaled though).
Consider the Fourier transform pair:
r
2 j
a(t) ∗ exp(jπrt ) ↔ Xa (f ) · exp(−jπf 2 /r), (2.72)
r
where a(t) ↔ Xa (f ). When the product rT is small, one can use similar con-
siderations as above and show that the signal a(t) ∗ exp(jπrt2 ) has an envelope,
whose shape is given by |Xa (f )|. |Xa (f )| must be appropriately delayed and
scaled.
Figure 2.10 shows an example of the result from the convolution operation
for a rectangular signal, whose duration is varied.
Notice that if a(t) is time-limited, i.e. starts at t = t1 and ends at t = t2 ,
then the a FM-signal with a “start” and “end” frequencies rt1 and rt2 is created.
This can also be achieved by applying the frequency-shift property of the Fourier
transform.
60 CHAPTER 2. COMPLEX SIGNALS
cos(π r t2)
1
g (t) (2 r)−0.5
0
R
−1
2
sin(π r t )
1
r−0.5
g (t)
0
I
−1
−4 −3 −2 −1 0 1 2 3 4
0.015
0.01
g (t)
1
0.005
0
−4 −3 −2 −1 0 1 2 3 4
|g (t) * exp(jπ r t2)|
0.5
1
0
−4 −3 −2 −1 0 1 2 3 4
t
−3
x 10
4
g2(t)
0
−4 −3 −2 −1 0 1 2 3 4
|g2(t) * exp(jπ r t2)|
0.6
0.4
0.2
0
−4 −3 −3 −2 −1 0 1 2 3 4
x 10
3
g3(t)
0
−4 −3 −2 −1 0 1 2 3 4
|g3(t) * exp(jπ r t2)|
0.4
0.2
0
−4 −3 −2 −1 0 1 2 3 4
t
Solution
Using the rules of linearity and time shift we obtain:
2
|H(f)| (V/Hz)
1.5
0.5
0
−0.5 0 0.5
f (oscillations/second)