VARDHAMAN COLLEGE OF ENGINEERING, HYDERABAD
An autonomous institute, affiliated to JNTUH
Digital Signal Processing (A4415)
Handout
Basic Elements of a Signal Processing System
Analog Signal Processing:
Digital Signal Processing:
Elementary Signals:
Classification of DT Signals
• Energy Signals vs. Power Signals
• Periodic Signals vs. Aperiodic Signals
• Symmetric (even) Signals vs. Antisymmtric (odd) Signals
Energy Signals vs. Power Signals:
E
2
x ( n)
n
N
1
P lim
2
x ( n)
N 2 N 1
n N
Energy Signal EN → E < ∞ P=0
Power Signal EN → ∞ 0<P<∞
but not faster than 2N+1
Neither EN → ∞ P→∞
but faster than 2N+1
Symmetric (Even) vs. Anti-Symmetric (Odd)
A DT signal is symmetric (even symmetric) if x[n]= x[-n]
A DT signal is Anti-symmetric (Odd symmetric) if x[n]= -x[-n]
Any signal x[n] can be broken into a sum of an even part % odd part like this
x[n] = xe [n] + xo [n]
Discrete-Time Systems:
Classification of Systems:
• Static(Memoryless) vs. Dynamic(Memory)
• Linear vs. non-linear
• Time invariant vs. time variant
• Causal vs. non-causal
• Stable vs. unstable
Response of an LTI System:
y[n] x[n]* h[n]
y[n] x[m]h[n m]
m
Discrete Fourier series:
Exponential DFS
N 1
1
x ( n)
N
X ( k )e
k 0
j 2 nk / N
Where
N 1
X (k ) x(n)e j 2 nk / N
n 0
Discrete Fourier Transform
N-point DFT can be defined as
N 1
X (k ) x(n)e j 2 nk / N
n 0
And IDFT is given by
N 1
1
x ( n)
N
X ( k )e
k 0
j 2 nk / N
Properties of DFT:
Periodicity x(n N ) x(n) n
X (k N ) X (k ) k
Linearity a1 x1 (n) a2 x2 (n) a1 X 1 (k ) a2 X 2 (k )
Circular time shift x1 (n)x2 (n) X (k )e j 2 lk / N
Time reversal x ( n) N X ( N k ) X ( k ) N
Circular frequency shift x(n)e j 2 ln/ N X (k l )
x* ( n ) X * ( k ) N
Circular convolution x1 (n)x2 (n) X 1 (k ). X 2 (k )
Fast Fourier Transform
Number of computations using direct evaluation
Additions: N(N-1)
Multiplications: N2
Number of computations using Radix-2 FFT
Additions: N log 2 N
N
Multiplications: log 2 N
2
8-point DIT FFT diagram:
8-point DIF FFT diagram:
IIR Filter Design
1. Butterworth Filters
2. Chebyshev Filters
a. Chebyshev Type – I
b. Chebyshev Type – II
Designing of IIR filters:
1. Direct Method
2. Indirect Method
o Digital specifications → Analog specifications → design analog filter H(s) →
Transform analog filter into digital.
Approaches to design IIR Filter:
1. Impulse Invariant
T
N N
Ci Ci
s p
i 1
i 1 1 e
PT
i
z 1
i
2. Bilinear transformation
2
tan
T 2
2 1 z 1
s
T 1 z 1
Design Procedure of low pass Butterworth filter:
Step 1: Choose the type of transformation i.e., impulse invariant or bilinear transformation
Step 2: Convert digital frequencies into analog and calculate the ratio of edge frequencies.
For Bilinear:
2 2 p
s tan s and p tan
T 2 T 2
For Impulse invariant:
s p
s and p
T T
Step 3: Calculate the order of the filter
1 1
log 2 1 2 1
1 A2 A1
N
2
log s
p
Step 4: Calculate the cut of frequency
p
c 1
1 2N
A2 1
1
Step 5: Determine the analog transfer function
For EVEN N:
N /2
c2
H (s) 2
k 1 s bk c s c
2
For ODD N:
N 1
c 2
c2
H ( s)
s c
k 1 s 2 bk c s c2
Where
2k 1
bk 2sin
2N
Step 6: Using the chosen transformation, transform analog transfer function H(s) into
digital i.e., H(z)
Design Procedure of low pass Chebyshev type-I filter:
Step 1: Choose the type of transformation i.e., impulse invariant or bilinear transformation
Step 2: Convert digital frequencies into analog and calculate the ratio of edge frequencies.
For Bilinear:
2 2 p
s tan s and p tan
T 2 T 2
For Impulse invariant:
s p
s and p
T T
Step 3: Calculate the attenuation constant
1
1 2
2 1
A1
Step 4: Calculate the order of the filter
1 1
1
2
cosh 1 2 1
A2
N
cosh 1 s
p
Step 5: Calculate the cut of frequency
p
c 1
1 2N
A2 1
1
Step 6: Determine the analog transfer function
For EVEN N:
N /2
Bk c2
H (s)
k 1 s 2 bk c s Ck c2
For ODD N:
N 1
B0c 2
Bk c2
H ( s)
s C0c
k 1 s 2 bk c s Ck c2
2k 1
Where bk 2 yN sin
2N
2k 1
Ck yN2 cos 2 and C0 yN
2N
1 1
N
1 1 2 1 1 2 1
1 N 1
y N 2 1 2 1
2
Step 7: Using the chosen transformation, transform analog transfer function H(s) into
digital i.e., H(z)
FIR Filters
Group Delay g :
d
g ( )
d
Phase delay p :
( )
p
Phase Response Linear phase with Condition Midpoint (α)
( ) Both g and p are h(n) h( N 1 n) N 1
constant 2
Only g constant h(n) h( N 1 n) N 1
( )
2 2
Window Functions
S.No Window Function
Name
N 1 N 1
Rectangular 1 n
1 wR (n) 2 2
window 0 else
2n N 1 N 1
Triangular 1 n
2 wT (n) N 1 2 2
window 0
else
2 n N 1 N 1
Raised cosine (1 ) cos n
3 w (n) N 1 2 2
window 0
else
2 n N 1 N 1
Hanning 0.5 0.5cos n
4 wHn (n) N 1 2 2
window 0
else
2 n N 1 N 1
Hamming 0.54 0.46cos n
5 wH (n) N 1 2 2
window 0
else
2 n 4 n N 1 N 1
Blackman 0.42 0.5cos 0.08cos n
6 wB (n) N 1 N 1 2 2
window 0
else
Procedure to design FIR filter:
Step 1: get the desired frequency response H d (e j )
Step 2: calculate hd (n) by applying IDTFT
H e e d
1
hd (n) i i n
2 d
Step 3: Obtain the finite duration impulse response by multiplying with window function
h(n) hd (n).w(n)
Step 4: Find the system function H(z)
N 1
2
H ( z)
N 1
h( n) z n
n
2
Step 5: Convert H(z) into realizable system function
N 1
H ' ( z) z 2
H ( z)
