Chapter 3

Linear Block Codes

3.1 (n, k) Linear Block Codes over GF(q)

Let the message m = ( m0 , m1 ,L , m k - 1 ) be an arbitrary k-tuple

from GF(q).

k
The linear (n, k) code over GF(q) is the set of q codeword of

row-vector form c = ( c0 , c 1 ,L , c n- 1 ) , where c j ∈ GF(q)

By linear transformation

k −1
c = m ⋅ G = ∑ mi ⋅ g i =m0 g0 + m1 g0 + L + mk - 1 g k - 1
i =0

Here G is a k × n matrix of rank k of elements from GF(q),

g i is the i-th row vector of G.

G is called a generator matrix of the code.

The rows of G are linearly independent since G is assumed to

have rank k.

The code C is called a k-dimensional subspace of the set of all

n-tuples.
Example:

(7, 4) Hamming code over GF(2)

The encoding equation for this code is given by

c0 = m0
c1 = m1
c2 = m2
c3 = m3
c4 = m0 + m 1 + m 2
c5 = m1 + m 2 + m 3
c6 = m0 + m 1 + m 3

that is,

1 0 0 0 1 0 1
0 1 0 0 1 1 1
G= 
0 0 1 0 1 1 0
 
0 0 0 1 0 1 1

An (n, k) block code is said to be linear if the vector sum of two

codeword is a codeword.

Linear Systematic Block Code:

In systematic from the codeword C is comprised of an

information segment and a set of n-k symbols that are linear

combinations of certain information symbols, determined by the

P matrix. That is
 c i = mi ; for 0 ≤ i < k

k −1
c i = ∑ m j p j, n- k - i ; for k ≤ i < n
j =0

message codeword

( m0 , m1 ,L , mk - 1 ) ↔ ( m0 , m1 ,L , mk -1 , c k , c k + 1 ,L , c n-1 )

The second set of equations, given above, is called the set of

parity-check equations.

An (n, k) linear systematic code is completely specified by a k × n

generator matrix of the following form

 g0 
 g 
G =  1  = [I k P ]
 M 
 
 g k -1 
where I k is the k × k identity matrix

 p0, ( n- k - 1) p0, ( n- k - 2 ) L p0, 0 

 p p1, ( n- k - 2 ) L p1, 0 
P= 
1, ( n - k - 1 )

 M M O M 
 
 p( k - 1 ) , ( n - k - 1 ) p( k - 1 ) , ( n - k - 2 ) L p( k - 1 ) , 0 

P -matrix is a k × ( n - k ) matrix.

Parity-check matrix

An (n, k) linear code can also be specified by an ( n - k ) × k

matrix H.

Let c = ( c0 , c 1 ,L , c n- 1 ) be an n-tuple

then c is a codeword if and only if

c ⋅ H T = (0,0,L ,0 )
14243
n− k

i.e. the inner product of c and each row of H is zero.

The matrix H is called a parity-check matrix.

Since G = [I k P ]

[
we can see that H = P I n- k
T
]
where P T is the transpose of P

and G ⋅ H T = 0 .

Note: For any given generator matrix G, many solution for H are

possible.
Example:

A (6, 3) code is generated by

 1 0 0 1 1 1
G = 0 1 0 1 0 1
 
0 0 1 0 1 1

The parity-check matrix is given by

1 1 0 1 0 0 
H = 1 0 1 0 1 0 
 
1 1 1 0 0 1

A code generated by H is called the dual code of the code

generated by G.

A dual code is denoted as C ⊥ .

3.2 Hamming Distance of Linear Block Code and
Error Protection Properties

Distance between two n-symbol vectors

u = ( u0 , u1 ,L , un- 1 )

v = ( v0 , v 1 ,L , v n- 1 )

(a) Euclidean distance

n-1
d E (u , v ) = ∑(u
i =0
i - vi ) 2

(b) Hamming distance

d H ( u , v ) = { i | ui ≠ v i , i = 0,1,L , n - 1 }

i.e. the number of places where u and v differ.

Hamming weight and Hamming distance of codewords

(a) For a linear code C, the Hamming distance between any two

codewords is simply described by

d H ( c 1 , c 2 ) = wt ( c 1 - c 2 ) = wt ( c 3 )

where c 3 is the difference between c 1 and c 2 .

wt ( c 3 ) is the Hamming weights of c 3 , or the number of

nonzero positions of c 3 .
(b) Triangle inequality

For codeword a , b and c

d H (a , c ) + d H (c , b ) ≥ d H (a , b )

a b

(c) d H ( a , b ) = wt ( a + b )
3.3 Minimum distance of a Block code
Let C be a linear block code. The minimum distance of C,

denoted as d min , is defined as follows:

d min ≡ min {d ( v , u ) : v , u ∈ C, v ≠ u }

The minimum weight of C, denoted as wmin , is defined as

follows:

wmin ≡ min { w( v ) : v ∈ C, v ≠ 0 }

Exercise:

Show that d min = wmin

Proof:

d min ≡ min {d ( v , u ) : v , u ∈ C, v ≠ u }
= min { d ( v + u ) : v , u ∈ C, v ≠ u }
= min { w( x ) : x ∈ C, x ≠ 0 }
= wmin
3.4 maximum Error-Correction Capability of a Block
Code

Suppose that c0 is selected for transmission and that the closest

codeword is d min in Hamming distance, as shown below (Fig. 3.5

page 87)



  
c1


  
 
c0 



   
c0
  



   

 c2
  
c1

c r =c +e

e = ( e0 , e 1 ,L , e n- 1 ) : error pattern.

c = ( c0 , c 1 ,L , c n- 1 ) : codeword transmitted.

r = ( r0 , r1 ,L , rn- 1 ) : received word.

  d min − 1 
If the channel-error pattern e has t =   or fewer errors,
 2 

one is guaranteed that r = c0 + e remains closer in Hamming

distance to c0 than to any other codeword and thus is decoded

correctly.

 d − 1
As a consequence, t =  min  is called the maximum
 2 

error-correction capability of the code.

Error-detection Capability

Suppose that the decoder’s task is only to detect the presence of

errors, and if errors are detected, to label the codeword (received

word) is unreliable.

The detector’s function can fail only if e takes the transmitted

codeword c0 into another codeword c 1 , that is c0 + e = c 1 .

This cannot occur if there are d min − 1 or fewer errors in the n

positions of the code.

That is, d min − 1 is the guaranteed error detection capability of

the code.

Hybrid modes of error control

One can correct t errors and still detect up to t d errors provided

that t + t d < d min .

3.5 Weight Distribution

Let C be an (n, k) linear block code and wmin denotes the

number of codewords in C with Hamming weight i .

Define W ( z ) = ∑ wi z as the weight enumerator polynomial.

i

i =0

Clearly, w0 = 1

w0 + w1 + L + wn = 2 k

Exercise:

Find the weight enumerator polynomial of the (7, 4) Hamming

code generated by

1 0 0 0 1 0 1
0 1 0 0 1 1 1
G= 
0 0 1 0 1 1 0
 
0 0 0 1 0 1 1

Answer: W ( z ) = 1 + 7z 3 + 7z 4 + z 7
3.6 Some Commonly-used Modifications of Linear
Codes

Shortened code

A code is shortened by deleting some message symbols (bits) from

the encoding process.

For example, by deleting one message symbol (or bit), an (n, k)

code becomes an (n-1, k-1) code.

Extended code

A code is extended by adding some additional redundant symbols

(or bits).

For example, by adding one parity symbol, a (n, k) code becomes

a (n+1, k) code.

In general, the error-control capability of the extended code can

be increased.

Punctured Code

A code is punctured by deleting some of its parity symbols (or

bits).

For example, by deleting one parity symbol, a (n, k) code

becomes (n-1, k) code.

In general, the error-control capability is reduced, but the code

rate is increased.