You are on page 1of 32

Interpolation & Decimation

jT
• Sampling period T , z  e at the output

INPUT OUTPUT

• Interpolation by m:
• Let the OUTPUT be Y (z ) [i.e. Samples
exist at all instants nT]
• then INPUT is X ( z ) [i.e. Samples exist
m

at instants mT]
1 Professor A G Constantinides
Interpolation & Decimation
• Let Digital Filter transfer function be H (.)
then Y ( z )  X ( z ).H (.)
m

• Hence H (.) is of the form H (z ) i.e. its


impulse response exists at the instants mT.
• Write
1 2 ( m 1)
H ( z )  h(0)  z .h(1)  h(2).z  ...  h(m  1).z
m ( m 1) ( 2 m 1)
 h(m).z  h(m  1).z  ...  h(2m  1).z
2m ( 2 m 1) ( 3m 1)
 h(2m).z  h(2m  1).z  ...  h(3m  1) z
2 Professor A G Constantinides
Interpolation & Decimation
1 2
• Or H ( z )  H1 ( z m
)  z H 2 ( z m
)  z H 3 ( z m
)  ...
• Where
H1 ( z m )  h(0)  h(m).z m  h(2m).z 2m  ...
m 2 m
H 2 ( z )  h(1)  h(m  1).z  h(2m  1).z
m
 ...
m 2 m
H 3 ( z )  h(2)  h(m  2).z
m
 h(2m  2).z  ...
• So that
1
Y ( z )  H1 ( z ). X ( z )  z H 2 ( z ). X ( z ) 
m m m m

 z 2 H 3 ( z m ). X ( z m )  ...
3 Professor A G Constantinides
Interpolation & Decimation
• Hence the structure may be realised as

H1 ( z m )

INPUT H2 (zm )
+ OUTPUT
H3 ( z m )
Samples across here are
phased
by T secs. i.e. they do not
interact in the adder.
Can be replaced by a
commutator switch.

4 Professor A G Constantinides
Interpolation & Decimation

• Hence
H1 ( z m )
m Commutator
INPUT
H2 (z )

H3 ( z m ) OUTPUT

5 Professor A G Constantinides
Interpolation & Decimation
• Decimation by m:
• Let Input be X (z ) (i.e. Samples exist at
all instants nT)
m
• Let Output be Y ( z ) (i.e. Samples exist at
instants mT)
• With digital filter transfer function H (z )
we have
Y ( z )  X ( z ).H ( z )
m

6 Professor A G Constantinides
Interpolation & Decimation
• Set 1 2
H ( z )  H1 ( z )  z H 2 ( z )  z H 3 ( z )  ...
m m m

( m1)
 ...  z m
.H m ( z )
• And X ( z )  X1 ( z m )  z 1 X 2 ( z m )  z 2 X 3 ( z m )  ...
 ...  z ( m1) X m ( z m )
• Where in both expressions the subsequences
are constructed as earlier. Then

Y ( z m )  H1 ( z m )  z 1H 2 ( z m )  ...  z ( m1) H m ( z m )  
7 X ( z
1
m 1
)  z X 2 ( z )  ...  z
m ( m1)
X (z ) m

mG Constantinides
Professor A
Interpolation & Decimation
• Any products that have powers of z 1 less
m
than m do not contribute to Y ( z ) , as this
is required to be a function of z m .
• Therefore we retain the products
H1 ( z ) X1 ( z )  z m H m ( z m ) X 2 ( z m )
m m

m
z m m
H m1 ( z ) X 3 ( z )...

...  z m H 2 ( z m ) X m ( z m )

8 Professor A G Constantinides
Interpolation & Decimation
• The structure realising this is
Commutator
H1 ( z m )

Hm (zm )
INPUT OUTPUT
H m1 ( z m ) +

H2 (zm )

9 Professor A G Constantinides
Interpolation & Decimation
• For FIR filters why Downsample and then
Upsample?
LOW PASS
fs fs

LENGTH N
#MULT/ACC  N . f s

DOWNSAMPLE M:1 UPSAMPLE 1:M


LOW PASS fs LOW PASS
fs M fs

LENGTH N LENGTH N
N. fs N. fs
#MULT/ACC  M #MULT/ACC  M

2. N . f s
TOTAL #MULT/ACC 
M
10 Professor A G Constantinides
Interpolation & Decimation
• A very useful FIR transfer function special
case is for : N odd, h(n) symmetric
• with additional constraints on h(n) to be
zero at the points shown in the figure.

11 Professor A G Constantinides
Interpolation & Decimation
• For the impulse response shown
1 3 5 7 9
H ( z )  h(0)  h(1).z  h(2).z  h(3).z  h(4).z  h(5).z
 h(1).z  h(2).z 3  h(3).z 5  h(4).z 7  h(5).z 9
• The amplitude response is then given
A( )  h(0)  h(1).2 cos(T )  h(2).2 cos(3T )
 h(3).2 cos(5T )
• In general r  1
A( )  h(0)  2  h 
. cos( rT )
r odd  2 
12 Professor A G Constantinides
Interpolation & Decimation
• Now consider 1, 2 
1  2 
T
• Then
 r  1
A(1 )  h(0)  2  h . cos( rT )
r odd  2 
 r  1    
A(2 )  h(0)  2  h . cos  r   1 T 
r odd  2   T  
 r  1
 h(0)   h . cos( r1T )
r odd  2 

13 Professor A G Constantinides
Interpolation & Decimation
• Hence A(1 )  A(2 )  2h(0)
• Also     r  1   
A   h(0)  2  h . cos  r. .T 
 2T  r odd  2   2T 
 h(0)
  
• Or A(1 )  A(2 )  2 A 
 2T 
• For a normalised response
 
A(0)  1   A   
T 
14 Professor A G Constantinides
Interpolation & Decimation
1
• Thus 2h(0)  1      1 h(0) 
2
• The shifted response
~ 1
A( )  A( ) 
2
is useful

15 Professor A G Constantinides
Design of Decimator and
Interpolator
• Example Develop the specs suitable for the
design of a decimator to reduce the
sampling rate of a signal from 12 kHz to
400 Hz
• The desired down-sampling factor is
therefore M = 30 as shown below

16 Professor A G Constantinides
Multistage Design of
Decimator and Interpolator
• Specifications for the decimation filter H(z)
are assumed to be as follows:
Fp  180 Hz , Fs  200 Hz ,
 p  0.002 ,  s  0.001

17 Professor A G Constantinides
Polyphase Decomposition
The Decomposition
• Consider an arbitrary sequence {x[n]} with
a z-transform X(z) given by

X ( z )  n x[n]z n
• We can rewrite X(z) as
M 1  k
X ( z )  k 0 z X k ( z )
M

where
 n  n
X k ( z )  n xk [n] z  n x[Mn  k ] z
0  k  M 1
18 Professor A G Constantinides
Polyphase Decomposition
• The subsequences {xk [n]} are called the
polyphase components of the parent
sequence {x[n]}
• The functions X k (z ), given by the
z-transforms of {xk [n]}, are called the
polyphase components of X(z)

19 Professor A G Constantinides
Polyphase Decomposition
• The relation between the subsequences {xk [n]}
and the original sequence {x[n]} are given
by
xk [n]  x[Mn  k ], 0  k  M  1
• In matrix form we can write
 X0(zM ) 
 
X ( z )  1  M
z 1 .... z ( M 1)  X1.( z ) 
 .. 
 X . ( z M )
20 Professor
M 1 
A G Constantinides
Polyphase Decomposition
• A multirate structural interpretation of the
polyphase decomposition is given below

21 Professor A G Constantinides
Polyphase Decomposition
• The polyphase decomposition of an FIR
transfer function can be carried out by
inspection
• For example, consider a length-9 FIR
transfer function:
8
n
H ( z)   h[n] z
n 0

22 Professor A G Constantinides
Polyphase Decomposition
• Its 4-branch polyphase decomposition is
given by
1 2 3
H ( z )  E0 ( z )  z E1( z )  z E2 ( z )  z E3 ( z )
4 4 4 4

where
1 2
E0 ( z )  h[0]  h[4]z  h[8]z
E1( z )  h[1]  h[5]z 1
1
E2 ( z )  h[2]  h[6]z
1
E3 ( z )  h[3]  h[7]z
23 Professor A G Constantinides
Polyphase Decomposition
• The polyphase decomposition of an IIR
transfer function H(z) = P(z)/D(z) is not that
straight forward
• One way to arrive at an M-branch polyphase
decomposition of H(z) is to express it in the
M
form P ' ( z ) / D '( z ) by multiplying P(z) and
D(z) with an appropriately chosen
polynomial and then apply an M-branch
polyphase decomposition to P '( z )
24 Professor A G Constantinides
Polyphase Decomposition
• Example - Consider 1 2 z 1
H ( z) 
13 z 1
• To obtain a 2-band polyphase decomposition we
rewrite H(z) as
(1 2 z 1 )(13 z 1 ) 15 z 1  6 z 2 1 6 z 2 5 z 1
H ( z)    
(13 z 1 )(13 z 1 ) 19 z 2 19 z 2 19 z 2
• Therefore,
1
where H ( z )  E0 ( z )  z E1( z )
2 2

1 6 z 1 5
E0 ( z )  1
, E1( z ) 
25 19 z 19 z 1
Professor A G Constantinides
Polyphase Decomposition
• The above approach increases the overall
order and complexity of H(z)
• However, when used in certain multirate
structures, the approach may result in a
more computationally efficient structure
• An alternative more attractive approach is
discussed in the following example

26 Professor A G Constantinides
Polyphase Decomposition
• Example - Consider the transfer function of
a 5-th order Butterworth lowpass filter with
a 3-dB cutoff frequency at 0.5:
0.0527864(1 z 1 )5
H ( z) 
1 0.633436854z  2  0.0557281z  4

• It is easy to show that H(z) can be expressed


as
1  0.105573 z  1 0.52786  z 
2 2
H ( z )     z  2 
2  1 0.105573z 2
 1 0.52786z 
27 Professor A G Constantinides
Polyphase Decomposition
• Therefore H(z) can be expressed as
1
H ( z )  E0 ( z )  z E1( z )
2 2

where
1
1  0.105573 z 
E0 ( z ) 
2  1 0.105573z 1 
1
1  0.52786  z 
E1( z ) 
2  1 0.52786 z 1 

28 Professor A G Constantinides
Polyphase Decomposition
• In the above polyphase decomposition,
branch transfer functions Ei (z ) are stable
allpass functions (proposed by
Constantinides)
• Moreover, the decomposition has not
increased the order of the overall transfer
function H(z)

29 Professor A G Constantinides
FIR Filter Structures Based on
Polyphase Decomposition
• We shall demonstrate later that a parallel
realization of an FIR transfer function H(z)
based on the polyphase decomposition can
often result in computationally efficient
multirate structures
• Consider the M-branch Type I polyphase
decomposition of H(z):
M 1  k
H ( z)   k 0 z Ek ( z M)
30 Professor A G Constantinides
FIR Filter Structures Based on
Polyphase Decomposition
• A direct realization of H(z) based on the
Type I polyphase decomposition is shown
below

31 Professor A G Constantinides
FIR Filter Structures Based on
Polyphase Decomposition
• The transpose of the Type I polyphase FIR
filter structure is indicated below

32 Professor A G Constantinides

You might also like