Processing
Multirate Digital Signal
Processing
What is multirate signal processing?
Processing of digital signal with
different sampling rates in the system.
Sampling Rate Conversion
Multirate Digital Signal
Processing
Upsampler  Used to increase
the sampling rate by an integer
factor
Downsampler  Used to decrease
the sampling rate by an integer
factor
Basic Sampling Rate Alteration Devices
Why sample rate conversion? (I)
Compatibility: convert sample frequencies of
different stds.
Efficiency: easier data processing
(computationally more efficient), less storage,
lower transmission speed,
Alldigital: Change sample frequency in an
efficient manner
Cost: Avoid need for expensive analogue anti
aliasing filters
Multirate Digital Signal
Processing
UpSampler
TimeDomain Characterization
An upsampler with an upsampling
factor L, where L is a positive integer,
develops an output sequence with
a sampling rate that is L times larger
than that of the input sequence x[n]
Blockdiagram representation
] [n x
u
L x[n]
] [n x
u
UpSampler
Upsampling operation is implemented by
inserting equidistant zerovalued
samples between two consecutive
samples of x[n]
Inputoutput relation
1 L
=
=
otherwise , 0
, 2 , , 0 ], / [
] [
L L n L n x
n x
u
UpSampler
In practice, the zerovalued samples
inserted by the upsampler are replaced
with appropriate nonzero values using
some type of filtering process
Process is called interpolation and will be
discussed later
DownSampler
TimeDomain Characterization
An downsampler with a downsampling
factor M, where M is a positive integer,
develops an output sequence y[n] with a
sampling rate that is (1/M)th of that of
the input sequence x[n]
Blockdiagram representation
M x[n]
y[n]
DownSampler
Downsampling operation is implemented
by keeping every Mth sample of x[n] and
removing inbetween samples to
generate y[n]
Inputoutput relation
y[n] = x[nM]
1 M
DownSampler
Figure below shows explicitly the time
dimensions for the downsampler
M
) ( ] [ nMT x n y
a
=
) ( ] [ nT x n x
a
=
Input sampling frequency
T
F
T
1
=
Output sampling frequency
'
1
'
T M
F
F
T
T
= =
UpSampler
Figure below shows explicitly the time
dimensions for the upsampler
Input sampling frequency
T
F
T
1
=
=
=
otherwise 0
, 2 , , 0 ), / ( L L n L nT x
a
L
) ( ] [ nT x n x
a
=
y[n]
Output sampling frequency
'
1
'
T
LF F
T T
= =
Basic Sampling Rate Alteration Devices
The upsampler and the downsampler are
linear but timevarying discretetime systems
Consider a factorofM downsampler defined
by
Its output for an input is
then given by
From the inputoutput relation of the down
sampler we obtain
y[n] = x[nM]
] [
1
n y ] [ ] [
0 1
n n x n x =
] [ ] [ ] [
0 1 1
n Mn x Mn x n y = =
)] ( [ ] [
0 0
n n M x n n y =
] [ ] [
1 0
n y Mn Mn x = =
UpSampler
FrequencyDomain Characterization
Consider first a factorof2 upsampler
whose inputoutput relation in the time
domain is given by
=
=
otherwise ,
, , , ], / [
] [
0
4 2 0 2 n n x
n x
u
UpSampler
In terms of the ztransform, the input
output relation is then given by
=
=
= =
even
] / [ ] [ ) (
n
n
n
n
n
u u
z n x z n x z X 2
2 2
[ ] ( )
m
m
x m z X z
=
= =
UpSampler
In a similar manner, we can show that
for a factorofL upsampler
On the unit circle, for , the input
output relation is given by
) ( ) (
L
u
z X z X =
e j
e z =
) ( ) (
L j j
u
e X e X
e e
=
UpSampler
Figure below shows the relation between
and for L = 2 in the
case of a typical sequence x[n]
) (
e j
e X
) (
e j
u
e X
UpSampler
As can be seen, a factorof2 sampling
rate expansion leads to a compression
of by a factor of 2 and a 2fold
repetition in the baseband [0, 2t]
This process is called imaging as we
get an additional image of the input
spectrum
) (
e j
e X
UpSampler
Similarly in the case of a factorofL
sampling rate expansion, there will be
additional images of the input spectrum in
the baseband
Lowpass filtering of removes the
images and in effect fills in the zero
valued samples in with interpolated
sample values
1 L
1 L
] [n x
u
] [n x
u
DownSampler
FrequencyDomain Characterization
Applying the ztransform to the inputoutput
relation of a factorofM downsampler
we get
The expression on the righthand side cannot
be directly expressed in terms of X(z)
=
=
n
n
z Mn x z Y ] [ ) (
] [ ] [ Mn x n y =
DownSampler
To get around this problem, define a
new sequence :
Then
=
=
otherwise ,
, , , ], [
] [
int
0
2 0 M M n n x
n x
] [
int
n x
=
=
= =
n
n
n
n
z Mn x z Mn x z Y ] [ ] [ ) (
int
) ( ] [
/
int
/
int
M
k
M k
z X z k x
1
= =
=
DownSampler
Now, can be formally related to x[n]
through
where periodic train c[n]
A convenient representation of c[n] is given
by
where
] [
int
n x
] [ ] [ ] [
int
n x n c n x =
=
=
otherwise ,
, , , ,
] [
0
2 0 1 M M n
n c
=
=
1
0
1
M
k
kn
M
W
M
n c ] [
M j
M
e W
/ t 2
=
DownSampler
Taking the ztransform of
and making use of
we arrive at
] [ ] [ ] [
int
n x n c n x =
=
=
1
0
1
M
k
kn
M
W
M
n c ] [
n
n
M
k
kn
M
n
n
z n x W
M
z n x n c z X
=


.

\

= = ] [ ] [ ] [ ) (
int
1
0
1
( )
=
=
=


.

\

=
1
0
1
0
1 1
M
k
k
M
M
k n
n kn
M
W z X
M
z W n x
M
] [
DownSampler
Consider a factorof2 downsampler
with an input x[n] whose spectrum is as
shown below
The DTFTs of the output and the input
sequences of this downsampler are
then related as
)} ( ) ( {
2
1
) (
2 / 2 / e e e
+ =
j j j
e X e X e Y
DownSampler
Now implying
that the second term in the
previous equation is simply obtained by
shifting the first term to the right
by an amount 2t as shown below
) ( ) (
2 / ) 2 ( 2 / t e e
=
j j
e X e X
) (
2 / e
j
e X
) (
2 / e j
e X
DownSampler
The plots of the two terms have an overlap,
and hence, in general, the original shape
of is lost when x[n] is downsampled
as indicated below
) (
e j
e X
DownSampler
This overlap causes the aliasing that takes
place due to undersampling
There is no overlap, i.e., no aliasing, only if
Note: is indeed periodic with a
period 2t, even though the stretched
version of is periodic with a period
4t
2 / 0 ) ( t > e =
e
for
j
e X
) (
e j
e X
) (
e j
e Y
DownSampler
For the general case, the relation between
the DTFTs of the output and the input of a
factorofM downsampler is given by
is a sum of M uniformly
shifted and stretched versions of
and scaled by a factor of 1/M
=
t e e
=
1
0
/ ) 2 (
) (
1
) (
M
k
M k j j
e X
M
e Y
) (
e j
e Y
) (
e j
e X
DownSampler
Aliasing is absent if and only if
as shown below for M = 2
2 / for 0 ) ( t > e =
e j
e X
M for e X
j
/ 0 ) ( t > e =
e
Cascade Equivalences
Two cascade equivalences are shown
below
L ] [n x
] [
2
n y
) (
L
z H
L
] [n x
] [
2
n y
) (z H
M ] [n x
] [
1
n y
) (z H
M
] [n x
) (
M
z H
] [
1
n y
Cascade equivalence #1
Cascade equivalence #2
Filters in Sampling Rate
Alteration Systems
The bandwidth of a critically sampled
signal must be reduced by lowpass
filtering before its sampling rate is
reduced by a downsampler to avoid
aliasing
Likewise, the zerovalued samples
introduced by an upsampler must be
interpolated by lowpass filtering to more
appropriate values for an effective
sampling rate increase
Filter Specifications
Since upsampling causes periodic
repetition of the basic spectrum, the
unwanted images in the spectra of the up
sampled signal must be removed by
using a lowpass filter H(z), called the
interpolation filter, as indicated below
The above system is called an interpolator
] [n x
u
L ] [n x
] [n y
) (z H
] [n x
u
Filter Specifications
On the other hand, prior to down
sampling, the signal v[n] should be
bandlimited to by means
of a lowpass filter, called the decimation
filter, as indicated below to avoid aliasing
caused by downsampling
The above system is called a decimator
M / t e <
M
] [n x ) (z H
] [n y
Interpolation Filter
Specifications
If we pass x[n] through a factorofL up
sampler generating , the relation
between the Fourier transforms of x[n] and
are given by
It therefore follows that if is passed
through an ideal lowpass filter H(z) with a
cutoff at t/L and a gain of L, the output of
the filter will be precisely y[n]
] [n x
u
] [n x
u
) ( ) (
L j j
u
e X e X
e e
=
] [n x
u
Interpolation Filter
Specifications
If is the highest frequency that needs
to be preserved in x[n], then
Summarizing the specifications of the
lowpass interpolation filter are thus given
by
c
e
L
c p
/ e e =
s s
s
=
t e t
e e
e
L
L L
e H
c
j
/ ,
/ ,
) (
0
Decimation Filter Specifications
In a similar manner, we can develop the
specifications for the lowpass decimation
filter that are given by
The design of the filter H(z) is a standard
IIR or FIR lowpass filter design problem
s s
s
=
t e t
e e
e
M
M
e H
c
j
/ ,
/ ,
) (
0
1
The FIR filter is realized using direct form
To avoid unnecessary calculations the decimator
is replaced with efficient transversal structure.
For the polyphase structure
] [ * ] [ ) (
1
0
n x n p n y
m m
M
m
=
=
Polyphase Decomposition
] [ ] [ ) (
1
0
n x m h n y
m
N
m
=
=
Polyphase Decomposition
Decomposition of H(z)=hm z
m
in blocks of M:
H (z) = ... + h(M )zM + h(M + 1) z M 1 + ... + h(1) z1
+ h(0)z0 + h(1) z1 + ... + h(M 1) z(M 1)
+ h(M )zM + h(M + 1) z(M +1) + ... + h(2M 1) z(2M 1)
+ h(2M )z2M + h(2M + 1) z(2M +1) + ... + h(3M 1) z(3M 1) + ...
= z0[... + h(0) z0 + h(M ) zM + ...] + z1[... + h(1) + h(M + 1) zM + ...]
+ z2[... + h(2) + h(M + 2) zM + ...] + ...
+ z(M 1) [... + h(M 1) + h(2M 1) zM + ...]
H (z) = z Pi z
i
i=0
M 1
( )
M
where Pi (z) =
n=
z h(nM + i)
n
+
Polyphase Decomposition
Implementation of Decimation
Using noble identity:
Operations performed at Operations at low rate
high rate more efficient
Using commutator:
Implementation of Decimation
one input per D pulses;
counterclockwise rotation
THANK YOU