Digital Filter Structures

N M
y n ak y n k bk x n k
k 1 k 0

M
v[n] bk x n k
k 0

N
y n ak y n k v[n]
k 1

Direct Form I implementation

On the z-domain

1 M

H ( z ) H 2 ( z ) H1 ( z ) ( N
) b z k
k

1 ak z k k 0
k 1

or equivalently
M

V ( z ) H1 ( z ) X ( z )

b z
k 0
k
k
X ( z)

1
Y ( z ) H 2 ( z )V ( z ) ( N
)V ( z )
1 ak z k
k 1
By changing the order of H1 and H2, onsider the equivalence on
the z-domain:
H ( z ) H1 ( z ) H 2 ( z )

1
k
M
where H 2 ( z) (
H1 ( z )

bk z
k 0
N
1 ak z k
)

k 1
Let
1
W ( z) H 2 ( z) X ( z) ( N
) X ( z)
1 ak z k
k 1
Then
M
k
Y ( z ) H1 ( z )W ( z )
k 0
bk z W ( z)

 N
In the time domain, w[n] ak w[n k ] x[n]
k 1
M
y[n] bk w[n k ]
k 0
We have the following equivalence for implementation:

We assume M=N
here
Note that the exactly the same signal, w[k], is stored in the two
chains of delay elements in the block diagram. The
implementation can be further simplified as follows:

Direct Form II (or

Canonic Direct Form)
implementation
By using the direct form II implementation, the number of
delay elements is reduced from (M+N) to max(M,N).

Example: 1 2 z 1
H ( z)
1 1.5 z 1 0.9 z 2

Direct form I implementation Direct form II implementation

Representing by signal-flow graph

Example: the signal-flow graph of direct form II.

Cascade Form

If we factor the numerator and denominator polynomials, we

can express H(z) in the form:
M1 M2
1 1 1
(1
k k
f z ) (1 g z )(1 g kz )
H ( z) k 1
N1
k 1
N2
1 1 1
(1
k k
c z ) (1 d z )(1 d kz )
k 1 k 1

where M=M1+2M2 and N=N1+2N2, gk and gk* are a complex

conjugate pair of zeros, and ck and ck* are a complex conjugate
pair of poles.

It is because that any N-th order real-coefficient polynomial

equation has n roots, and these roots are either real or complex
conjugate pairs.
 b0 k b1k z 1 b2 k z 2
Ns
A general form is H ( z ) 1 2
k 1 1 a 1k z a 2k z
where N s ( N 1) / 2
and we assume that MN. The real poles and zeros have been
combined in pairs. If there are an odd number of zeros, one of
the coefficients b2k will be zero.
It suggests that a difference equation can be implemented via
the following structure consisting of a cascade of second-order
and first-order systems:

cascade form of implementation (with a direct form II realization of

each second-order subsystems)
Example:

1 2 z 1 z 2 (1 z 1 )(1 z 1 )
H ( z) 1 2

1 0.75 z 0.125 z (1 0.5 z 1 )(1 0.25 z 1 )

Cascade structure: direct form I implementation

Cascade structure: direct form II implementation

Parallel Form

If we represent H(z) by additions of low-order rational

systems: Np 1
N 1 N
Ak 2
Bk (1 ek z )
H ( z ) Ck z k 1
1 1
k 0 k 1 1 c k z k 1 (1 d k z )(1 d kz )

where N=N1+2N2. If MN, then Np = MN.

Alternatively, we may
N p group the Nreal poles in pairs,
1 so that
s
e0 k e1k z
H ( z ) Ck z k
1 2
k 0 k 1 1 a1k z a 2k z
Illustration of parallel-form
structure for six-order
system (M=N=6) with the
real and complex poles
grouped in pairs.
Example: consider still the same system

1 2 z 1 z 2 7 8 z 1
H ( z) 1 2
8
1 0.75 z 0.125 z 1 0.75 z 1 0.125 z 2
another alternation of the same system

1 2 z 1 z 2 18 25
H ( z) 1 2
8 1

1 0.75 z 0.125 z 1 0.5 z 1 0.25 z 1

Hence, given a system

function, there are many
ways to implement it.

There are equivalent when

infinite-precision arithmetic
is used. However, their
behavior with finite-
precision arithmetic can be
quite different.
While the signal flow graph is an efficient way to
represent a difference equation, not all of its instances
are realizable:
If a system function has poles, a corresponding signal
flow graph will have feedback loops.
A signal flow graph is computable if all loops contain at
least one unit delay element. Eg.

A non-computable system

Computable systems
 Circular Convolution (for DFT)
Time-domain convolution implies frequency domain
multiplication. This property is valid for continuous Fourier
transform, Fourier series, and DTFT, but is not exactly true for
DFT.
The DFT pair considered hereafter (following Openheims
book, where the 1/N is put on the inverse-transform side):
N 1
X [k ] x[n]WNkn , k 0,1..., N 1
n 0
N 1
1
x[n]
N
X [ k
k 0
]W kn
N , n 0,1..., N 1

where WN = ej2/N is a root of the equation WN=1.

 Circular Convolution (for DFT)
For DFT, time domain circular convolution implies frequency
domain multiplication, and vice versa.
Consider a periodic sequence. Its DTFT is both periodic
and discrete in frequency. Multiplication in the frequency
domain results in a convolution of the two corresponding
periodic sequences in the time domain. Now lets consider a
single period of the resulted sequence. Since the two
sequences are both periodic, the convolution appears as
folding the rear of a sequence to the front one by one, and
superimposing the inner products so obtained, in a single
period.
Convolution of two periodic sequences
Circular convolution (definition)
x3 [n] x1[n] x2 [n] x2 [n] x1[n]
N 1 N 1
x2 [m]x1[((n m)) N ] x2 [m]x1[(n m) mod N ]
m 0 m0

circular convolution

Symbol for representing circular convolution: or N .

If the DFT of x1[n], x2[n], and x3[n] are X1[k], X2[k], and X3[k],
respectively.
Time domain circular convolution implies frequency domain
multiplication: x [ n] x [ n] x [n] X [ k ] X [ k ] X [ k ]
3 1 2 3 1 2

Time domain multiplication implies frequency domain circular

convolution (with 1/N amplitude reduction): 1
x3 [n] x1[n]x2 [n] X 3 [k ] X 1[ k ] X 2 [ k ]
N
Example: circular
convolution with a delayed
impulse sequence
x1[n] [n n0 ], 0 n0 N
Example: circular convolution of two rectangular pulses

N-point circular convolution of two sequences of length N.

Example: circular convolution of two rectangular pulses
(continue)

Given two sequences of length L, assume that we add L zeros

on its end, making an N=2L point sequence referred to as
zero padding

N-point circular convolution of two sequences of length L, where N=2L.

N-point circular convolution of two sequences of length L, where N=2L (continue).
Note that by zero padding, we can use circular convolution to
compute convolution of two finite length sequences.
Some other properties involving circulation:

Time domain circular shift implies frequency domain phase

shift:
x[((n m)) N ], 0 n N 1 e j ( 2k / N ) m X [k ] W km X [k ]

Duality property of DFT:

Since DFT and IDFT has very similar form, we have a duality
property for DFT:
DFT
If x[n] X [k ]

DFT
Then X [n] Nx[((k )) N ], 0 k N 1
DFT Properties:
Sampling the DTFT spectrum
We have seen that, for an M-point sequence, if we uniformly
sample M points in its DTFT spectrum within [0, 2], we can
equivalently obtain its DFT.
What happen when we sample N points in the frequency domain
[0, 2], instead of M points? Let the samples be
~
X [k ] X ( z ) z e j ( 2 / N ) k X (e j ( 2 / N ) k ), k 0,..., N 1

We will have the following property:

x [n] x[n] [n rN ]
~
x[n rN ]
r r
which means that frequency-domain sampling implies time domain
aliasing.
when N M, no aliasing will happen, and particular when N=M,
we obtain DFT.
when N < M, aliasing (of nonzero values) happens.
 Linear Convolution Using DFT
Recall that linear convolution is

x3[n] x [m]x [n m]
m
1 2

when the lengths of x1[n] and x2[n] are L and P, respectively the
length of x3[n] is L+P-1.
Thus a useful property is that the circular convolution of two
finite-length sequences (with lengths being L and P respectively)
is equivalent to linear convolution of the two N-point (N L+P1)
sequences obtained by zero padding.
Another useful property is that we can perform circular
convolution and see how many points remain the same as those of
linear convolution. When P < L and an L-point circular convolution is
performed, the first (P1) points are corrupted by circulation, and
the remaining points from n=p1 to n=L1 (ie. The last LP+1
points) are not corrupted (ie., the last LP+1 points remain the
same as the linear convolution result).
Block convolution (for implementing an FIR filter)

FIR filtering is equal to the linear convolution of a (possibly)

infinite-length sequence.
To avoid delay in processing, and also to make efficient
computation, we would like to segment the signal into sections of
length L. Each L-length sequence can then be convolved with the
finite-length impulse response and the filtered sections fitted
together in an appropriate way. called block convolution.

When each section is sufficiently large, we usually use circular

convolution (instead of linear convolution) to compute each
section. (since it will be shown that there are fast algorithms,
fast Fourier transform (FFT), to compute circular convolutions
highly efficiently)
Two methods for circular-convolution-based block convolution:
Overlapping-add method and overlapping-save method.
Overlapping-add method (for implementing an FIR filter)
When segmenting into L-length segments, the signal x[n] can be
represented as
x[n] x [n rL]
m
r

where x[n rL] 0 n L 1

xr [ n ]
0 otherwise
Because convolution is an LTI operation, it follows that

y[n] x[n] h[n] y [n rL]
m
r

where yr [n] xr [n] h[n]

Since xr[n] is of length L and h[n] is of length P, each yr[n] has
length (L+P1).
So, we can use zero-padding to form two N point sequences,
N=L+P1, for both xr[n] and h[n]. Performing N-point circular
convolution (instead of linear convolution) to compute yr[n].
For example, consider two sequences h[n] and x[n] as follows.
Segmenting x[n] into L-
length sequences.
Each segment is
padded by P1 zero
values.

Fir filtering by using the

overlapping-add
method.
Overlapping-save method (for implementing an FIR filter)
Can we perform L-point circular convolution, instead of (L+P1)-
point circular convolution?
If a P-point sequence is circularly convolved with a P-point
sequence (P<L), the first (P1) points of the result are
incorrect, while the remaining points are identical to those
that would be obtained by linear convolution.

Separating x[n] as overlapping sections of length L, so that

each section overlaps the preceding section by (P1) points.
xr [n] x[n r ( L P 1) P 1], 0 n L 1

Then y[n] yr [n r ( L P 1) P 1]
r 0
Example of overlapping-save method

Decompose x[n] into

overlapping sections
of length L
Example of overlapping-save method (continue)

Result of circularly convolving

each section with h[n]. The
portions of each filter section
to be discarded in forming the
linear convolution are indicated
See the following reference for the suggestion of L and P:
M. Borgerding, Turning Overlap-Save into a Multiband Mixing, Downsampling
Filter Bank, IEEE Signal Processing Magazine, pp. 158-161, 2006.

Fast Fourier Transform (FFT)

N 1
DFT pairs: X [k ] x[n]W , kn
N k 0,1..., N 1
n 0
N 1
1
x[n]
N
X
k 0
[ k ]W kn
N , n 0,1..., N 1

WN = ej2/N is a root of the equation WN=1.

It requires N2 complex multiplications and (N1)N complex
additions for computation.
Each complex multiplication needs four real multiplications and
two real additions, and each complex addition requires two real
additions. It requires 4N2 real multiplications and N(4N2) real
additions.
Goertzel algorithm
N 1
Since W kN
N 1 X [k ] WN kN x[r ]WNkr
r 0
N 1
x[r ]WN k ( N r )
r 0

Define
y k [ n] x[ r
r
]W k ( nr )
N u[n r ]

The above equation can be interpreted as a discrete convolution

of the finite-duration sequence x[n], 0 n N1, with the
kn
sequence N u[n] , which is the impulse response of the LTI
W
system.
Note that x[r] is nonzero only when 0 r < N. We can easily
verify that
X [ k ] y k [ n] n N
The Goertzel algorithm computes DFT by implementing the above
kn
LTI system. The system function is the z transform of WN u[n]

1
H k ( z)
1 WN k z 1

The signal-flow graph of the LTI system for obtaining yk[n] is

The implementation can be further simplified as
1 1 WNk z 1
H k ( z) k 1

1 WN z (1 WN k z 1 )(1 WNk z 1 )
1 WNk z 1

1 2 cos(2k / N ) z 1 z 2
vk[n]

Flow graph of second-order computation of X[k] (Goertzel algorithm)

Since we only need to bring the system to a state from which
yk[n] can be computed, the complex multiplication by Wnk
required to implement the zero of the system need not be
performed at every iteration, but only after the N-th iteration,
by the following difference equation:
vk [n] x[n] 2 cos(2k / N )vk [n 1] vk [n 2], 0 k N.
X [k ] yk [ N ] vk [ N ] WNk vk [ N 1]

It requires 2 real multiplications and 4 real additions to compute

vk[n] (that may be a complex sequence). The multiplication by Wnk
is performed only when n=N, which requires 4 real multiplications
and 4 real additions. Finally, a total of 2N+4 real multiplications
and 4N+4 real additions are required.
To compute all the X[k], k=0, , N1, we need 2N(N+2) real
multiplications and 4N(N+1) real additions, where the number of
multiplications are reduced by almost a half.
The Goertzel algorithm is usually used to compute X[k] for which
Decimation-in-time FFT algorithm
Most conveniently illustrated by considering the special case of N
an integer power of 2, i.e, N=2v.
Since N is an even integer, we can consider computing X[k] by
separating x[n] into two (N/2)-point sequence consisting of the
even numbered point in x[n] and the odd-numbered points in x[n].

X [k ] x[
n even
n ]W N
nk
x[
n odd
n ]W nk
N

or, with the substitution of variable n=2r for n even and n=2r+1
for n odd ( N / 2 ) 1 ( N / 2 ) 1
X [k ] x[
r 0
2 r ]W 2 rk
N x[
r 0
2 r 1]W ( 2 r 1) k
N
( N / 2 ) 1 ( N / 2 ) 1
x[
r 0
2 r ](W 2 rk
N ) W k
N x[
r 0
2 r 1](W 2 rk
N)
Since WN2 e 2 j ( 2 / N ) e j 2 /( N / 2 ) WN / 2

That is, WN2 is the root of the equation WN/2=1

Consequently,
( N / 2 ) 1 ( N / 2 ) 1
X [k ] x[
r 0
2 r ](W 2 rk
N ) W k
N x[
r 0
2 r 1](W 2 rk
N)

G[k ] WNk H [k ], k 0,1,..., N 1

Both G[k] and H[k] can be computed by (N/2)-point DFT, where

G[k] is the (N/2)-point DFT of the even numbered points of the
original sequence and the second being the (N/2)-point DFT of
the odd-numbered point of the original sequence.
Although the index ranges over N values, k = 0, 1, , N-1, each of
the sums must be computed only for k between 0 and (N/2)-1,
since G[k] and H[k] are each periodic in k with period N/2.
Decomposing N-point DFT into two (N/2)-point DFT for the
case of N=8
We can further decompose the (N/2)-point DFT into two (N/4)-
point DFTs. For example, the upper half of the previous diagram
can be decomposed as
Hence, the 8-point DFT can be obtained by the following
diagram with four 2-point DFTs.
Finally, each 2-point DFT can be implemented by the following
signal-flow graph, where no multiplications are needed.

Flow graph of a 2-point DFT

Flow graph of complete decimation-in-time decomposition of an 8-point DFT.
In each stage of the decimation-in-time FFT algorithm, there
are a basic structure called the butterfly computation:
X m [ p ] X m 1[ p ] WNr X m 1[q ]
X m [q ] X m 1[ p ] WNr X m 1[q ]

Flow graph of a basic butterfly

computation in FFT.

The butterfly computation can be simplified as follows:

Simplified butterfly computation.

Flow graph of 8-point FFT using the simplified butterfly computation
In the above, we have introduced the decimation-in-time
algorithm of FFT.
Here, we assume that N is the power of 2. For N=2v, it requires
v=log2N stages of computation.
The number of complex multiplications and additions required was
N+N+N = Nv = N log2N.
When N is not the power of 2, we can apply the same principle
that were applied in the power-of-2 case when N is a composite
integer. For example, if N=RQ, it is possible to express an N-point
DFT as either the sum of R Q-point DFTs or as the sum of Q R-
point DFTs.
In practice, by zero-padding a sequence into an N-point sequence
with N=2v, we can choose the nearest power-of-two FFT algorithm
for implementing a DFT.
The FFT algorithm of power-of-two is also called the Cooley-
Tukey algorithm since it was first proposed by them.
For short-length sequence, Goertzel algorithm might be more
efficient.
 Two-dimensional Fourier Transform
Two-dimensional transforms can be formulated by directly
extending the one-dimensional transform. Eg.

DFT of two-dimensional signal (eg., an image):

N 1 N 1
1
v[k , l ]
N
u[ m,
m 0 n 0
n ]W km
N W ln
N

N 1 N 1
1
u[m, n]
N
v[ k ,
k 0 l 0
l ]W mk
N W nl
N

Two-dimensional convolution (circular convolution):

N 1 N 1
u2 [m, n] h[(m m' ) mod N , (n n' ) mod N ]u [m' , n' ]
m ' 0 n ' 0
1