Académique Documents
Professionnel Documents
Culture Documents
Gerhard Schmidt
Christian-Albrechts-Universitt zu Kiel
Faculty of Engineering
Institute of Electrical and Information Engineering
Digital Signal Processing and System Theory
Introduction
Digital processing of continuous-time
signals
signal processing
Fast computation of the DFT: The FFT
Transformation of real-valued sequences
Digital filters
Multi-rate digital signal processing
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-2
and
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-3
and
with
The parameter
is defines as
with
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-4
. Thus, we have
with
This means that a circular convolution can be
performed very efficiently (see next slides) in the
DFT domain!
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-5
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-6
Periodic
extension
Periodic
extension
Periodic
extension
Periodic
extension
Periodic
extension
Periodic
extension
Periodic
extension
Periodic
extension
Periodic
extension
Periodic
extension
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-7
can also be carried out in the frequency domain using the DFT.
This is very attractive, since fast algorithms (fast Fourier transforms) exist.
with a length
The frequency
with length
and
with length
is defined as:
domain equivalent is
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-8
with length
and
with length
(continued):
In order to represent
the sequence
uniquely in the frequency domain by
samples of its spectrum
, the number of samples must be equal or exceed
. Thus, a DFT of size
is required.
is
Explanation on the blackboard
(if required)
with length
and
with length
leads to the same result as the linear convolution
when the lengths
of
and
are increased to
by zero padding.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-9
Alternative interpretation:
The circular convolution can be
Linear convolution
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-10
of length
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-11
point DFTs of
and
the results:
The
is terminated with
zeros, the last
signal samples from each output block
must be overlapped with (added to) the
first
signal samples of the succeeding block (linearity of convolution):
Slide IV-12
zeros
samples
added together
Output signal
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
samples
added together
Slide IV-13
.
Compute
point DFTs of
and
the results:
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-14
In order
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-15
Copy
samples
Copy
samples
Input signal
(all elements are filled)
Output signal
Discard
samples
Discard
samples
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Discard
samples
Slide IV-16
..
..
Are there
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-17
in practice:
Anti-aliasing lowpass
For
( : number of samples
input signal
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-18
can be obtained as
is
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-19
for
and
any
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-20
crossing of
at
Decreasing the frequency resolution (making the window width smaller) leads to
an increase of the time resolution and vice versa.
Duality of time and frequency domain.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-21
except for
, rectangular window
Results:
DFT of
except for
since
is exactly an integer multiple of
The periodic repetition of
leads to a pure cosine.
DFT of
:
The periodic repetition of
for
is not a cosine sequence anymore.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-22
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-23
and smeared
in general?
..
..
What can you do in order
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-24
where
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-25
and
: The two maxima (the main lobes) for both window spectra
and
can be separated.
: Correct values for the spectral samples, but the main lobes cannot
be separated anymore.
and
overlap.
The ability to resolve spectral lines of different frequencies is limited by the main lobe width,
which also depends on the length of the window impulse response .
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-26
with
and
for
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-27
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-28
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-29
of the cosine-function
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
after windowing
Slide IV-30
The reduction
of the sidelobes
and the reduced
resolution compared to the
rectangular
window can be
clearly observed.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-31
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-32
Remarks:
The sampling grid can be arbitrarily fine by increasing the length of the windowed
An increase
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-33
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-34
..
..
How
do you select a window function? What prior information might be useful to know
before you chose a window function?
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-35
and complex
1 complex addition
When looking closer
we see that not all operations require the above mentioned complexity:
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-36
Memory
hardware.
Basic idea for reducing the computational complexity:
The basic idea for reducing the complexity of a DFT is to decompose the big problem into
several smaller problems. This usually leads to a reduction in complexity. However, this
trick can not always be applied.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-37
and for
For a so-called radix 2 realization of the DFT we decompose the input series into two series
of half length:
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-38
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-39
DFT of length
DFT of length
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
and
also
Slide IV-40
the decomposition:
1 DFT of order
operations.
2 DFTs of order
operations and
operations.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-41
DFT
of
order
4
DFT
of
order
4
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-42
As a result
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-43
DFT
of order
2
DFT
of order
2
DFT
of order
2
DFT
of order
2
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-44
As we can see, also over here we need the same basic scheme that we have used also in the
previous decompositions:
This basic scheme is called butterfly of a radix-2 FFT. The abbreviation FFT stands for
Fast Fourier Transform.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-45
This leads to a further reduction in terms of multiplications (only 50 % of them are really required).
In total we were able to reduce the required operations for computing a DFT from
by using efficient radix-2 approaches.
down to
Examples:
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-46
Please keep in mind that in each stage only in-place operations are required.
This means that no now memory has to be allocated for a new stage!
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-47
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-48
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-49
, are
obtained as
Only
In-place algorithm
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-50
Example for
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-51
[
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
that is
With additional scaling and index permutations the IDFT can be calculated with the
same FFT algorithm as the DFT.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-53
without bit-reversal, or
(b)
The approach (a) has the disadvantage that it is a non-inplace algorithm, because the
butterfly-structure does not continue past the first stage.
Next slides: the flow graphs for both approaches.
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-54
Slide IV-55
Slide IV-56
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-57
Slide IV-58
and
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-59
Slide IV-60
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-61
..
..
Which operations should be counted besides multiplications and additions when
How
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-62
with
and
and
by
Define
leading to the DFT
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-63
into
and
Corresponding DFTs:
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-64
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-65
with
the relation
with
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-66
have to be calculated.
can be derived from the values for
Application:
Fast convolution of two real-valued sequences with the DFT/FFT!
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-67
with
Hermitian symmetry since
for all :
Define
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-68
and
and
for
In order to calculate
for
.
from
and
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-69
Finally we have:
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-70
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-71
many memory elements do you need to store the result of a DFT of order M
if the input sequence was real-valued?
..
..
(General
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-72
Introduction
Digital processing of continuous-time
signals
signal processing
Fast computation of the DFT: The FFT
Transformation of real-valued sequences
Digital filters
Multi-rate digital signal processing
Digital Signal Processing and System Theory| Advanced Digital Signal Processing | DFT and FFT
Slide IV-73