Review of Discrete Fourier Transform: Cwliu@twins - Ee.nctu - Edu.tw

Review of Discrete Fourier Transform
cwliu@twins.ee.nctu.edu.tw
4 Forms of Fourier Transform
Sampled frequency
Continuous-Time and Continuous-Frequency
Continuous Aperiodic
Continuous Aperiodic
Continuous-Time and Discrete-Frequency
Fourier series of periodic continuous signals
Periodic Continuous
Discrete Aperiodic
4
Discrete-Time and Continuous-Frequency
Fourier transform of aperiodic discrete signals
Discrete Aperiodic
Continuous Periodic
5
Discrete Fourier Transform
DFT is identical to samples of Fourier transforms In DSP applications, we are able to store only a finite number of samples we are able to compute the spectrum only at specific discrete values of
cwliu@twins.ee.nctu.edu.tw 6
Discrete Fourier Transform

Discrete Fourier transform (DFT) pairs
kn X [k ] = x[n]WN , k = 0,1, K , N 1 n =0 N 1
1 x[n] = N
kn X [ k ] W n = 0,1, K , N 1, N , k =0 j 2 kn N
N 1
N complex multiplications N-1 complex additions
kn where WN =e
DFT/IDFT can be implemented by using the same hardware It requires N2 complex multiplications and N(N-1) complex additions
More About DFT

Properties of Discrete Fourier Transform Linear Convolution and Discrete Fourier Transform Discrete Cosine Transform
Periodic Sequence
Consider a periodic sequence x [n] of period N The sequence can be represented by Fourier series 1 ~ ~ x [n] = X [k ]e j (2 / N )kn N k The Fourier series for any discrete-time signal with period N requires only N harmonically related complex exponentials.
Q ek +lN [n] = e j (2 / N )(k +lN )n = e j (2 / N )kn = ek [n]

1 ~ x [ n] = N ~ j ( 2 / N )kn X [ k ] e
k =0
9
N 1
Apply the Orthogonality property, we have
Interchange the order of summation
The coefficients are also periodic with period N cwliu@twins.ee.nctu.edu.tw
10
DFS Representation of a Periodic Sequence
Synthesis equation
Analysis equation
~ X [k ] and ~ x [n] are periodic sequence of period N

Physical Significance
Let One period
Then
~ j X[k] vs X (e )
Example
13
Sampling the Fourier Transform

Suppose Then exists
2 N
or The sampling sequence is periodic with period N Since
unit circle
14
~ x [n] vs x[n]
By adding together an infinite number of shifted replicas of x[n]

Aliasing Problem 1
x[n] is infinite-length sequence
~ x [ n]
16
Aliasing Problem 2
If x[n] is finite-length sequence, 0nM-1 Consider the case N<M
~ x [ n]
~ x [n] x[n]
Concluding Remarks
The case NM
~ x [ n]
or
Property of DFT
Linearity
if of length N1 of length N2 then of length N3=max[N1 , N2]
19
Circular Shift of a Sequence

N=15
x[n]
~ x [ n]
~ x [ n 2]
~ x [ n 2] R N [ n ]
A rotation of the cylinder

20
Circular Shift of a Sequence

N=15
x[n]
~ x [ n]
~ x [n + 13]
~ x [n + 13]R15 [n]
A rotation of the cylinder

21
Property of DFT
Circular Shift
if then of length N
m
that is On the other hand
A rotation of the sequence in the interval
0 n N-1 0 l N-1
Other Properties of DFT

Duality
8.6.3
Symmetry
8.6.4
23
More About DFT

Properties of Discrete Fourier Transform Linear Convolution and Discrete Fourier Transform Discrete Cosine Transform
24
Review of Convolution
x L-point sequence h N-point sequence s (L+N-1)-point sequence
Given two sequences:

Data sequence xi, 0 i N-1, of length N Filter sequence hi, 0 i L-1, of length L
Linear convolution
Direct computation, for example 2-by-2 convolution
yi = xi hi = hi xi , i = 0,1, K, L + N 2
s0 h0 0 s = h h x0 0 1 1 x 1 s 0 h cwliu@twins.ee.nctu.edu.tw 1 2
NL multiplications
require 4 multiplications and 1 addition

25
Linear Convolution
Linear Shift
Linear Shift Linear Shift
26
Linear Shift vs Circular Shift
Conventional shift (linear shift)

Circular Shift Example
28
Periodic/Circular Convolution
Circular Shift
29
Circular Convolution Definition

Suppose two finite-length duration sequences: x1[n] and x2[n] of length N
x3[n] is also a finite-length duration sequences of length N
30
Computation for Circular Convolution

1. To period the two sequence with period N (large enough) 2. To compute the periodic convolution of the two periodic sequences 3. To get out the duration sequence between [0, N-1]
31
Example
Step 2
Step 1
Step 3
32
Circular Convolution Property

Usually, we use the following notation to represent the circular convolution of length N
Circular convolution property
33
Circular Convolution Implementation

Direct Implementation
x N-point sequence 44 cyclic convolution Circular Convolution h N-point sequence s N-point sequence
~ O(N2) 16 multiplications 12 additions

Using Circular Convolution to Implement Linear Convolution

Consider two sequences x1[n] of length L and x2[n] of length P, respectively The linear convolution x3=x1[n] x2[n]
a sequence of length L+P-1
Choose N, such that NL+P-1, then
The same concept related to Winogrand Algorithm
35
Linear Convolution
36
Circular Convolution with N=L+P-1
Time aliasing in the circular convolution of two finite-length sequence can be avoided if N L+P-1
Concluding Remarks
The convolution of two finite-length sequences can be interpreted by circular convolution with large enough length Circular convolution can be implemented by DFT/FFT
However, in real applications.

A large delay in processing An indefinite memory
For an FIR system, the input sequence is of indefinite duration To store the entire input signal requires ? Block convolution
Block Convolution
Step1: To segment a sequence into sections of length L Step2: Each section is convolved with the finite-length impulse response of length P by using DFT/FFT of length N=L+P-1 Step3: The filtered sections are fitted together in an appropriate way Overlap-add method Overlap-save method
Overlap-Add Method
h[n] x h y x[n]
Step1
Zero padding
Zero padding Zero padding

Step2 & Step3
yr [n] = xr [n] h[n] = xr [n] N h[n]

Time shift
with L+P-1 length
Time shift
41
Overlap-Save Method
Suppose L > P. Consider an L-point circular convolution of a P-point impulse response h[n] with an L-point input sequence xr[n]
Due to aliasing problem, the first (P-1)-point of the result is incorrect the remaining points [P, L-1] are identical to those that would be obtained by linear convolution
Step1: To segment a sequence into sections of length L such that each section overlaps the preceding section by (P-1) points Step2: Each section is convolved with the finite-length impulse response of length P by using DFT/FFT of length L Step3: The first (P-1)-point of each filtered sequence must be discarded. The remaining samples from successive sections are then abutted to construct the final output.
42
Step1
43
Step2 & Step3
44
Fast Convolution with the FFT

Given two sequences x1 and x2 of length N1 and N2 respectively Consider using FFT to convolve two sequences:
Direct implementation requires N1N2 complex multiplications
Pick N, a power of 2, such that NN1+N2-1 Zero-pad x1 and x2 to length N Compute N-point FFTs of zero-padded x1 and x2, one obtains X1 and X2 Multiply X1 and X2 Apply the IFFT to obtain the convolution sum of x1 and x2 Computation complexity: 2(N/2) log2N + N + (N/2)log2N
Example
A sequence x[n] of length 1024 FIR filter h[n] of length 34 Direct computation: 341024=34816 Using radix-2 FFT: 35840 (N=2048) Using overlap-add radix-2 FFT:
x[n] is segmented into a set of contiguous blocks of equal length 95 Apply radix-2 FFT of length 128 Each segment requires 1472 multiplications This algorithm requires total 16192 multiplications
46

Review of Discrete Fourier Transform: Cwliu@twins - Ee.nctu - Edu.tw

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Review of Discrete Fourier Transform: Cwliu@twins - Ee.nctu - Edu.tw

Diunggah oleh

Hak Cipta:

Format Tersedia

Review of Discrete Fourier Transform

4 Forms of Fourier Transform

Continuous-Time and Continuous-Frequency

Continuous-Time and Discrete-Frequency

Fourier series of periodic continuous signals

Discrete-Time and Continuous-Frequency

Fourier transform of aperiodic discrete signals

Discrete Fourier Transform

Discrete Fourier Transform

N complex multiplications N-1 complex additions

More About DFT

Q ek +lN [n] = e j (2 / N )(k +lN )n = e j (2 / N )kn = ek [n]

Apply the Orthogonality property, we have

Interchange the order of summation

The coefficients are also periodic with period N cwliu@twins.ee.nctu.edu.tw

DFS Representation of a Periodic Sequence

~ X [k ] and ~ x [n] are periodic sequence of period N

Sampling the Fourier Transform

or The sampling sequence is periodic with period N Since

By adding together an infinite number of shifted replicas of x[n]

Circular Shift of a Sequence

A rotation of the cylinder

Circular Shift of a Sequence

A rotation of the cylinder

A rotation of the sequence in the interval

Other Properties of DFT

More About DFT

Given two sequences:

Direct computation, for example 2-by-2 convolution

require 4 multiplications and 1 addition

Linear Shift Linear Shift

Linear Shift vs Circular Shift

Conventional shift (linear shift)

Circular Shift Example

Circular Convolution Definition

x3[n] is also a finite-length duration sequences of length N

Computation for Circular Convolution

Circular Convolution Property

Circular convolution property

Circular Convolution Implementation

~ O(N2) 16 multiplications 12 additions

Using Circular Convolution to Implement Linear Convolution

a sequence of length L+P-1

Choose N, such that NL+P-1, then

The same concept related to Winogrand Algorithm

Circular Convolution with N=L+P-1

However, in real applications.

Zero padding Zero padding

Step2 & Step3

yr [n] = xr [n] h[n] = xr [n] N h[n]

with L+P-1 length

Step2 & Step3

Fast Convolution with the FFT

Anda mungkin juga menyukai