UCLA
Dept. of Electrical Engineering
EE 114, Winter 2013
Problem Set 2
Due: January 23, 2013
1. During spectral analysis of speech, a pre-emphasis lter is generally used to boost higher
frequencies, since they tend to contain discriminative information. A
EE114, Winter 2011
Computer Assignment 5 Solution
_
1 Matlab functions
1.1 ScaleFTAmp.m
function FAmp = ScaleFTAmp(F)
% shift the FT with shift, log, and rescale
c = 4.;
% take the log of the amplitude
FAmp = log10(abs(F)/max(max(abs(F) + 0.000001);
% nor
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2013
Computer Assignment 5: Solutions
Matlab code We need to write the following of m-les:
GenerateGaussian.m
function h = GenerateGaussian (size, lambda)
% this function generates a Gaussian matrix
% ge
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Computer Assignment 6: 2D DFT
Due: February 26, 2014
Introduction: In this assignment, you will experiment with the two-dimensional Fourier Transform of digital images.
The following tools in Matlab
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2013
Computer Assignment 1: Introduction to Frequency Analysis
Due: January 16, 2013
EE 114 Computer Assignment Report Format: When completing computer assignment
reports, we recommend that you start eac
1
Review of Digital Signal Processing Fundamentals
1.1
Discrete-Time Signals
Speech signals occur naturally as continuous-time acoustic signals, xa (t). However, speech signals
1
can be transduced into electrical signals, and sampled at period T = Fs to b
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2013
Problem Set 5 Solutions
1. Find the 2D Fourier Transform of the two-dimensional function
f (x, y ) = rect(ax + b) sinc(cy ) .
Answer:
The function f (x, y ) is separable, i.e., f (x, y ) = g (x) h (
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 5
1. Find the 2D Fourier Transform of the two-dimensional function
f (x, y ) = rect (ax + b) sinc (cy ) .
2. Let f (x, y ) be a function below with value 1 in the shaded region and 0 oth
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 4
1. Consider the signal:
x (n) = n for n 0
Using the autocorrelation method of linear prediction analysis, nd the 1st order prediction coecient, i.e., a1 . Find the corresponding error,
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 1 - Solutions
1. Consider the discrete time sequence:
x (n) = [4, 1, 2]
(a) Compute the Z-transform X (z ).
Solution:
x (n) z n
X (z ) =
n=
1
= 4z
+ 2 z 2
(b) Let y (n) = x (n) x (n). De
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2013
Problem Set 3
Due: January 30, 2013
1. The Fourier Transform can be expressed as:
Xn ej = an ( ) jbn ( ) = |Xn ej |ejn ()
(1)
If x (n) is real, nd the relationship between the following:
(a) an ( )
EE114 Speech and Image Processing Systems Design Computer Assignment 2
Speech and Image Processing System Design
Computer Assignment 2
- Pre-emphasis filter -
Subject
Professor
Department
Name
Student number
Due date
:
:
:
:
:
:
EE114
Flavio Lorenzelli
EE
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Computer Assignment 1 - Solutions
Abstract: We consider the analysis of a speech signal in both the time domain and the frequency domain. First we look at the time domain representation of the sente
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 7
1. The four basis vectors of the N = 4 1D DFT form an orthonormal set, i.e.,
aH a =
k
where
H
1, k = ,
0, otherwise,
denotes conjugate transpose.
(a) Conrm that the above relationship
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Computer Assignment 3: Temporal Analysis of Speech
Due: January 29, 2014
Introduction: This assignment focuses on the temporal analysis of speech.
Short-time energy analysis:
Download the function c
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Computer Assignment 2 Solutions
Abstract: This assignment explored the eect of pre-emphasis lters on speech lters. Specically,
it studied the eect of pre-emphasis on perception and spectral shape. A
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 7 - Solutions
1. The four basis vectors of the N = 4 1D DFT form an orthonormal set, i.e.,
aH a =
k
where
H
1, k = ,
0, otherwise,
denotes conjugate transpose.
(a) Conrm that the above r
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 1
1. Figure 1 represents the spectrum of a steady-state vowel, |X e j |. The approximated vocal
tract transfer function |H e j | is shown as a dashed line.
Figure 1: The spectrum of a st
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 6 - Solutions
1. Consider a continuous 2D signal of the form
f (x, y ) = cos (2 (4x + 3y ) .
Suppose you wish to design a sampling/reconstruction system that fails to satisfy the
Nyquist
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 2 - Solutions
1. During spectral analysis of speech, a pre-emphasis lter is generally used to boost higher
frequencies, since they tend to contain discriminative information. A common fo
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2014
Problem Set 6
1. Consider a continuous 2D signal of the form
f (x, y ) = cos (2 (4x + 3y ) .
Suppose you wish to design a sampling/reconstruction system that fails to satisfy the
Nyquist requirement
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2013
Computer Assignment 7: Image Enhancement
Due: March 6, 2013
Introduction: In this assignment, you will experiment with image enhancement tools.
Use the image from the previous assignment, ca6 image.
EE114 Speech and Image Processing Systems Design Computer Assignment 6
Speech and Image Processing System Design
Computer Assignment 6
- Simple Image compression(LPF) -
Subject
Professor
Department
Name
Student number
Due date
:
:
:
:
:
:
EE114
Flavio Lor
1
Review of Digital Signal Processing Fundamentals
1.1
Discrete-Time Signals
Speech signals occur naturally as continuous-time acoustic signals, xa (t). However, speech signals
1
can be transduced into electrical signals, and sampled at period T = Fs to b
UCLA
Dept. of Electrical Engineering
EE 114, Winter 2013
Problem Set 7 Solutions
1. The four basis vectors of the N = 4 1D DFT form an orthonormal set, i.e.,
1, k = ,
0, otherwise,
aH a =
k
where
H
denotes conjugate transpose.
a) Conrm that the above rela
Dylan Ler | 104172087 | EE114 Computer Assignment 5
2-D Convolution
For MATLAB Code:
GenerateGaussian.m
function h = GenerateGaussian (size, lambda)
this function generates a Gaussian matrix
generate the meshgrid and the Gaussian matrix
[x,y]= meshgrid([-
Dylan Ler | 104172087 | EE114 Computer Assignment 1
The figure above is the time waveform of phoneme /a/
1.
MATLAB code for dft.m:
for k=0:N-1
for n=0:N-1
real(k+1) = real(k+1)+x(n+1)*cos(2*pi*(k+1)*(n+1)/N);
imag(k+1) = imag(k+1)-x(n+1)*sin(2*pi*(k+1)*(n