CS3570 Introduction to Multimedia

Homework #2
Due: 11:59pm, 4/14/2014

1. Create your own FIR filters to filter audio signal:
With the ideal impulse responses given in Table 5.2 (slide #82) and the
windowing functions in Table 5.3 (slide #85), we can design a filter according to
algorithm 5.1 (slide #86). In this homework, you need to design different filters
and test their effects.
hw2_mix.wav is a mix of 3 songs, and you need to apply appropriate filters
to separate the three audio signals from this audio file.
Apply appropriate FIR filters in time domain, and matlab built-in function
conv is not allowed to use in this homework. You need to implement 1-D
convolution by yourself.
Store the filtered audio files, and plot
(1) The spectrum of the input signal
(2) The spectrums of the output signals
(3) The spectrums of the filters
(4) The shapes of the filters (time domain)
Implement the following 3 kinds of
windowing functions, and compare their
results.
(1) Hanning windowing function
(2) Hamming windowing function
(3) Blackman windowing function
Discuss how you determine the filters for this homework problem.
You have to upload m-files and three filtered audio files. Show the plots and
the discussions in your report.
Hint:
You can use the following code to plot spectrum:
% y1: signal, Fs1: sampling rate
%{
L = 2^nextpow2(max(size(y1)));
y1_FFT = fft(y1,L);
xx = Fs1/2*linspace(0,1,L/2+1);
figure, plot(xx,abs(y1_FFT(1:L/2+1)));
%}