PRAKTIKUM IV PEMROSESAN SINYAL
Nama : Muhammad Faisol Haq NRP
: 2408 100 010
I.
Fungsi Window dan Filter
Ketik command line berikut ini % Sampling frequency in Hz Fs = 16000; % contoh contoh Rectangular Rectangular and Hamming window, banyak fungsi fungsi jendela1 = rectwin(51); jendela2 = hamming(51);
window lainnya
% Magnitudo FFT dari fungsi window fftLength = 1024; magFJendela2 = abs(fft(jendela2, fftLength)); magFJendela1 = abs(fft(jendela1, fftLength)); %Ganti %Gan ti namaJendela namaJendela dengan dengan window window function function y ang anda pakai pakai subplot(2,1,1); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB' 'dB'); ); legend( 'Rectangular Window' ); subplot(2,1,2); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela2(1:ceil(fftLen 20*log10(magFJen dela2(1:ceil(fftLength gth /2)))); ylabel( 'dB' 'dB'); ); xlabel( 'Normalized Frequency' ); legend( 'Hamming Window' ); % Window visualization tool by MATLAB wvtool(jendela1, jendela2);
maka akan muncul grafik sebagai berikut berikut 40 Rec tangular tangular W indow indow
20 B d
0 -2 0 -4 0 -6 0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
50 Hamming W indow indow 0 B d
-5 0
-100
0
0.05
0.1
0.15
0.2 0.25 0.3 0.35 Normalized Frequency
0.4
0.45
0.5
I.
Fungsi Window dan Filter
Ketik command line berikut ini % Sampling frequency in Hz Fs = 16000; % contoh contoh Rectangular Rectangular and Hamming window, banyak fungsi fungsi jendela1 = rectwin(51); jendela2 = hamming(51);
window lainnya
% Magnitudo FFT dari fungsi window fftLength = 1024; magFJendela2 = abs(fft(jendela2, fftLength)); magFJendela1 = abs(fft(jendela1, fftLength)); %Ganti %Gan ti namaJendela namaJendela dengan dengan window window function function y ang anda pakai pakai subplot(2,1,1); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB' 'dB'); ); legend( 'Rectangular Window' ); subplot(2,1,2); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela2(1:ceil(fftLen 20*log10(magFJen dela2(1:ceil(fftLength gth /2)))); ylabel( 'dB' 'dB'); ); xlabel( 'Normalized Frequency' ); legend( 'Hamming Window' ); % Window visualization tool by MATLAB wvtool(jendela1, jendela2);
maka akan muncul grafik sebagai berikut berikut 40 Rec tangular tangular W indow indow
20 B d
0 -2 0 -4 0 -6 0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
50 Hamming W indow indow 0 B d
-5 0
-100
0
0.05
0.1
0.15
0.2 0.25 0.3 0.35 Normalized Frequency
0.4
0.45
0.5
Tim doma dom a in
F re que qu e n
d o ma in
40 1 20 0 .8 )
0
d (
d u 0 . i l p m A
e
d u i n
g a M
0 .4
0 .2
-2 0 -4 0
-6 0
80
0
10
20
30 Sa mpl s
40
50
0
0 .2 Nor ma li
0 .4 d F re que qu e n
0. 0 .8 (vT ra d/s a mple) mple)
Cari fungsi jendela selain yang diatas (minimal tiga fungsi window selain diatas). Plot masing-masing lalu bandingkan dengan fungsi filter : fir1, ellip, cheby1. Kesimpulan apa yang bisa anda peroleh ? Fungsi window lain Bartlett
dan Blackman
% Sampling frequency in Hz Fs = 16000; % contoh Bartlett and Blackman window, banyak fungsi window lainnya jendela1 = bartlett(51); jendela2 = blackman(51); % Magnitudo FFT dari fungsi window fftLength = 1024; magFJendela2 = abs(fft(jendela2, fftLength)); magFJendela1 = abs(fft(jendela1, fftLength)); %Ganti namaJendela dengan window function yang anda pakai subplot(2,1,1); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB' 'dB'); ); legend( 'Bartlett Window' ); subplot(2,1,2); plot(linspace(0,0.5,ceil(fftLength plot(linspace(0, 0.5,ceil(fftLength /2)), 20*log10(magFJendela2(1:ceil(fftLength/2)))); ylabel( 'dB' 'dB'); ); xlabel( 'Normalized Frequency' ); legend( 'Blackman Window' ); % Window visualization tool by MATLAB wvtool(jendela1, jendela2);
Diperoleh hasil
sebagai berikut
50 Bartlett W indow 0 B d
-5 0
-100
0
0.0 5
0.1
0.1 5
0.2
0.25
0.3
0 .35
0.4
0.45
0.5
50 Blackm an W indow 0 B d
-5 0 -100 -150
0
0.0 5
0.1
0.1 5
0.2 0.25 0.3 0 .35 Normalized Frequency
Time domain
0.4
0.45
0.5
Frequency domain
1
50
0. 8 0 )
B d (
e 0. 6 d u t i l p m A 0. 4
e d u t i n g a
-5 0
M
-100 0. 2
0
-150 10
20
30 Samples
40
50
0
0.2
0.4
0 .6
0.8
Normalized Frequency (vT rad /sample )
Chebyshev
dan Hann
% Sampling frequency in Hz Fs = 16000; % contoh Chebyshev and Hann window, banyak fungsi window lainnya jendela1 = chebwin(51); jendela2 = hann(51); % Magnitudo FFT dari fungsi window fftLength = 1024; magFJendela2 = abs(fft(jendela2, fftLength)); magFJendela1 = abs(fft(jendela1, fftLength)); %Ganti namaJendela dengan window function yang anda pakai subplot(2,1,1); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB'); legend( 'Chebyshev Window' ); subplot(2,1,2); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela2(1:ceil(fftLength/2)))); ylabel( 'dB'); xlabel( 'Normalized Frequency' ); legend( 'Hann Window' ); % Window visualization tool by MATLAB wvtool(jendela1, jendela2);
diperoleh hasil berikut
50 Cheby hev W indow 0 B d
-5 0 -100 -150
0
0.05
0 .1
0.15
0 .2
0 .25
0.3
0.35
0.4
0.45
0.5
50 Ha nn W indow 0 B d
-5 0 -100 -150
0
0.05
0 .1
0.15
0 .2 0 .25 0.3 0.35 Normalized Frequency
0.4
0.45
0.5
Time domain
Fre que n
doma in
50 1 0
0 .8 )
B d (
e
d u 0 .6 i l p m A
e
d u i n
-5 0
g a M
0 .4
-100 0 .2
0
-150 10
20
30 Sa mples
40
50
0
0.2 0.4 Nor mali ed F re quen
0.6 0.8 (vT ra d/s ample)
Kaiser dan Taylor % Sampling frequency in Hz Fs = 16000; % contoh Kaiser and Taylor window, banyak fungsi window lainnya jendela1 = kaiser(51); jendela2 = taylorwin(51); % Magnitudo FFT dari fungsi window fftLength = 1024; magFJendela2 = abs(fft(jendela2, fftLength)); magFJendela1 = abs(fft(jendela1, fftLength)); %Ganti namaJendela dengan window function yang anda pakai subplot(2,1,1); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB'); legend( 'Kaiser Window' ); subplot(2,1,2); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela2(1:ceil(fftLength/2)))); ylabel( 'dB'); xlabel( 'Normalized Frequency' ); legend( 'Taylor Window' ); % Window visualization tool by MATLAB wvtool(jendela1, jendela2);
didapatkan hasil berikut
40 Kai er W indow
20 B d
0 -2 0 -4 0 - 0
0
0.05
0.1
0.15
0.2
0.2 5
0.3
0.35
0.4
0 .45
0 .5
40 aylor W indow
20 B d
0 -2 0 -4 0 - 0
0
0.05
0.1
0.15
0.2 0.2 5 0.3 0.35 Normalized Frequency
ime domain 1.
0.4
0 .45
Fre quen
domain
0 .5
40
1. 4 20 1. 2 )
B d (
e 1 d u i l p m 0. 8 A
0
e
d u i n
g a -2 0 M
0. -4 0 0. 4 0. 2
- 0 10
20
30 Sample
40
50
0
0.2 0 .4 0. 0. 8 Normalized Frequency (vT ra d/s ample)
Triang dan Boxcar % Sampling frequency in Hz Fs = 16000; % contoh Triang and Boxcar window, banyak fungsi window lainnya jendela1 = triang(51); jendela2 = rectwin(51); % Magnitudo FFT dari fungsi window fftLength = 1024; magFJendela2 = abs(fft(jendela2, fftLength)); magFJendela1 = abs(fft(jendela1, fftLength)); %Ganti namaJendela dengan window function yang anda pakai subplot(2,1,1); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJen dela1(1:ceil(fftLength/2)))); ylabel( 'dB'); legend( 'Triang Window' ); subplot(2,1,2); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela2(1:ceil(fftLength/2)))); ylabel( 'dB'); xlabel( 'Normalized Frequency' ); legend( 'Boxcar Window' ); % Window visualization tool by MATLAB wvtool(jendela1, jendela2);
didapatkan hasil berikut
50 rian W indow 0 B d
-5 0 -100 -150
0
0.0 5
0.1
0.1 5
0.2
0.25
0 .3
0 .35
0 .4
0.45
0.5
40 Box ca r W indow
20 B d
0 -2 0 -4 0 - 0
0
0.0 5
0.1
0.1 5
0.2 0.25 0 .3 0 .35 Normalized Frequency
0 .4
0.45
0.5
Time domain
Fre que n
doma in
40 1
20 0
0 .8 )
e
d u 0 .6 i l p m A
B d (
-2 0
d u i n
-4 0
e
g a M
0 .4
-6 0 -80
0 .2 -100 -120
0
10
20
30 Sa mples
40
50
0
0.2 0.4 Nor mali ed F re quen
0.6 0.8 (vT ra d/s ample)
Kesimpulan: Setiap fungsi window punya karakteristik masing masing. Secara garis besar, perbedaan masing masing fungsi ini akan berpengaruh pada nilai dB dan frekuensi normal, sehingga hal itu dapat dilihat dampak pada amplitude dan waktu yang dihasilkan.
Perbandingan 3 fungsi window dengan fungsi filter Fungsi window % Sampling frequency in Hz Fs = 16000; jendela1 jendela2 jendela3 jendela4 jendela5 jendela6
= = = = = =
hann(51); flatt opwin(51); Chebwin(51); Gausswin(51); Tukeywin(51); Kaiser(51);
% Magnitudo FFT darifungsi window fftLength = 1024; magFJendela6 = abs(fft(jendela6, fftLength)); magFJendela5 = abs(fft(jendela5, fftLength)); magFJendela4 = abs(fft(jendela4, fftLength)); magFJendela3 = abs(fft(jendela3, fftLength)); magFJendela2 = abs(fft(jendela2, fftLength)); magFJendela1 = abs(fft(jendela1, fftLength));
%GantinamaJendeladengan window function yang andapakai subplot(6,1,1 ); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB'); legend( 'Hann Window' ); subplot(6,1,2); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB'); legend( 'Flattopwin Window' ); subplot(6,1,3); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB'); legend( 'Chebyshev Window' ); subplot(6,1,4); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLen gth/2)))); ylabel( 'dB'); legend( 'Gausswin Window' ); subplot(6,1,5); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela1(1:ceil(fftLength/2)))); ylabel( 'dB'); legend( 'Tukeywin Window' ); subplot(6,1,6); plot(linspace(0,0.5,ceil(fftLength/2)), 20*log10(magFJendela2(1:ceil(fftLength/2)))); ylabel( 'dB'); xlabel( 'Normalized Frequency' ); legend( 'Kaiser Window' ); % Window visualization tool by MATLAB
wvtool(jendela1, jendela2, jendela3, jendela4, jendela5, jendela6);
dan didapatkan hasil berikut
B d
2 00 0 -200
2 00 B 0 d -200 2 00 B 0 d -200 2 00 B 0 d -200 2 00 B 0 d -200 2 00 B 0 d -200
Ha nn W indow 0
0.05
0.1
0.15
0.2
0.2 5
0.3
0.35
0.4
0 .45
0 .5
Flatto win W indow 0
0.05
0.1
0.15
0.2
0.2 5
0.3
0.35
0.4
0 .45
0 .5
Cheby he v W indow 0
0.05
0.1
0.15
0.2
0.2 5
0.3
0.35
0.4 G au
0
0.05
0.1
0.15
0.2
0.2 5
0.3
0.35
0.4
0 .45
0 .5
win W indow 0 .45
0 .5
uke ywin W indow 0
0.05
0.1
0.15
0.2
0.2 5
0.3
0.35
0.4
0 .45
0 .5
Kai er W indow 0
0.05
0.1
0.15
0.2 0.2 5 0.3 0.35 Normalized Frequency
ime domain
0.4
0 .45
0 .5
Frequency domain
1. 2
50
1 0 0. e d u t i l m A
) B d -5 0 ( e d u t i n g a -100 M
0. 6 0. 4 0. 2
-150 0 -0.2
-200 10
20
30 Sam le
40
50
0
0.2 0.4 0 .6 0. Normalized Frequency (vT rad/ am le)
Fungsi filter LPF = fir1(50,[0.2 0.5]); freqz(LPF,0.5,1025)
50 ) B d ( e d u t i n a
0
-5 0
-100
0
0 .1
0.2
0.3 0.4 0.5 0. 0.7 0. Normalized Frequency (vT rad am le)
0 .9
1
0
0 .1
0.2
0.3 0.4 0.5 0. 0.7 0. Normalized Frequency (vT rad am le)
0 .9
1
1000 ) e e r e d ( e a h P
0
-1000
-2000
[b,a] = ellip(20,3,30,200/500); freqz(b,a,1025,1000) title('n=20 Lowpass Elliptic Filter' ) n=20 Lo
a
Elli tic Filter
0 ) B d ( e d u t i n
-5 0
a
-100
0
50
100
150
200 250 300 Frequency ( z )
350
400
450
500
0
50
100
150
200 250 300 Frequency ( z )
350
400
450
500
20 0 ) e e r e d ( e a h P
0 -200 -400 - 00
[b,a] = cheby1(20,3,200/450); freqz(b,a,1025,1000) 0 ) d ( - 2 0 0 d u t i n
-4 0 0
-
00 0
50
100
150
200
250
F r qu n
300
(
350
400
450
500
350
400
450
500
)
0 )
-5 0 0 r d ( - 1 0 0 0 h - 1 5 0 0 P
-2 0 0 0
0
50
100
150
200
250
F r qu n
300
(
)
Kesimpulan: Setelah membandingkan antara window function dengan fungsi filter, maka kesimpulan yang dapat diambil adalah bahwa pada window function, dapat diperoleh grafik yang menunjukkan noise asli yang rapat dan konstan. Namun pada grafik filter, noise yang diberikan, mendapat filter dari masing masing fungsi filter sehingga grafik yang dihasilkan tidak konstan.
Bagian mana yang dikehendaki dan bagian filter/window mana yang tidak dikehendaki, Mengapa ?
-
Bagian yang dikehendaki oleh window adalah grafik yang rapat dan konstan.
-
Bagian yang dikehendaki oleh filter adalah bagian yang renggang.
-
Bagian yang tidak dikehendaki window dan filter yaitu bagian noise
II.
Time-Frequency Analysis
Spectrogram adalah analisa frekuensi yang bergantung pada waktu. Spectrogram merupakan visualisasi dari kekuatan spektrum sinyal suara dengan menggunakan metode estimasi kekuatan spektrum periodogram. Ketik command line seperti berikut T
= 0:0.001:2;
X = chirp(T,100,1,200,'q'); spectrogram(X,128,120,128,1E3); title('Quadratic Chirp');
dan didapatkan hasil seperti ini Quadratic Chir 1. 1. 1.4 1.2 e m i
1 0. 0. 0.4 0.2 0
50
1 00
150
2 00 2 50 30 0 Frequency (Hz )
350
40 0
450
500
Pengertian Narrowband dan Wideband Spectrogram
merupakan jenis spectrogram yang memiliki bandwith 45-50 Hz dengan kekuatan yang berbeda beda sehingga dapat memilih masing-masing harmonic. Narrowband
merupakan jenis spectrogram yang memiliki bandwith 300-500 Hz. Pada Wideband ini ketika digunakan untuk berbicara normal dengan frekuensi dasar sekitar 100200 Hz, akan mengambil energi dari beberapa harmonic. Wideband
Modifikasi source code diatas agar mendapatkan kedua jenis spectrogram itu. Terkait dengan pertanyaan no.3, jelaskan mengapa narrowband dan wideband spectrogram tidak dikehendaki. %Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = window_narrowband; noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' );
%Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = window_wideband;
noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapatkan hasil berikut narrowband
ectro ram
) 4 00 z H ( i n e u k e r F
2 00
0
0 .2
0.4
0.
0 .2
0.4
0.
0.
1 1.2 T ime( ec) wideband ectro ram
1.4
1.
1.
2
1.4
1.
1.
2
) 4 00 z H ( i n e u k e r F
2 00
0
0.
1 1.2 time( ec)
Ubah jenis window pada spectrogram, lihat soal no. I Fungsi window dan filter diatas. Urutkan window mana yang paling cocok, sertai dengan plot dan alasan mengapa. D engan Rectangular Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024;
window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = rectwin (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap) ; xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = rectwin (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut narrowband
ectro ram
) 4 00 z H ( i n e u k e r F
2 00
0
0 .2
0.4
0.
0 .2
0.4
0.
0.
1 1.2 T ime( ec) wideband ectro ram
1.4
1.
1.
2
1.4
1.
1.
2
) 4 00 z H ( i n e u k e r F
2 00
0
0.
1 1.2 time( ec)
D engan
Hamming window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = hamming (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .00 1; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = hamming (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
narrowband
ectro ram
) 4 00 z H ( i n e u k e r F
2 00
0
0 .2
0.4
0.
0.
1 T ime(
wideband
1.2 1.4 ec ) ectro ram
1.
1.
2
1.
1.
2
) 4 00 z H ( i n e u k e r F
2 00
0
0 .2
0.4
0.
0.
1 1.2 time( ec)
1.4
D engan Bartlett Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = bartlett(51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = bartlett (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
n rro b nd
tro r
) 4 0 0 ( i n u 2 0 0 r F
0 0.2
0.4
0.
0.
1 Ti
(
id b nd
1.2 ) tro r
1.4
1.
1.
2
1.2
1.4
1.
1.
2
) 4 0 0 ( i n u 2 0 0 r F
0 0.2
0.4
0.
0.
1
ti
(
)
D engan
Blackman Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = blackman(51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideban d = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = blackman (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz )' ); title('wideband spectrogram' );
dan didapat hasil berikut
narrowband
ectro ram
) 400 z H ( i n e u k e r F
200
0
0.2
0 .4
0.
0.
1 T ime(
wideband
1.2 1.4 ec) ectro ram
1.
1.
2
1.
1.
2
) 400 z H ( i n e u k e r F
200
0
0.2
0 .4
0.
0.
1 1.2 time( ec )
1.4
D engan
Chebyshev Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = chebwin (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi ( Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = chebwin (51); noverlap = noverlap_wideband; nfft = nfft_ wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
narrowband
ectro ram
) 4 00 z H ( i n e u k e r F
2 00
0
0 .2
0.4
0.
0.
1 T ime(
wideband
1.2 1.4 ec ) ectro ram
1.
1.
2
1.
1.
2
) 4 00 z H ( i n e u k e r F
2 00
0
0 .2
0.4
0.
0.
1 1.2 time( ec)
1.4
D engan Hann Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = hann (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = hann (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
narrowband
ectro ram
) 40 0 z H ( i n e u k e r F
20 0
0
0.2
0.4
0.
0.2
0.4
0.
0.
1 1.2 T ime( ec) wideband ectro ram
1 .4
1.
1.
2
1 .4
1.
1.
2
) 40 0 z H ( i n e u k e r F
20 0
0
0.
1 1.2 time( ec)
D engan Kaiser Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = kaiser (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = kaiser (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_w ideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
n rro b nd
tro r
) 4 0 0 ( i n u 2 0 0 r F
0 0.2
0.4
0.
0.
1
1.2
Ti ( id b nd
) tro r
1.4
1.
1.
2
1.4
1.
1.
2
) 4 0 0 ( i n u 2 0 0 r F
0 0.2
0.4
0.
0.
1.2
1
ti
(
)
D engan
Taylor Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = taylorwin (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = taylorwin (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
narrowband
ectro ram
) 400 z H ( i n e u k e r F
200
0
0 .2
0.4
0.
0.
1 T ime(
wideband
1.2 1.4 ec ) ectro ram
1.
1.
2
1.
1.
2
) 400 z H ( i n e u k e r F
200
0
0 .2
0.4
0.
0.
1 1.2 time( ec)
1.4
D engan
Triang Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = triang (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = triang (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap ); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
n rro b nd
tro r
) 4 0 0 ( i n u 2 0 0 r F
0 0.2
0.4
0.
0.
1
1.2
Ti ( id b nd
) tro r
1.4
1.
1.
2
1.4
1.
1.
2
) 4 0 0 ( i n u 2 0 0 r F
0 0.2
0.4
0.
0.
1.2
1
ti
(
)
D engan
Boxcar Window
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = chirp(T,100,1,200, 'q'); nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = rectwin (51); noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap ); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = rectwin (51); noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram' );
dan didapat hasil berikut
narrowband
ectro ram
) 40 0 z H ( i n e u k e r F
20 0
0
0.2
0.4
0.
0.2
0.4
0.
0.
1 1.2 T ime( ec) wideband ectro ram
1 .4
1.
1.
2
1 .4
1.
1.
2
) 40 0 z H ( i n e u k e r F
20 0
0
0.
1 1.2 time( ec)
Penjelasan:
III.
Speech Analysis
Ketik command line berikut %Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 1000; y = 'iconk2.wav' ; nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = window_narrowband; noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram wav' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = window_wideband; noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram wav' ); narrowband
didapat hasil berikut
ectro ram wav
) 40 0 z H ( i n e u k e r F
20 0
0
1
2
3
4
5 T ime( ec )
wideband
7 x 10
-3
ectro ram wav
) 40 0 z H ( i n e u k e r F
20 0
0
1
2
3
4 5 time( ec)
7 x 10
-3
Analisis
spectrogram (narrowband dan wideband)
Mengganti
Fs = 8000
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 8000; y = 'iconk2.wav' ; nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = window_narrowb and; noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram wav' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = window_wideband; noverlap = noverlap_wideband; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram wav' );
didapat hasilnya adalah n rro b nd
tro r
4000 )
3000
( i n 2 0 0 0 u r 1 0 0 0 F
0
-0 . 1
-0 . 0 5
0
0.05 Ti
0.1
(
) tro r
(
)
id b nd
0.15
0.2
0.25
4000 )
3000
( i n 2 0 0 0 u r 1 0 0 0 F
0
-1
-0 .
0
0.
ti
1
1.
2
Mengganti
Fs = 16000
%Narrowband t_window_narrowband = .005; t_overlap_narrowband = .001; T = 0:0.001:2; Fs = 16000; y = 'iconk2.wav' ; nfft_narrowband = 1024; window_narrowband = t_window_narrowband * Fs; noverlap_narrowband = t_overlap_narrowband * Fs; jendela = window_narrowband; noverlap = noverlap_narrowband; subplot(2,1,1); specgram(y,nfft_narrowband,Fs,jendela,noverlap); xlabel( 'Time(sec)' ); ylabel( 'Frekuensi (Hz)' ); title('narrowband spectrogram wav' ); %Wideband t_window_wideband = .005; t_overlap_wideband = .001; window_wideband = t_window_wideband*Fs; noverlap_wideband = 1; nfft_wideband = 9600; jendela = window_wideband; noverlap = noverlap_wide band; nfft = nfft_wideband; subplot (2,1,2) specgram(y,nfft_wideband,Fs,jendela,noverlap); xlabel ( 'time(sec)' ); ylabel ( 'Frekuensi (Hz)' ); title('wideband spectrogram wav' );
dan didapat
narrowband
ectro ram wav
0 00 ) z H ( i n e u k e r F
0 00 4000 2000 0 -0.0
-0.04
-0 .02
0
0.02 0.04 0.0 T ime( ec) wideband ectro ram wav
-0.4
-0 .2
0
0.0
0.1
0.12
0 00 ) z H ( i n e u k e r F
0 00 4000 2000 0
0.2 0.4 time( ec )
0.
0.
1
1.2
