www.pudn.com > voiceprocessingtoolbox.rar > fft021.m

% 此範例展示一個簡單正弦波的傅立葉轉換，以雙邊頻譜來顯示 
% 此正弦波的頻率恰巧是 freqStep 的非整數倍，所以雙邊頻譜會「散開」(Smearing) 
 
k = 10.5;			% 非整數倍  
N = 256;			% 點數 
fs = 8000;			% 取樣頻率 
freqStep = fs/N;		% 頻域的頻率的解析度 
f = k*freqStep;			% 正弦波的頻率 
freq = freqStep*(-N/2:N/2-1);	% 頻域的頻率刻度 
time = (0:N-1)/fs;		% 時域的時間刻度 
y = cos(2*pi*f*time);		% Signal to analyze 
Y = fft(y);			% Spectrum 
Y = fftshift(Y); 
 
% Plot time data 
subplot(3,1,1); 
time2 = (0:0.1:N-1)/fs;		% Interpolated time axis 
plot(time2, cos(2*pi*f*time2)); 
line(time, y, 'color', 'r', 'marker', '.', 'linestyle', 'none'); 
title('Sinusoidal signals'); 
xlabel('Time (seconds)'); ylabel('Amplitude'); 
axis tight 
 
% Plot spectral magnitude 
magY = abs(Y); 
subplot(3,1,2); 
plot(freq, magY, '.-b'); grid on 
xlabel('Frequency)');  
ylabel('Magnitude (Linear)'); 
 
% Plot phase 
phaseY = unwrap(angle(Y)); 
subplot(3,1,3); 
plot(freq, phaseY, '.-b'); grid on 
xlabel('Frequency)');  
ylabel('Phase (Radian)');