How to do a fft from a sine signal with hanning function

Hello, i want to do a fast fourier transformation on a sine signal. By using the hanning function. Btw i am a newbie, concering matlab programming so, dont be to hard too me. If you could also help me explain the hanning function properly, i would really appreciate it.

Answer 1

[Gareth Thomas]

MATLAB has awesome help. If you type

doc fft

You should get a nice page with examples that should be enough to get you going: FFT in MATLAB

[Jack Walker]

Thank you, i will try to read the link you provided it. Although tt is a little bit complicated to me xD

Answer 2

[Jan Orwat]

You can try starting with something like this

fs = 44100; % sampling frequency
T = 1; % period, 1 second
n = T*fs;
t = linspace(1/fs, T, n);

% calculate hanning window, if you have Signal Processing Toolbox just use hann(n)
window = 0.5 - 0.5*cos(2*pi*linspace(0, 1, n));

% generate simplest sine signal
freq = 1000; % 1kHz sine wave
s = sin(2*pi*freq);

% calculate fft
z = fft(s.*window);

hanning window is just 1 - one period of cosine function, it's usually normalised to 1.

[Jack Walker]

Thank you for helping me! I think you made a mistake at the 4th line. I think you have to write fs in capital letters, otherwise i get an error in my commandwindow. Btw is it possible to make it visible? i tried to plot it, but i think i did something wrong..

[Jan Orwat]

That's right, corrected.

Show your plotting attempt, let's investigate what's wrong there.

[Jack Walker]

To be honest, i am a little bit embarrassed about it, i just tried to make it as simple as possible. I wrote plot(z) at end of your code, because i was curious, if can understand more about your code...

[mehdi fathizad]

hi everone

Now i am doing a project about damage detection of structure based on ARMA model. I have already taken all the accelerations for damaged and undamaged system and I want to use pre-whitening procedure to remove correlation between the story's responses. I read somewhere that for doing pre-whitening its better to do hanning window at first on my response time series. then carry out the fft fallowing the normalization. at the end, inverse FFT and then start the ARMA procedure on the obtained new time series.

I am wondering if its correct (whole concept and the arrangement)?

and also normalization means dividing all magnitude by the maximum amount?

if so, after producing the new time series, how can I check to find out the procedure was rigth or not?(for example checking the variance, mean and so on)

thanks

[David Pinho]

s = sin(2*pi*freq); should be sin(2*pi*freq*t)