How do I get the fft from the audio wav file?

Hi guys, I am new to MATLAB. I need help on how to get FFT plot from an audio file that i have?

Answer 1

[Walter Roberson]

[data, fs] = audioread('YourFile.wav');
data_fft = fft(data);
plot(abs(data_fft(:,1)))

If you want better markings such as frequency on the fft plot, then see the examples in the fft documentation to see how to construct the frequency graph.

[Aidil AA]

Hi Walter Roberson,

[data,fs]=audioread('1khz test1.wav');

data_fft = fft(data);

plot(abs(data_fft(:,1)));

title('1khz-fft');

I tested the code above on a pure 1khz sound file, I got the graph shown in the attachment. Shouldn't I just get one spike instead of two? Am I doing something wrong? Many thanks, cheers.

[Walter Roberson]

No. fft() is the two-sided fast fourier transform, returning both "negative" time and positive time. The positive time is the first half of the output and the negative time is the second half of the output. fftshift() would move the data to be symmetric around time 0.

Most people are only interested in positive time, so they mostly only plot the first half of the data. See the example plotting "Single-Sided Amplitude Spectrum of X(t)" at https://www.mathworks.com/help/matlab/ref/fft.html

[Aidil AA]

Oh I get it now, thank you. So what should I add to the code so that it will only plot the first half of the data?

[data,fs]=audioread('1khz test1.wav');
 data_fft = fft(data);

 plot(abs(data_fft(:,1)));

 title('1khz-fft');

Answer 2

[Emin ORUÇ]

Hi,how can we shift the signal to the left or right and then return to the time domain?

[Walter Roberson]

For your purposes, what does it mean to shift the signal to the left or right?

To some extent, increasing or decreasing the "bin number" of data within an fft result corresponds to frequency shifting; however this idea tends to break down when you think about negative time or think about the frequencies between 0 and the shift.

Answer 3

[Frantz Bouchereau]

Here is a popular MATLAB doc page that explains the relationship between FFT and true power spectra: Power Spectral Density Estimates Using FFT. USe this to scale the FFT to obtain true power values.