Why my fft graph gives different output on again running the code?

y=audioread('speech.wav');
%sound(y);
subplot(2,3,1);
plot(y);
xlabel('Samples');
ylabel('Magnitude');
title('Original speech signal');

%adding noise
x=awgn(y,5);
z=y+x;
z=z/2;
%sound(z);
subplot(2,3,2);
plot(z);
xlabel('Samples');
ylabel('Magnitude');
title('Noise added to speech signal');

% framing
f_duration = 0.025;
fs=8000;
f_size = (f_duration.*fs);
n = length(y);
n_f = floor(n/f_size);  %no. of frames
temp = 0;

fr=z(1 :  f_size);
window=hamming(length(fr));
window_framing=fr.*window;

%Taking FFT
fft_sig=fft(window_framing,256);
abs_val=abs(fft_sig);
N=256;
w_axis=(2*pi*[0:N-1])/N;
subplot(2,3,3);
plot(w_axis,abs_val);

When I plot the graph of FFT, I always get different on running the again and again. Please tell me what's wrong in my code?

Answer 1

[Cris LaPierre]

I suspect the reason it changes is because you are adding randon noise to your signal with awgn. Since that changes every time the code is run, the FFT result will look slightly different every time.

[Jaya Sodhani]

Thank You. My doubt is clear now