The x-axis will be frequency, that being defined as cycles/(time unit of the original independent varialble). If that is in seconds, the frequency will be in Hz.
The best way to do that is:
[y,Fs] = audioread('speech.wav'); % Return Sampling Frequency As The Second Output
L = size(y,1);
Fn = Fs/2; % Nyquist Frequency
tv = linspace(0, L-1, L)/Fs; % Time Vector
One way to compute the frequency vector:
Fv = linspace(0, 1, fix(length(fft)/2)+1)*Fn; % One-Sided 'fft'
Iv = 1:numel(Fv); % Index Vector
The fft result to be plotted would then use ‘Fv’ and ‘Iv’ to create an equal-length fft vector.
Fv = linspace(-Fs/2, Fs/2-Fs/length(s), length(s)); % Two-Sided 'fft': EVEN 'length(s)' (Asymmetric)
Fv = linspace(-Fs/2, Fs/2, length(s)); % Two-Sided 'fft': ODD 'length(s)' (Symmetric)
.
