Hi, If the signal is real-valued, it is better to not use the whole result of fft() since you are looking at the absolute value.
The spacing of the DFT bins is Fs/length(input) as you note in your code above.
t = 0:0.001:1-0.001;
Fs = 1000;
x = cos(2*pi*100*t)+randn(size(t));
xdft = fft(x);
xmag = abs(xdft(1:length(xdft)/2+1));
freq = 0:Fs/length(x):Fs/2;
Now, how do we find the value of xmag that corresponds to 100 Hz?
The spacing between the DFT bins is Fs/length(x), so
df = Fs/length(x);
freqbin = round(100/df)+1;
xmag(freqbin)
Of course if you want the actual Fourier coefficient at that frequency
xdft(freqbin)
