If we sample a signal let say sine(2 * pi * f) with f=1Hz and a sampling Frequency of Fs = 8Hz, is it right that the length of the data should be N = Fs/f or multiple of Fs/f like N= d*(Fs/f) with d=1,2,3,.., is that right ?

I read often, that it is better to increase N, my question : we can only increase N with Nnew= d * Nold, right ? and not like Nnew= 1,2 * Nold?

for my example , with f=1HZ, Fs= 8Hz, T=1s, Ts=0,125s, N=8

we can increase to f=1HZ, Fs= 8Hz, T=2s, Ts=0,125s, Nnew=16,

but it would not be right to increase ist to f=1HZ, Fs= 8Hz, T=1,5s, Ts=0,125s, Nnew=12, wich is not right because x[n]=x[n+N] doesnt fit, right ?

...... if it is right above, then where do I know how to get the right length N if i down know my signal ?

Do I have to Look the x[n] Series of the Signal one by one to see where N (full Cycle) of the signal ist ?


  • $\begingroup$ Does the provided answer below gives you what you want? $\endgroup$
    – Gilles
    Jan 6, 2021 at 9:35

1 Answer 1


For a time domain signal the length or the number of samples in the sampled signal can be computed as shown in Equation $(1)$ $$ T = NT_s \equiv T = \frac NF_s \implies N = TF_s\tag{1} $$ Where $T$ is the signal total duration in [seconds] and $F_s$ is the sampling frequency in [Hertz], and $N$ the total number of samples.

The length of the DFT depends on the number of point taken in computing the DFT. For a $N$-point DFT on a signal of $N$ total samples you have $N$ as the length of your DFT.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.