The discrete Fourier series is very related to the Fourier transform.
This post shows that one can be derived from the other.
A periodic signal , with fundamental period N, can be represented as a sum of complex exponentials with fundamental angular frequency :
The coefficients represent the Fourier Series. These coefficients can be obtained from the inner product of the periodic signal with the exponential basis:
For a given period N, and letting we can derive these equations in matrix forms:
The coefficients can be obtained similarly:
Let us generate an aperiodic signal by taking N samples from the periodic signal and setting to zero the remaining samples. By making , we can derive the Fourier transform of this aperiodic signal:
From this equation we can easily see that the coefficients of the periodic signal can be obtained by sampling the Fourier transform of the associated aperiodic signal at frequencies :
Moreover, any signal, periodic or aperiodic, written as:
has a Fourier transform, given as:
.
If is periodic with period N and , then
.
From these results we may conclude:
- From the Fourier transform of an aperiodic signal we derive the Fourier series coefficients of an equivalent periodic signal of chosen period N, by sampling the Fourier transform.
- From the Fourier series coefficients of a periodic signal we easily derive the Fourier transform of the same signal.