The sampling theorem

If a signal is not sampled frequently enough, its digital representation will be totally different from the original signal. This effect is called Aliasing. It can not be corrected once the signal has been sampled!

The sampling frequency must be at least twice the highest frequency present in the signal. (Nyquist's theorem) - In other words: Frequencies higher than half the sampling frequency must not occur in the analog signal.

Sampling theorem and aliasing in the frequency domain

Example: if a signal is sampled 128 times/second, one must make sure in advance that the signal does not contain oscillations higher than 64 Hz. This can be achieved by filtering.

 

 

 

©1998 BITS OF SLEEP (Demo) : Human sleep investigation: Signal analysis basics.