Sampling Continuous Time Signal

Impulse-train sampling uses a periodic impulse train—a series of delta functions spaced at multiples of the sampling period—to sample a continuous-time signal. When the signal is multiplied by this impulse train, impulses are produced with amplitudes matching the signal's value at each sampling point. This technique converts a continuous signal into discrete samples at regular intervals.

In the frequency domain, sampling results in the convolution of the original signal's spectrum with the impulse train's spectrum. This produces shifted replicas of the signal's spectrum spaced at multiples of the sampling frequency. The resulting sampled spectrum is periodic and scaled by the inverse of the sampling period.

The zero-order hold method retains each sampled value constant until the next sampling period, creating a piecewise constant or staircase-like waveform. This method approximates the original continuous-time signal and is particularly useful in digital-to-analog conversion, providing a simple way to generate a continuous signal from discrete samples.

The piecewise constant signal from zero-order hold is processed through a system with a rectangular impulse response. This system smooths out the transitions between held values, resulting in a steady output that better approximates the original signal and reduces discontinuities in the reconstructed waveform.

The sampling period determines the time interval between consecutive samples, while the sampling frequency is the reciprocal of the sampling period. In the frequency domain, shifted replicas of the signal's spectrum are spaced at intervals equal to the sampling frequency, directly reflecting this inverse relationship.

Impulse-train sampling and the zero-order hold method form a fundamental process in digital signal processing. Sampling converts continuous signals into discrete samples, while zero-order hold reconstructs a continuous approximation by holding each sample value. Together, they enable conversion of analog signals into the digital domain and their subsequent reconstruction using interpolation in applications like digital audio and telecommunications.

The sampled spectrum is periodic because convolution of the original signal's spectrum with the impulse train's spectrum produces shifted replicas at regular intervals. These replicas repeat at multiples of the sampling frequency, creating a periodic function. This periodicity is fundamental to understanding how sampling affects signal representation in the frequency domain.