Nyquist Sampling Theorem

Definition

The Nyquist sampling theorem states that a continuous-time bandlimited signal can be completely represented and perfectly reconstructed from its samples if it is sampled at a rate greater than twice its highest frequency component. This minimum required sampling rate is called the Nyquist rate.

If the highest frequency present in the signal is $f_m$ , then the sampling frequency $f_s$ must satisfy:

$f_s \ge 2f_m$

The quantity $\frac{f_s}{2}$ is called the Nyquist frequency.

This theorem is fundamental in digital communication because it explains how analog signals such as voice, audio, and sensor signals can be converted into digital form without losing information, provided sampling is done correctly.

Main Content

1. First Concept: Sampling and the Need for Discrete Representation

In digital communication, many real-world signals are originally analog, meaning they vary continuously with time. However, digital systems work with discrete values, so these signals must be converted into a sampled form.

Sampling

is the process of measuring the amplitude of a continuous signal at regular time intervals. If a signal is sampled too slowly, important information is lost and the original signal cannot be accurately reconstructed.

For example, consider a voice signal whose highest frequency is 3.4 kHz, which is typical in telephone communication. According to the theorem:

$f_s \ge 2 \times 3.4\text{ kHz} = 6.8\text{ kHz}$

So, a sampling rate of at least 6.8 kHz is required. In practice, telephone systems use 8 kHz, which satisfies the theorem and provides a safety margin.

Why sampling matters in digital communication:

It converts analog information into a form suitable for processing, storage, encryption, transmission, and error correction.
It enables devices such as ADCs (Analog-to-Digital Converters) to interface with analog-world signals.
It forms the basis of PCM (Pulse Code Modulation), one of the most important techniques in digital communication.

Sampling representation:

Consider a smooth analog waveform:

Signal
  ^
  |        ~~~~~~~~
  |     ~~~        ~~~
  |   ~~              ~~
  |__~____________________> t

If sampled at equal time intervals:

Signal
  ^
  |        *     *     *
  |     *     *     *   
  |   *     *     *     
  |______________________> t

The stars represent the sampled values taken at specific time instants.

2. Second Concept: Nyquist Rate, Nyquist Frequency, and Aliasing

The Nyquist rate is the minimum sampling rate required to avoid loss of information. If the maximum frequency component in the signal is $f_m$ , then the Nyquist rate is $2f_m$ .
The Nyquist frequency is half the sampling frequency, $f_s/2$ . Any signal frequency above this limit cannot be uniquely represented after sampling.

A major issue related to the theorem is aliasing. Aliasing occurs when the sampling frequency is too low, causing high-frequency components to appear as lower frequencies in the sampled data. This leads to distortion and makes recovery of the original signal impossible.

Example of aliasing: Suppose a signal has a frequency of 7 kHz and is sampled at 10 kHz. The Nyquist frequency is:

$\frac{f_s}{2} = \frac{10}{2} = 5\text{ kHz}$

Since 7 kHz is greater than 5 kHz, aliasing occurs. The 7 kHz component may be incorrectly interpreted as a lower frequency in the reconstructed signal.

Why aliasing is harmful:

It introduces false frequency components.
It causes distortion in voice, audio, image, and communication signals.
Once aliasing occurs during sampling, it cannot be perfectly corrected afterward.

How to avoid aliasing:

Sample at or above the Nyquist rate.
Use an anti-aliasing filter before sampling to remove frequency components higher than $f_s/2$ .
Ensure practical systems allow a margin above the theoretical minimum because real signals may not be perfectly bandlimited.

Frequency folding idea: When sampling is insufficient, higher frequencies “fold” into lower frequencies. This is why aliasing is often described as a folding effect around the Nyquist frequency.

3. Third Concept: Reconstruction of the Original Signal and Practical Communication Systems

The Nyquist theorem does not only concern sampling; it also guarantees that a sampled signal can be perfectly reconstructed from its samples if the theorem’s condition is satisfied.
Reconstruction is typically done using an ideal low-pass filter or mathematical interpolation, which uses the sampled points to rebuild the continuous waveform.

The theoretical reconstruction formula is based on the sinc interpolation principle:

$x(t) = \sum_{n=-\infty}^{\infty} x(nT_s)\,\text{sinc}\left(\frac{t-nT_s}{T_s}\right)$

where:

$x(nT_s)$ are the sampled values,
$T_s$ is the sampling period,
$\text{sinc}(x) = \frac{\sin(\pi x)}{\pi x}$

This formula shows that the original continuous signal can be rebuilt from its samples if the sampling rate is high enough.

Role in digital communication:

In PCM, the analog waveform is sampled, quantized, and encoded into binary bits.
In voice and audio systems, correct sampling ensures faithful reproduction of speech and music.
In data acquisition, sensors measure physical quantities such as temperature, pressure, and voltage, which are then digitized for processing and transmission.

Practical notes:

Real signals are rarely perfectly bandlimited, so designers often use filters to limit the bandwidth before sampling.
Real reconstruction filters are not ideal, so systems typically sample above the Nyquist rate to reduce reconstruction errors.
Oversampling is common in modern systems because it simplifies filtering and improves signal quality.

Simple view of sampling and reconstruction:

Analog signal  ->  Sampling  ->  Discrete samples  ->  Reconstruction  ->  Analog signal

This chain is the foundation of converting continuous-time information into digital form and back again.

Working / Process

1. Determine the highest frequency component of the signal

Identify the maximum frequency $f_m$ present in the analog signal.
This is essential because the sampling rate must be based on the highest frequency, not the average or dominant frequency.
Example: if a signal contains frequencies up to 4 kHz, then $f_m = 4$ kHz.

2. Choose a sampling frequency at least twice the highest frequency

Apply the Nyquist condition: $f_s \ge 2f_m$
For $f_m = 4$ kHz, the minimum sampling frequency is 8 kHz.
In practice, engineers usually choose a slightly higher rate to account for non-ideal filters and real-world imperfections.

3. Sample the signal and reconstruct it properly

Take samples at uniform intervals $T_s = 1/f_s$ .
Pass the input signal through an anti-aliasing filter before sampling if needed.
Use interpolation or low-pass filtering during reconstruction to recover the original waveform.
If sampling was below the Nyquist rate, aliasing will occur and the original signal cannot be recovered exactly.

Advantages / Applications

Enables analog-to-digital conversion

It provides the theoretical basis for converting analog signals into digital form, which is essential for modern communication and signal processing systems.

Prevents information loss when correctly applied

By choosing the correct sampling rate, the original message signal can be represented and reconstructed with high accuracy.

Widely used in communication and multimedia systems

It is applied in PCM telephony, audio recording, digital broadcasting, image processing, radar, biomedical instrumentation, and sensor networks.

Summary

The Nyquist sampling theorem says a signal can be perfectly represented if sampled at least twice its highest frequency.
Sampling too slowly causes aliasing and loss of information.
It is the foundation of digital communication, especially in converting analog signals into digital form.
Important terms to remember: sampling rate, Nyquist rate, Nyquist frequency, aliasing, anti-aliasing filter, reconstruction