Understanding the Nyquist Theorem in Digital Signal Processing
Digital(Discrete-time) signal processing relies on one fundamental principle that governs how continuous signals can be represented in digital form: the Nyquist Theorem.
Whenever an analog signal is converted into a digital signal through sampling, the sampling rate determines whether the original information can be reconstructed accurately. If the sampling rate is chosen incorrectly, distortions known as aliasing appear, permanently altering the signal's frequency content.
In this article, we will explore
What the Nyquist theorem is
Why it is essential for digital(discrete-time) signal processing
What happens when the sampling rule is violated
How engineers apply the theorem in practical systems
What Is the Nyquist Theorem?
The Nyquist TheoremNyquist–Shannon Sampling Theorem, states that a continuous signal can be perfectly reconstructed from its samples if the sampling frequency is greater than twice the highest frequency component of the signal.
Mathematically
Fs ≥ 2Fmax
- Fs: Sampling frequency
- Fmax: Maximum frequency of signal
Commonly used concept is the Nyquist frequency, which is defined as
This represents the highest frequency that can be represented without aliasing.
Why Sampling Is Necessary
Real-world signals are often continuous in time, meaning they exist at every moment. However, digital systems cannot process continuous signals directly. Instead, signals must be converted into a sequence of discrete samples.
Sampling is the process of measuring the signal amplitude at regular time intervals.
For example
| Signal Type | Representation |
|---|
| Analog signal | Continuous waveform |
| Digital signal | Discrete samples |
The goal is to choose a sampling rate high enough to preserve the original signal content.
Continuous signal(assumed) and its sampled representation
The Nyquist Frequency
The Nyquist frequency represents the maximum frequency that can be captured without distortion.
If a signal contains frequencies higher than the Nyquist frequency, those components will fold back into the spectrum, producing incorrect frequency components.
For example
| Sampling Frequency | Nyquist frequency |
|---|
| 1000Hz | 500 Hz |
| 2000Hz | 1000 Hz |
| 8000Hz | 4000 Hz |
This means that if the signal contains frequencies above the Nyquist frequency, aliasing will occur.
Frequency spectrum showing components below the Nyquist limit
MALMIJAL example
What Is Aliasing?
Aliasing occurs when a signal contains frequency components that exceed the Nyquist frequency.
Instead of appearing at their true frequency, these components appear at a different frequency within the measurable range.
This phenomenon happens because the sampling process cannot distinguish between certain high-frequency components.
The aliasing frequency can be calculated as
- fsignal is actual signal frequency
- Fs is sampling frequency
- n is integer
Example of frequency folding caused by insufficient sampling rate
MALMIJAL example
Visualizing Aliasing
A useful way to understand aliasing is to compare two sampling scenarios.
| Case | Sampling Frequency | Result |
|---|
| Proper sampling | Above Nyquist frequency | Accurate frequency |
| Undersampling | Below Nyquist frequency | Frequency-folding |
When the sampling frequency is too low, the waveform still appears periodic, but it actually represents the wrong signal.
Same signal sampled at different rates demonstrating aliasing
MALMIJAL example
Signal: 400 Hz sine wave
Case 1: Fs = 2000 Hz
Case 2: Fs = 600 Hz
In this example, the signal frequency is 400 Hz, the Nyquist condition (400 < Fs/2 = 1000) is satisfied and the waveform is represented correctly. However, when the same signal is sampled at 600 Hz, the Nyquist frequency becomes 300 Hz, which is lower than the signal frequency.
As a result, aliasing occurs.
The aliasing frequency can be calculated as:
For this example:
This means that a 400 Hz signal sampled at 600 Hz will appear as a 200 Hz signal in the sampled data.
frequency domain. When the FFT is computed for the properly sampled signal, the spectrum shows a peak at 400 Hz. However, when the signal is sampled at 600 Hz, the FFT instead shows a peak at 200 Hz, which corresponds to the alias frequency predicted by the equation.
FFT comparison showing the original 400 Hz signal and its 200 Hz alias caused by undersampling
Nyquist Theorem in Real Engineering Systems
In real systems, engineers rarely use the Nyquist rate exactly.
Instead, the sampling rate is usually chosen higher than the theoretical limit
Common examples include
| Application | Sampling Rate |
|---|
| Audio CD | 44.1 kHz |
| Speech processing | 16 kHz |
| Vibration analysis | several kHz |
| RF systems | MHz or higher |
Anti-aliasing filter
Before sampling, analog signals are typically passed through a low-pass filter that removes frequencies above the Nyquist frequency.
This prevents aliasing from occurring during the sampling process.
Digital anti-aliasing filtering ensures that the signal satisfies the Nyquist condition before downsampling
(refer to Samples/anti-aliasing.mmj)
MALMIJAL example
Original signal
LPF applied
Downsampling result
Practical Rule Used by Engineers
In practice, engineers rarely place the cutoff frequency exactly at the Nyquist frequency.
Instead, the cutoff is usually set slightly below the Nyquist frequency
For example
| Sampling Frequency (Fs ) | Nyquist Frequency | Typical LPF cutoff |
|---|
| 2000Hz | 1000Hz | 800 ~ 900 Hz (0.4Fs ~ 0.45Fs) |
This margin ensures that unwanted high-frequency components are sufficiently attenuated.
Summary
The Nyquist theorem is one of the most fundamental principles in digital signal processing.
It defines the minimum sampling rate required to preserve the frequency content of a signal.
Key takeaways
A signal must be sampled at at least twice its highest frequency
The Nyquist frequency is half of the sampling rate
Frequencies above the Nyquist limit cause aliasing
Anti-aliasing filters are used to prevent this problem
Practical systems use a margin below the Nyquist frequency
Understanding the Nyquist theorem is essential for designing reliable digital signal processing systems, from audio recording to advanced engineering measurements.
Suggested Further Reading
#You may also find these topics helpful:
Understanding the Nyquist Theorem in Digital Signal Processing
Digital(Discrete-time) signal processing relies on one fundamental principle that governs how continuous signals can be represented in digital form: the Nyquist Theorem.
Whenever an analog signal is converted into a digital signal through sampling, the sampling rate determines whether the original information can be reconstructed accurately. If the sampling rate is chosen incorrectly, distortions known as aliasing appear, permanently altering the signal's frequency content.
In this article, we will explore
What the Nyquist theorem is
Why it is essential for digital(discrete-time) signal processing
What happens when the sampling rule is violated
How engineers apply the theorem in practical systems
What Is the Nyquist Theorem?
The Nyquist TheoremNyquist–Shannon Sampling Theorem, states that a continuous signal can be perfectly reconstructed from its samples if the sampling frequency is greater than twice the highest frequency component of the signal.
Mathematically
Fs ≥ 2Fmax
Commonly used concept is the Nyquist frequency, which is defined as
This represents the highest frequency that can be represented without aliasing.
Why Sampling Is Necessary
Real-world signals are often continuous in time, meaning they exist at every moment. However, digital systems cannot process continuous signals directly. Instead, signals must be converted into a sequence of discrete samples.
Sampling is the process of measuring the signal amplitude at regular time intervals.
For example
The goal is to choose a sampling rate high enough to preserve the original signal content.
Continuous signal(assumed) and its sampled representation
The Nyquist Frequency
The Nyquist frequency represents the maximum frequency that can be captured without distortion.
If a signal contains frequencies higher than the Nyquist frequency, those components will fold back into the spectrum, producing incorrect frequency components.
For example
This means that if the signal contains frequencies above the Nyquist frequency, aliasing will occur.
Frequency spectrum showing components below the Nyquist limit
MALMIJAL example
Signal: 100 Hz + 700 Hz
Sampling rate: 2000 Hz → Nyquist frequency: 1000Hz
FFT display
What Is Aliasing?
Aliasing occurs when a signal contains frequency components that exceed the Nyquist frequency.
Instead of appearing at their true frequency, these components appear at a different frequency within the measurable range.
This phenomenon happens because the sampling process cannot distinguish between certain high-frequency components.
The aliasing frequency can be calculated as
Example of frequency folding caused by insufficient sampling rate
MALMIJAL example
Original signal: 700 Hz
Sampling rate: 1000 Hz → Nyquist frequency: 500Hz
FFT result showing alias around 300 Hz
Visualizing Aliasing
A useful way to understand aliasing is to compare two sampling scenarios.
When the sampling frequency is too low, the waveform still appears periodic, but it actually represents the wrong signal.
Same signal sampled at different rates demonstrating aliasing
MALMIJAL example
Signal: 400 Hz sine wave
Case 1: Fs = 2000 Hz
Case 2: Fs = 600 Hz
In this example, the signal frequency is 400 Hz, the Nyquist condition (400 < Fs/2 = 1000) is satisfied and the waveform is represented correctly. However, when the same signal is sampled at 600 Hz, the Nyquist frequency becomes 300 Hz, which is lower than the signal frequency.
As a result, aliasing occurs.
The aliasing frequency can be calculated as:
For this example:
This means that a 400 Hz signal sampled at 600 Hz will appear as a 200 Hz signal in the sampled data.
frequency domain. When the FFT is computed for the properly sampled signal, the spectrum shows a peak at 400 Hz. However, when the signal is sampled at 600 Hz, the FFT instead shows a peak at 200 Hz, which corresponds to the alias frequency predicted by the equation.
FFT comparison showing the original 400 Hz signal and its 200 Hz alias caused by undersampling
Nyquist Theorem in Real Engineering Systems
In real systems, engineers rarely use the Nyquist rate exactly.
Instead, the sampling rate is usually chosen higher than the theoretical limit
Common examples include
Anti-aliasing filter
Before sampling, analog signals are typically passed through a low-pass filter that removes frequencies above the Nyquist frequency.
This prevents aliasing from occurring during the sampling process.
Digital anti-aliasing filtering ensures that the signal satisfies the Nyquist condition before downsampling
(refer to Samples/anti-aliasing.mmj)
MALMIJAL example
Original signal
LPF applied
Downsampling result
Practical Rule Used by Engineers
In practice, engineers rarely place the cutoff frequency exactly at the Nyquist frequency.
Instead, the cutoff is usually set slightly below the Nyquist frequency
For example
This margin ensures that unwanted high-frequency components are sufficiently attenuated.
Summary
The Nyquist theorem is one of the most fundamental principles in digital signal processing.
It defines the minimum sampling rate required to preserve the frequency content of a signal.
Key takeaways
A signal must be sampled at at least twice its highest frequency
The Nyquist frequency is half of the sampling rate
Frequencies above the Nyquist limit cause aliasing
Anti-aliasing filters are used to prevent this problem
Practical systems use a margin below the Nyquist frequency
Understanding the Nyquist theorem is essential for designing reliable digital signal processing systems, from audio recording to advanced engineering measurements.
Suggested Further Reading
#You may also find these topics helpful: