What Is an Anti-Aliasing Filter?
An anti-aliasing filter is a filter designed to prevent aliasing, a distortion that occurs when a signal is sampled or resampled without satisfying the Nyquist condition. Formally, its role is to limit the signal bandwidth so that it does not exceed half of the sampling rate (Nyquist frequency).
What is Aliasing?
Aliasing is a spectral folding phenomenon.
When a signal contains frequency components above the Nyquist frequency
- Those components do not disappear
- They are mapped (folded) into lower frequencies
- This produces false frequency components
Example: How Aliasing Occurs in the Signal
Consider a signal composed of two sine waves
100 Hz
700 Hz
with an original sampling rate of Fs = 2000 Hz
Both frequencies are below the Nyquist limit (1000 Hz), so they are represented correctly.
Without aliasing: 100Hz and 700Hz (refer to Samples/anti-aliasing.mmj)
However, if we downsample the signal to 1000 Hz, the new Nyquist frequency becomes 500 Hz
Now the 700 Hz component is above the new Nyquist limit.
Instead of disappearing, it folds back into the spectrum and appears as a false lower frequency (300Hz in this case).
This is aliasing.
Aliased waveform after downsampling (refer to Samples/anti-aliasing.mmj)
Aliasing:300Hz instead of 700Hz (refer to Samples/anti-aliasing.mmj)
Why Anti-Aliasing Is Necessary
Aliasing is irreversible distortion
- Once folding occurs, original frequency information is lost
- No post-processing can recover the true spectrum
Therefore, aliasing must be prevented before it happens, not corrected afterward.
Two Types of Anti-Aliasing Filters
1. Analog Anti-Aliasing Filter (Before ADC)
Location
- Applied to continuous-time (analog) signal
- Placed before Analog-to-Digital Converter (ADC)
Purpose
Features
- Implemented using analog circuits (RC, active filters, etc.)
- Cannot be perfectly sharp → requires margin (guard band)
2. Digital Anti-Aliasing Filter (Before Downsampling)
Location
- Applied to discrete-time signal
- Used before resampling (downsampling)
Purpose
- Adjust signal bandwidth to match the new (lower) sampling rate
- Safe resampling, not filtering itself
Features
- Implemented using digital filters (FIR, IIR)
- Cannot be perfectly sharp → requires margin (guard band)
Example: Apply Digital Anti-Aliasing Filter
Frequency response of digital anti-aliasing filter (refer to Samples/anti-aliasing.mmj)
Waveform of the signal after applying a digital anti-aliasing (low-pass) filter while keeping the original sampling rate
(refer to Samples/anti-aliasing.mmj)
Downsampled signal after applying an digital anti-aliasing filter (refer to Samples/anti-aliasing.mmj)
Key Takeaway
AA(Anti-aliasing) filters must be applied before any sampling or resampling stage
- Analog anti-aliasing filtering before ADC
- Digital anti-aliasing filtering before downsampling
| Point of View | Analog AA filter | Digital AA filter |
|---|
| Location | Before ADC | Before downsampling |
| Domain | Continuous-time | Discrete-time |
| Purpose | Prevent initial aliasing | Prevent resampling aliasing |
| Recoverability | Aliasing is irreversible | Preventable before downsampling |
| Design flexibility | Limited | Higher (FIR/IIR digital low-pass filters) |
Conclusions
An anti-aliasing filter is not defined by its shape (low-pass), but by its role in protecting the sampling theorem.
Therefore, anti-aliasing filters are often introduced as low-pass filters, but their real purpose is much more specific.
Without an anti-aliasing filter
high-frequency components fold into lower frequencies
the signal spectrum becomes distorted
important information may be corrupted
With an anti-aliasing filter
high-frequency components are removed
the signal satisfies the new Nyquist limit
downsampling can be performed safely
In practice, anti-aliasing filters exist in both analog and digital forms. Analog anti-aliasing filters are used before ADC, while digital anti-aliasing filters are used before resampling operations such as downsampling.
Suggested Further Reading
#You may also find these topics helpful:
What Is an Anti-Aliasing Filter?
An anti-aliasing filter is a filter designed to prevent aliasing, a distortion that occurs when a signal is sampled or resampled without satisfying the Nyquist condition. Formally, its role is to limit the signal bandwidth so that it does not exceed half of the sampling rate (Nyquist frequency).
What is Aliasing?
Aliasing is a spectral folding phenomenon.
When a signal contains frequency components above the Nyquist frequency
Example: How Aliasing Occurs in the Signal
Consider a signal composed of two sine waves
700 Hz
with an original sampling rate of Fs = 2000 Hz
Both frequencies are below the Nyquist limit (1000 Hz), so they are represented correctly.
Without aliasing: 100Hz and 700Hz (refer to Samples/anti-aliasing.mmj)
However, if we downsample the signal to 1000 Hz, the new Nyquist frequency becomes 500 Hz
Now the 700 Hz component is above the new Nyquist limit.
Instead of disappearing, it folds back into the spectrum and appears as a false lower frequency (300Hz in this case).
This is aliasing.
Why Anti-Aliasing Is Necessary
Aliasing is irreversible distortion
Therefore, aliasing must be prevented before it happens, not corrected afterward.
Two Types of Anti-Aliasing Filters
1. Analog Anti-Aliasing Filter (Before ADC)
Location
Purpose
Ensure that the input signal satisfies:
Bandwidth ≤ Fs / 2
Prevents aliasing during the sampling process itself
Features
2. Digital Anti-Aliasing Filter (Before Downsampling)
Location
Purpose
Features
Example: Apply Digital Anti-Aliasing Filter
Frequency response of digital anti-aliasing filter (refer to Samples/anti-aliasing.mmj)
Waveform of the signal after applying a digital anti-aliasing (low-pass) filter while keeping the original sampling rate
(refer to Samples/anti-aliasing.mmj)
Downsampled signal after applying an digital anti-aliasing filter (refer to Samples/anti-aliasing.mmj)
Key Takeaway
AA(Anti-aliasing) filters must be applied before any sampling or resampling stage
Conclusions
An anti-aliasing filter is not defined by its shape (low-pass), but by its role in protecting the sampling theorem.
Therefore, anti-aliasing filters are often introduced as low-pass filters, but their real purpose is much more specific.
Without an anti-aliasing filter
high-frequency components fold into lower frequencies
the signal spectrum becomes distorted
important information may be corrupted
With an anti-aliasing filter
high-frequency components are removed
the signal satisfies the new Nyquist limit
downsampling can be performed safely
In practice, anti-aliasing filters exist in both analog and digital forms. Analog anti-aliasing filters are used before ADC, while digital anti-aliasing filters are used before resampling operations such as downsampling.
Suggested Further Reading
#You may also find these topics helpful: