A notch filter is a type of electronic filter that removes a very narrow band of frequencies from a signal while letting everything above and below that band pass through unchanged. Think of it as the opposite of a laser pointer: instead of picking out one frequency to keep, it picks out one frequency to eliminate. The specific frequency it targets is called the “notch frequency,” and the filter’s defining trait is how surgically narrow that removal is, often achieving 50 to 60 dB of attenuation (meaning the unwanted frequency is reduced to a tiny fraction of its original strength).
How a Notch Filter Works
Every electronic signal, whether it’s audio, a heartbeat recording, or a radio broadcast, is made up of many frequencies layered on top of each other. Sometimes one specific frequency is pure noise: a hum from a power line, a screech of feedback from a speaker, or interference from a nearby radio station. A notch filter zeros in on that one frequency and pulls it out, leaving the rest of the signal mostly intact.
What makes a notch filter different from a broader “band-stop” filter is width. A band-stop filter can block a wide range of frequencies, which is useful when the problem spans a large portion of the spectrum. A notch filter, by contrast, works on only a narrow slice. This precision is its greatest strength, but it also means notch filters aren’t the right tool when interference is spread across many frequencies. They’re built for situations where the problem has a known, fixed frequency.
Where Notch Filters Are Used
Medical Equipment
One of the most common places you’ll encounter a notch filter is inside an ECG or EEG machine. Electrical power lines radiate a constant hum at either 50 Hz (in most of Europe, Asia, and Africa) or 60 Hz (in North America and parts of South America). That hum gets picked up by the sensitive electrodes attached to a patient’s skin and shows up as wavy interference on the recording. A notch filter tuned to 50 or 60 Hz strips out the power-line noise so doctors can read the actual heart or brain activity underneath.
For human ECG recordings, this works well with no noticeable distortion. Interestingly, the same approach can cause problems in animal research. A study published in the Journal of Electrocardiology found that applying a 50/60 Hz notch filter to rat ECG recordings severely deformed the QRS complex, the sharp spike that represents each heartbeat. Rat hearts beat much faster than human hearts, so more of the signal’s energy sits near 50 or 60 Hz, making it vulnerable to the filter. It’s a good reminder that the same tool can behave very differently depending on what signal you feed it.
Live Sound and Audio
If you’ve ever heard a piercing squeal from a microphone pointed at a speaker, that’s acoustic feedback, sometimes called “howling.” It happens when a specific frequency bounces between the microphone and the speaker, amplifying itself in a loop. Sound engineers use notch filters to kill that exact frequency without dulling the rest of the audio. The alternative, simply turning down the overall volume, works but sacrifices sound quality across all frequencies.
Modern feedback suppression systems can detect the howling frequency automatically and apply a notch filter in real time. Research into segmented notch filtering for hearing aids has shown that these filters can achieve extremely deep attenuation at the problem frequency while producing a high-quality output waveform, making them especially promising for hearing aid development where both clarity and comfort matter.
Radio Receivers
In radio frequency engineering, notch filters help reject “image frequencies,” unwanted signals that sneak into a receiver at a predictable offset from the desired station. By placing a notch filter at the image frequency, designers can suppress that sideband noise and improve the receiver’s overall signal quality. This technique is used in everything from Wi-Fi chipsets operating at 2.4 GHz to traditional AM/FM radios.
Passive vs. Active Notch Filters
Notch filters come in two basic hardware flavors. Passive notch filters are built from resistors, capacitors, and sometimes inductors, with no external power source. The classic design is the “Twin-T” circuit, named for its two T-shaped networks of resistors and capacitors. Passive filters are simple and reliable, but inductors can be expensive when high accuracy or small physical size is required, and the filter’s performance is limited by the components themselves.
Active notch filters add an amplifying element, typically an operational amplifier (op amp), with resistors and capacitors in its feedback loop. The op amp eliminates the need for inductors entirely, which is a significant practical advantage. Active designs also offer better control over how deep and narrow the notch is, and they can be cascaded (chained together) to build more complex filtering behavior. Most modern notch filters in consumer and medical electronics are active designs.
In software, digital notch filters accomplish the same thing mathematically. A digital filter processes a stream of numbers representing the signal and calculates an output with the target frequency removed. Digital filters are common in EEG and ECG processing, audio software, and communications systems where the signal is already in digital form.
The Tradeoff: Phase Distortion
No filter is perfectly transparent to the frequencies it’s supposed to leave alone. Every notch filter introduces small shifts in the timing of nearby frequencies, an effect called phase distortion. For a simple audio application like killing feedback, this is rarely noticeable. But in scientific signal processing, where the precise timing between different frequency components carries meaning, phase shifts can be a real problem.
Research published in eNeuro found that filters cause time lags in the filtered signal, disrupting the timing relationship between different frequencies within the same recording and between different recordings. These time shifts typically increase with a steeper (higher-order) filter, and they get worse for frequencies closer to the notch. In practical terms, a very aggressive notch filter applied to a brain recording could make two neural events appear to happen at slightly different times than they actually did.
For most everyday applications, this isn’t something you need to worry about. But it explains why engineers choose the gentlest filter that gets the job done. A narrower, shallower notch introduces less phase distortion than a deep, steep one.
Key Specifications to Know
If you’re shopping for or designing a notch filter, three numbers matter most:
- Notch frequency: the center frequency the filter removes. This needs to match the interference you’re targeting, such as 50 Hz, 60 Hz, or whatever frequency is causing problems.
- Stopband attenuation: how much the filter reduces the target frequency, measured in decibels (dB). A typical specification is 50 to 60 dB, meaning the unwanted frequency is reduced by a factor of roughly 100,000 to 1,000,000 in power.
- Q factor (quality factor): how narrow the notch is. A higher Q means a tighter notch that affects fewer neighboring frequencies. A lower Q means a wider notch that’s easier to implement but removes more of the signal you might want to keep.
Passband ripple is also worth checking. This describes how much the filter disturbs frequencies outside the notch. A well-designed filter keeps passband ripple below about 0.5 to 1 dB, meaning the “good” parts of your signal stay essentially flat.