What Are Notch Filters?

Notch filters are a class of band-stop filters designed to suppress a narrow band of frequencies while leaving the rest of the spectrum essentially unaltered. Where a conventional band-stop filter attenuates a broad range of frequencies, a notch filter targets a single frequency or a very tight band, producing a sharp "notch" in the frequency response. The depth of that notch and the width of the attenuated region are the two principal design parameters, and the balance between them determines how well the filter performs in practice.

The concept draws from classical filter theory, which traces through Butterworth, Chebyshev, and elliptic designs, but notch filters occupy a specific niche: selective interference rejection. They appear in both analog circuits and digital signal processing (DSP) implementations, and the same mathematical framework applies in both domains. The IEEE has published extensively on notch filter design, including feedback-based architectures explored in narrowband notch filter research on IEEE Xplore.

Transfer Function and Quality Factor

A notch filter's frequency response is characterized by its center frequency, its attenuation depth at that frequency, and its quality factor Q. The Q factor is the ratio of the center frequency to the bandwidth of the notch; a high-Q filter produces a very narrow notch, while a low-Q filter attenuates a wider band. For interference rejection applications, a high Q is desirable because it minimizes distortion to nearby frequencies, but high-Q analog designs are sensitive to component tolerances and can drift with temperature. In digital implementations, the Q can be held precisely by coefficient selection and is not subject to the same aging effects.

The transfer function of a second-order notch filter in the z-domain places a pair of zeros on the unit circle at the target frequency, with poles positioned just inside the unit circle at the same angle to control bandwidth. This pole-zero placement is a direct expression of the Q: moving the poles closer to the zeros narrows the notch and sharpens the selectivity.

Filter Topologies

On the analog side, the twin-T network is the most widely used passive topology. Two T-shaped RC networks connected in parallel create cancellation at the target frequency when component values are matched precisely. The twin-T suffers from limited Q without additional circuitry, so active versions using operational amplifiers feed a portion of the output back to the filter to increase selectivity. Wien-bridge and state-variable active filter designs offer similar functionality with better tunability.

In digital systems, notch filters are commonly implemented as infinite impulse response (IIR) filters. A second-order IIR notch filter requires only a few multiplications and additions per sample, making it computationally efficient even on constrained hardware. Field-programmable gate array (FPGA) implementations have demonstrated high-speed operation suitable for real-time signal conditioning, as shown in high-speed IIR notch filter work published through IEEE. Finite impulse response (FIR) notch filters are also used when linear phase response is required, though they demand substantially more coefficients to achieve comparable attenuation.

Adaptive Notch Filters

When the interference frequency is unknown or varies over time, adaptive notch filters estimate and track the notch center frequency automatically. The least mean squares (LMS) algorithm and recursive least squares (RLS) are both applied to this problem; the filter coefficients update continuously based on the error between the filter output and a reference signal. Adaptive designs are common in active noise control, echo cancellation, and applications where power-line frequency deviates slightly from its nominal 50 or 60 Hz value. The ScienceDirect overview of notch filter applications documents uses across biomedical, communications, and control domains.

Applications

Notch filters have applications in a wide range of fields, including:

• Biomedical signal processing, where 50/60 Hz power-line noise is removed from ECG and EEG recordings
• Audio engineering, for eliminating hum, feedback tones, or narrowband interference
• Industrial control systems, for suppressing resonance at mechanical natural frequencies
• Wireless communications receivers, for rejecting a known interferer at a fixed frequency
• Image and video processing, for removing periodic noise patterns in scanned or transmitted images