What Is Noise Bandwidth?

Noise bandwidth, also called equivalent noise bandwidth (ENBW), is a measure of the frequency range over which a filter or receiver system admits thermal noise power. It is defined as the bandwidth of a hypothetical ideal rectangular filter that would pass the same total noise power as the actual filter when driven by white Gaussian noise. Because real filters do not have perfectly sharp rolloff characteristics, the noise bandwidth differs from the more commonly cited 3 dB bandwidth, and it is the noise bandwidth that correctly determines the noise floor of a system.

The concept is foundational in receiver design, signal processing, and communications engineering. Thermal noise power is proportional to both bandwidth and temperature, following the relation P = kTB, where k is Boltzmann's constant, T is the absolute temperature of the system, and B is the noise bandwidth. A narrower noise bandwidth lowers the noise floor and enables greater receiver sensitivity, which is why the selection of filter characteristics is one of the primary design levers in radio-frequency, microwave, and analog signal chain engineering.

Mathematical Basis

For a filter with a power transfer function H(f), the equivalent noise bandwidth is computed by integrating the transfer function over all frequencies and normalizing by the peak value: ENBW = (1/|H(fmax)|^2) * integral of |H(f)|^2 df. This integral can be evaluated analytically for common filter topologies such as Butterworth, Chebyshev, and Bessel designs, each of which yields a characteristic ratio between the noise bandwidth and the 3 dB bandwidth. For a first-order RC lowpass filter, the noise bandwidth is exactly (pi/2) times the 3 dB bandwidth. For higher-order Butterworth filters, the ratio converges toward unity as order increases because the rolloff steepens and the passband approximates a rectangle more closely. ScienceDirect's overview of noise equivalent bandwidth provides tabulated ratios for standard analog filter families.

Noise Bandwidth in Receiver Design

Receiver sensitivity is a function of the thermal noise floor, the receiver noise figure, and the required carrier-to-noise ratio for a given modulation scheme. These factors combine in the sensitivity equation: Sensitivity = 10 log(kTB) + NF + C/N, where B is the equivalent noise bandwidth. As described in technical coverage from High Frequency Electronics, a receiver with a 10 kHz ENBW operating at room temperature has a thermal noise floor near -134 dBm; adding a 5 dB noise figure and a 4 dB carrier-to-noise requirement sets the reference sensitivity at -125 dBm. Designers reduce noise bandwidth by selecting filters with tighter passband characteristics, but must weigh this against group delay distortion and inter-symbol interference in digital communications systems.

Measurement and Computation

In practice, the ENBW of a realized filter is determined by numerically integrating measured frequency response data. RF engineers convert S-parameter measurements from logarithmic dB scale to linear power before integration, then apply the normalization formula. Digital filter implementations in software-defined radio or digital signal processing chains use the filter coefficient response to compute ENBW analytically. The NIST technical note on signal-to-noise ratios in RF systems documents measurement methodology relevant to system-level noise bandwidth assessment.

Applications

Noise bandwidth has applications in a range of fields, including:

• Radio-frequency receiver design and sensitivity specification
• Analog and digital filter selection for communications hardware
• Radar signal processing and clutter rejection analysis
• Wireless communications systems and spectrum management
• Instrumentation design for low-noise signal acquisition