Coherence bandwidth ($B_c$) is the frequency-domain dual of delay spread, quantifying the frequency interval over which the channel's transfer function remains highly correlated. A channel is considered flat fading if the transmitted signal bandwidth is significantly less than $B_c$, meaning all frequency components experience the same attenuation and phase shift. Conversely, when signal bandwidth exceeds $B_c$, the channel induces frequency-selective fading, distorting the waveform through inter-symbol interference.
Glossary
Coherence Bandwidth

What is Coherence Bandwidth?
Coherence bandwidth is the statistical measure of the range of frequencies over which a wireless channel passes all spectral components with approximately equal gain and linear phase, defining the maximum signal bandwidth that experiences flat rather than frequency-selective fading.
The most common engineering approximation defines coherence bandwidth as $B_c \approx 1/(5\sigma_\tau)$, where $\sigma_\tau$ is the RMS delay spread. This threshold ensures a correlation coefficient above 0.5 between frequency components. In OFDM systems, subcarrier spacing is deliberately designed to be narrower than $B_c$, ensuring each subcarrier experiences flat fading that can be corrected by a single-tap equalizer.
Key Characteristics of Coherence Bandwidth
Coherence bandwidth is the statistical measure of the frequency range over which a wireless channel exhibits a correlated amplitude and linear phase response. It defines the boundary between flat and frequency-selective fading.
Inverse Relationship with Delay Spread
Coherence bandwidth (Bc) is fundamentally inversely proportional to the root-mean-square delay spread (στ). A channel with a small delay spread (e.g., a small office) has a large coherence bandwidth, while a channel with large delay spread (e.g., a hilly urban macrocell) has a small coherence bandwidth. The approximate relationship is often defined as:
- Bc ≈ 1 / (5 * στ) for 50% correlation
- Bc ≈ 1 / (50 * στ) for 90% correlation This means a delay spread of 1 µs yields a coherence bandwidth of roughly 200 kHz at the 50% correlation threshold.
Flat vs. Frequency-Selective Fading Boundary
Coherence bandwidth is the critical threshold that determines fading type:
- Flat Fading: Occurs when the signal bandwidth (Bs) is significantly less than Bc. All spectral components experience the same gain and linear phase shift, preserving the signal's spectral shape.
- Frequency-Selective Fading: Occurs when Bs > Bc. Different frequency components experience uncorrelated amplitude and phase distortion, causing inter-symbol interference (ISI). For robust digital design, a channel is considered flat if Bs < Bc / 10.
Correlation Threshold Definitions
Coherence bandwidth is not a single absolute value but is defined by a chosen correlation coefficient threshold between the complex channel gains at two frequencies:
- 50% Coherence Bandwidth (Bc,50): The frequency separation at which the envelope correlation drops to 0.5. This is a looser bound.
- 90% Coherence Bandwidth (Bc,90): The separation at which correlation drops to 0.9. This is a stricter bound, typically an order of magnitude smaller than Bc,50. Engineers select the threshold based on the required fidelity for their specific modulation scheme.
OFDM Subcarrier Spacing Design Rule
In Orthogonal Frequency Division Multiplexing (OFDM) systems like LTE and 5G NR, the subcarrier spacing (Δf) is deliberately designed to be much smaller than the coherence bandwidth. This ensures that each narrowband subcarrier experiences flat fading, simplifying equalization to a single-tap complex multiplication per subcarrier.
- LTE uses Δf = 15 kHz, targeting coherence bandwidths typical of pedestrian environments.
- 5G NR supports scalable numerologies (15, 30, 60 kHz) to adapt to varying delay spreads from indoor to high-mobility scenarios.
Frequency Correlation Function
The coherence bandwidth is derived from the frequency correlation function, which is the Fourier transform of the multipath intensity profile (power delay profile). As frequency separation (Δf) increases, the correlation between the channel response at f and f + Δf decays. The coherence bandwidth quantifies the Δf at which this correlation falls below a specified threshold, directly linking the channel's time-dispersion characteristics to its frequency-domain behavior.
Pilot Symbol Density in Channel Estimation
Coherence bandwidth dictates the minimum density of pilot symbols required for accurate channel estimation. Pilots must be inserted in the time-frequency grid at intervals smaller than both the coherence bandwidth and coherence time to satisfy the 2D Nyquist sampling theorem.
- In the frequency domain, pilot spacing must be less than Bc.
- If pilots are spaced too far apart (greater than Bc), the interpolated channel estimate between them becomes unreliable, degrading equalization and increasing bit error rate.
Frequently Asked Questions
Clear, technically precise answers to the most common questions about coherence bandwidth, its relationship to delay spread, and its critical role in OFDM system design and channel estimation.
Coherence bandwidth (Bc) is the range of frequencies over which the wireless channel's frequency response is considered highly correlated or approximately flat, meaning two sinusoids separated by less than Bc will experience strongly correlated amplitude fading. It is formally defined as the frequency separation at which the channel's frequency correlation function drops below a specified threshold, typically 0.5 or 0.9. The coherence bandwidth is inversely proportional to the channel's delay spread (τ_rms): Bc ≈ 1/(k·τ_rms), where k is a constant depending on the correlation threshold (k≈50 for 0.9 correlation, k≈5 for 0.5 correlation). This parameter fundamentally determines whether a transmitted signal experiences flat fading or frequency-selective fading, making it a cornerstone metric for physical layer design in wideband systems.
Enabling Efficiency, Speed & Accuracy
Intelligent Analysis, Decision & Execution
We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.
Talk to Us
Search across company data
Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.
Useful when people spend too long searching or get different answers from different systems.

Automate internal workflows
Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.
Useful when repetitive work moves across multiple tools and teams.

Add AI to products and internal tools
Build assistants, guided actions, or decision support into the software your team or customers already use.
Useful when AI needs to be part of the product, not a separate tool.
Related Terms
Coherence bandwidth is fundamentally linked to the channel's time dispersion and multipath structure. These related terms define the physical phenomena and statistical models that determine the frequency selectivity of a wireless channel.
Delay Spread
The root-mean-square (RMS) delay spread is the primary physical parameter that determines coherence bandwidth. It quantifies the time dispersion of multipath components arriving at the receiver.
- Inverse relationship: Coherence bandwidth (Bc) ≈ 1 / (k × τ_rms), where k is a constant (typically 5-50 depending on correlation threshold)
- Typical values: Indoor environments exhibit delay spreads of 10-100 ns, while outdoor macrocellular environments can reach 10-20 μs
- Measurement: Derived from the power delay profile (PDP) of the channel impulse response
A larger delay spread indicates more severe time dispersion, resulting in a narrower coherence bandwidth and greater frequency selectivity.
Flat vs. Frequency-Selective Fading
Coherence bandwidth defines the boundary between two fundamental fading regimes that dictate how a signal is distorted.
- Flat fading: Occurs when signal bandwidth < coherence bandwidth. All frequency components experience correlated fading with minimal distortion. The received signal simply fluctuates in amplitude.
- Frequency-selective fading: Occurs when signal bandwidth > coherence bandwidth. Different spectral components fade independently, causing intersymbol interference (ISI) and severe linear distortion.
This classification directly determines the complexity of equalization required at the receiver. Wideband systems like LTE and 5G NR intentionally operate in frequency-selective regimes and use OFDM to convert the wideband channel into multiple parallel flat-fading subcarriers.
Power Delay Profile
The power delay profile (PDP) is the empirical measurement from which both delay spread and coherence bandwidth are derived. It represents received signal power as a function of propagation delay.
- Structure: Consists of distinct multipath clusters arriving at different delays, each with its own amplitude and phase
- Derivation: The Fourier transform of the PDP's autocorrelation yields the frequency correlation function, from which coherence bandwidth is directly computed
- Standard thresholds: Coherence bandwidth is typically defined at correlation levels of 0.9 (high correlation) or 0.5 (moderate correlation)
PDP measurements are fundamental to channel sounding campaigns and serve as input to stochastic channel models like the 3GPP spatial channel model (SCM).
OFDM and Subcarrier Design
Orthogonal Frequency Division Multiplexing (OFDM) is the primary modulation technique designed around coherence bandwidth constraints. It partitions a wideband frequency-selective channel into many narrowband flat-fading subchannels.
- Subcarrier spacing: Must be significantly smaller than the coherence bandwidth to ensure each subcarrier experiences flat fading
- Cyclic prefix: A guard interval longer than the maximum excess delay of the channel eliminates ISI between OFDM symbols
- Pilot placement: Reference signals are inserted at intervals in both time and frequency, with frequency-domain spacing determined by coherence bandwidth to enable accurate channel estimation
5G NR supports scalable numerology with subcarrier spacings from 15 kHz to 240 kHz, allowing adaptation to diverse delay spread environments from rural macrocells to indoor millimeter-wave.
Frequency Correlation Function
The frequency correlation function is the mathematical tool that directly quantifies coherence bandwidth. It measures the normalized correlation between the complex channel gain at two frequencies separated by Δf.
- Definition: R_H(Δf) = E[H(f) · H*(f + Δf)], where H(f) is the channel transfer function
- Coherence bandwidth extraction: Bc is the value of Δf at which |R_H(Δf)| drops below a specified threshold (commonly 0.5 or 0.9)
- Relationship to PDP: By the Wiener-Khinchin theorem, the frequency correlation function is the Fourier transform of the power delay profile
This function is essential for designing pilot symbol patterns in OFDM systems, ensuring that reference signals are spaced closely enough in frequency to accurately reconstruct the channel response across all subcarriers.
Channel Sounding
Channel sounding is the measurement campaign that empirically determines coherence bandwidth and other channel parameters for a specific environment. It involves transmitting known probing signals and analyzing the received waveform.
- Techniques: Include swept-frequency chirps, pseudo-noise sequences, and multi-tone OFDM probes
- Outputs: Produces the time-variant channel impulse response, from which PDP, delay spread, Doppler spread, and coherence bandwidth are extracted
- Applications: Essential for validating ray-tracing simulations, calibrating digital twins, and parameterizing stochastic channel models
Modern channel sounders use software-defined radios (SDRs) and GPU-accelerated post-processing to capture wideband measurements with high delay resolution, enabling accurate coherence bandwidth estimation even in highly dynamic vehicular or high-speed rail environments.

About the author
Prasad Kumkar
CEO & MD, Inference Systems
Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.
His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.
Partnered with leading AI, data, and software stack.
How We Work
Custom AI workflows for your Business
One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.
01
Review the use case
We understand the task, the users, and where AI can actually help.
Read more02
Pick the right approach
We define what needs search, automation, or product integration.
Read more03
Build the first useful version
We implement the part that proves the value first.
Read more04
Improve from there
We add the checks and visibility needed to keep it useful.
Read moreThe first call is a practical review of your use case and the right next step.
Talk to Us