Inferensys

Glossary

OFDM Numerology

OFDM numerology is the set of scalable physical-layer parameters in 5G NR, including subcarrier spacing and cyclic prefix length, that define the frame structure for different frequency ranges and use cases.
Strategy consultant facilitating AI use case discovery workshop, sticky notes on glass wall, casual corporate meeting.
5G NR PHYSICAL LAYER STRUCTURE

What is OFDM Numerology?

OFDM numerology defines the scalable set of physical-layer parameters—specifically subcarrier spacing and cyclic prefix duration—that determine the frame structure of a 5G New Radio (NR) waveform for different frequency ranges and deployment scenarios.

OFDM numerology is the configuration of subcarrier spacing (SCS) and cyclic prefix (CP) length that defines the fundamental time-frequency grid of a 5G NR transmission. Unlike LTE's fixed 15 kHz SCS, 5G NR introduces a scalable numerology with SCS values derived by multiplying 15 kHz by powers of two (15, 30, 60, 120, and 240 kHz), enabling the waveform to adapt to diverse spectrum allocations from sub-1 GHz to millimeter-wave bands.

The choice of numerology directly governs the slot duration, symbol length, and cyclic prefix overhead, creating a trade-off between latency and spectral efficiency. Wider subcarrier spacings produce shorter symbol periods, reducing latency for ultra-reliable low-latency communication (URLLC) but increasing CP overhead. Narrower spacings maximize efficiency for enhanced mobile broadband (eMBB) in macro-cell deployments. Multiple numerologies can coexist on the same carrier through bandwidth parts (BWPs).

SCALABLE OFDM PARAMETER SETS

5G NR Numerology Configurations (μ Values)

Comparison of physical-layer parameters for each 5G NR numerology index (μ), defining subcarrier spacing, slot duration, and cyclic prefix characteristics across frequency ranges and use cases.

Parameterμ = 0μ = 1μ = 2μ = 3μ = 4

Subcarrier Spacing (Δf)

15 kHz

30 kHz

60 kHz

120 kHz

240 kHz

OFDM Symbol Duration (useful part)

66.67 μs

33.33 μs

16.67 μs

8.33 μs

4.17 μs

Slot Duration

1 ms

0.5 ms

0.25 ms

0.125 ms

0.0625 ms

Slots per Subframe (1 ms)

1

2

4

8

16

Slots per Radio Frame (10 ms)

10

20

40

80

160

OFDM Symbols per Slot

14

14

14

14

14

Normal CP Length (symbol 0)

5.2 μs

2.86 μs

1.69 μs

1.11 μs

0.81 μs

Normal CP Length (symbols 1-6)

4.69 μs

2.34 μs

1.17 μs

0.59 μs

0.29 μs

Maximum Carrier Bandwidth

50 MHz

100 MHz

200 MHz

400 MHz

400 MHz

Applicable Frequency Range

FR1 (sub-6 GHz)

FR1 (sub-6 GHz)

FR1 & FR2

FR2 (mmWave)

FR2 (mmWave)

Extended CP Supported

Typical Use Case

Wide-area coverage, LTE coexistence

Urban macro, enhanced MBB

Dense urban, low-latency URLLC

mmWave hotspots, fixed wireless

Ultra-low latency, short-range

Phase Noise Sensitivity

Low

Low

Moderate

High

Very High

Doppler Resilience

Low

Moderate

High

Very High

Very High

PHYSICAL LAYER PARAMETERIZATION

How OFDM Numerology Scaling Works

OFDM numerology defines the scalable physical-layer parameters—subcarrier spacing and cyclic prefix duration—that allow a unified waveform to adapt to diverse 5G NR frequency ranges and service requirements.

OFDM numerology scaling is the mechanism by which 5G NR adapts a single waveform framework across sub-1 GHz to millimeter-wave bands. The fundamental scaling principle is $\Delta f = 2^\mu \cdot 15$ kHz, where $\mu \in {0,1,2,3,4}$ is the numerology index. As subcarrier spacing doubles, the OFDM symbol duration halves proportionally, preserving the time-frequency resource grid structure while enabling latency reduction for higher frequencies.

Each numerology pairs a specific subcarrier spacing with a corresponding cyclic prefix (CP) length to maintain multipath resilience. Wider subcarrier spacings use shorter CPs, suited for small-cell deployments with limited delay spread. This scaling also determines the slot duration—from 1 ms at 15 kHz to 0.0625 ms at 240 kHz—allowing the frame structure to support both enhanced mobile broadband and ultra-reliable low-latency communication services within a single radio interface.

5G NR FRAME STRUCTURE

Key Characteristics of OFDM Numerology

OFDM numerology defines the scalable physical-layer parameters—subcarrier spacing and cyclic prefix—that adapt the 5G NR air interface to diverse spectrum and use cases.

OFDM NUMEROLOGY

Frequently Asked Questions

Clear answers to common questions about the scalable physical-layer parameters that define 5G NR frame structures, subcarrier spacing, and cyclic prefix configurations.

OFDM numerology defines the set of scalable physical-layer parameters—primarily subcarrier spacing (SCS) and cyclic prefix (CP) duration—that determine the frame structure of an Orthogonal Frequency-Division Multiplexing transmission. In 5G New Radio (NR), numerology is the foundational design principle that enables a single air interface to support diverse use cases across vastly different frequency ranges. Unlike 4G LTE, which uses a fixed 15 kHz SCS, 5G NR introduces a scalable numerology indexed by μ (mu), where SCS = 15 × 2^μ kHz. This scaling allows the OFDM symbol duration to halve with each increment of μ, making the waveform adaptable: lower numerologies (μ=0, 15 kHz) provide long symbol durations ideal for wide-area coverage, while higher numerologies (μ=3, 120 kHz) produce short symbols suited for low-latency applications in millimeter-wave bands. The cyclic prefix scales inversely, maintaining a consistent CP overhead of approximately 7% per slot. This parametric flexibility is what allows 5G NR to simultaneously serve enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC), and massive machine-type communications (mMTC) on a unified waveform.

Prasad Kumkar

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.