A Demodulation Reference Signal (DMRS) is a UE-specific or cell-specific pilot sequence transmitted within the resource block grid to enable coherent demodulation of the associated data channel. Unlike broadcast reference signals, DMRS is precoded identically to the data layers it supports, allowing the receiver to estimate the effective channel experienced by the transmission without explicit precoding matrix indicator feedback.
Glossary
Demodulation Reference Signal (DMRS)

What is Demodulation Reference Signal (DMRS)?
A Demodulation Reference Signal (DMRS) is a known pilot sequence embedded within a specific resource block allocation to provide the phase and amplitude reference required for coherent demodulation of the associated physical data channel.
In 5G NR and LTE, DMRS is multiplexed in the time-frequency domain with specific patterns defined by the mapping type and number of antenna ports. The receiver performs channel estimation by comparing the received DMRS symbols against the known Zadoff-Chu or Gold sequence, deriving amplitude and phase corrections to equalize the data resource elements and recover the transmitted symbols.
Key Characteristics of DMRS
The Demodulation Reference Signal (DMRS) is a UE-specific or cell-specific pilot sequence embedded within the resource block allocation to provide the phase and amplitude reference essential for coherent demodulation of the associated data channel.
Channel Estimation Foundation
DMRS provides the complex channel coefficient estimate for each allocated resource element. The receiver performs interpolation between DMRS positions in the time-frequency grid to reconstruct the channel response across the entire allocation. Without accurate DMRS-based estimation, higher-order modulations like 256QAM cannot be reliably decoded due to their tight error vector magnitude (EVM) requirements.
Precoded vs. Non-Precoded DMRS
In 5G NR, DMRS can be transmitted with the same precoding applied to the data layer, enabling transparent channel estimation without the UE needing to know the precoding matrix. This contrasts with cell-specific reference signals (CRS) in LTE, which are non-precoded and require the UE to reconstruct the effective channel from separate precoding matrix indicator (PMI) feedback.
Configurable Time-Domain Density
DMRS supports multiple mapping types to balance channel estimation accuracy against pilot overhead:
- Single-symbol DMRS: Occupies 1 OFDM symbol per slot, suitable for low-Doppler scenarios
- Double-symbol DMRS: Uses 2 adjacent OFDM symbols, supporting up to 12 orthogonal antenna ports via length-2 orthogonal cover codes (OCC)
- Additional DMRS positions: Up to 3 extra DMRS instances per slot for high-mobility use cases above 500 km/h
Gold Sequence Generation
The DMRS sequence is a pseudo-random Gold sequence initialized with a seed that depends on:
- Scrambling ID (SCID): 0 or 1, configured by higher layers
- Cell ID: The physical-layer cell identity
- Slot number and OFDM symbol index within the radio frame This ensures low cross-correlation between co-channel cells and enables MU-MIMO pairing with quasi-orthogonal reference signals.
DMRS Port Multiplexing
Multiple antenna ports are multiplexed onto the same time-frequency resources using:
- Frequency-domain OCC (FD-OCC): Length-2 codes applied across adjacent subcarriers for comb-based DMRS patterns
- Time-domain OCC (TD-OCC): Length-2 codes across consecutive OFDM symbols
- Code Division Multiplexing (CDM) groups: Sets of ports sharing the same REs but separated by orthogonal codes This enables up to 8 layers for SU-MIMO and 12 orthogonal ports for MU-MIMO in 5G NR.
DMRS Configuration Types
5G NR defines two DMRS mapping types for the PDSCH and PUSCH:
- Type 1: Comb-based structure with 2 CDM groups, supporting up to 8 ports. Higher density per port, optimized for SU-MIMO
- Type 2: Non-comb structure with 3 CDM groups, supporting up to 12 ports. Lower density per port, optimized for MU-MIMO with many co-scheduled UEs The gNB selects the type based on the deployment scenario and UE capability.
DMRS vs. Other OFDM Reference Signals
Functional comparison of the Demodulation Reference Signal against other physical-layer reference signals in 5G NR and LTE OFDM systems.
| Feature | DMRS | PTRS | CSI-RS |
|---|---|---|---|
Primary Function | Coherent data demodulation | Phase noise compensation | Channel state information acquisition |
Associated Channel | PDSCH, PUSCH, PDCCH, PBCH | PDSCH, PUSCH | Downlink only |
Precoding | Same precoding as associated data | Same precoding as associated data | Transmitted without data precoding |
Time-Frequency Density | High (up to 4 symbols per slot) | Low (every 2 or 4 OFDM symbols) | Configurable (periodic, aperiodic, semi-persistent) |
UE-Specific Configuration | |||
Beam Management Support | |||
Frequency Domain Occupancy | Scheduled resource blocks only | Scheduled resource blocks only | Full or partial bandwidth |
Overhead (Typical) | 4.7% to 14.3% | 0.5% to 2.5% | 1% to 5% |
Frequently Asked Questions
Clear, technically precise answers to the most common questions about the Demodulation Reference Signal and its role in coherent data recovery in 4G LTE and 5G NR physical layers.
A Demodulation Reference Signal (DMRS) is a known pilot sequence embedded within a specific user's resource block allocation that provides the phase and amplitude reference required for coherent demodulation of the associated physical data channel. Unlike cell-wide reference signals, the DMRS is UE-specific and precoded with the same precoding matrix applied to the data layers, ensuring that the receiver can estimate the effective channel experienced by the transmission. The DMRS sequence is generated from a pseudo-random Gold sequence or a Zadoff-Chu sequence, initialized by parameters including the scrambling identity (SCID) and slot number. In 5G NR, the DMRS design is highly flexible, supporting front-loaded patterns for low latency and additional symbols for high-Doppler scenarios, enabling accurate channel estimation per resource block and per MIMO layer.
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
Key reference signals and estimation techniques that work alongside or provide context for the Demodulation Reference Signal in OFDM-based systems.
Phase Tracking Reference Signal (PTRS)
A 5G NR reference signal specifically designed to compensate for phase noise introduced by local oscillators at high carrier frequencies. Unlike DMRS, which provides a baseline channel estimate, PTRS is inserted with high density in the time domain to track rapid phase variations within a slot. It is primarily used in millimeter wave (FR2) deployments where oscillator imperfections become significant. PTRS configuration is tightly coupled to the scheduled MCS and bandwidth, with its presence signaled via RRC and DCI.
Channel State Information Reference Signal (CSI-RS)
A downlink reference signal in LTE and 5G NR used for channel state information acquisition rather than data demodulation. While DMRS provides UE-specific phase references for PDSCH decoding, CSI-RS enables the UE to measure and report Channel Quality Indicator (CQI), Precoding Matrix Indicator (PMI), and Rank Indicator (RI). CSI-RS supports multi-port configurations for beam management and can be configured with zero power to enable interference measurement in neighboring cells.
Sounding Reference Signal (SRS)
An uplink reference signal transmitted by the UE to enable the gNB to estimate the uplink channel quality across a wide bandwidth. Unlike DMRS, which is confined to the UE's allocated resource blocks, SRS can be configured to sound the entire carrier bandwidth. This information is critical for frequency-selective scheduling, uplink timing estimation, and reciprocity-based downlink precoding in TDD systems where the uplink channel estimate is used to derive downlink beamforming weights.
DMRS Configuration Type 1 vs Type 2
5G NR defines two DMRS mapping types that trade off multiplexing capacity against channel estimation density. Type 1 uses a comb-based structure with 2 CDM groups, supporting up to 4 orthogonal ports for single-symbol DMRS and 8 ports for double-symbol DMRS. Type 2 uses a frequency-domain OCC structure with 3 CDM groups, supporting up to 6 ports (single-symbol) or 12 ports (double-symbol). Type 2 is preferred for MU-MIMO scenarios requiring higher spatial multiplexing layers.
Precoding Matrix Indicator (PMI)
A feedback parameter reported by the UE to recommend a specific precoding matrix from a standardized codebook for downlink transmission. The gNB uses this recommendation to shape the transmitted signal toward the UE's spatial direction. DMRS undergoes the same precoding as the associated PDSCH data layers, enabling the UE to estimate the effective precoded channel directly without needing explicit knowledge of the precoding matrix. This transparent precoding is a fundamental principle of 5G NR's DMRS design.
Physical Resource Block (PRB) Bundling
A technique where the UE assumes that the same precoding is applied across a contiguous set of PRBs, allowing joint channel estimation across multiple DMRS instances. This improves estimation accuracy, especially at cell edges. The PRB bundle size is configured by RRC and can be set to wideband (entire allocation) or a specific number of PRBs (2 or 4). PRB bundling directly impacts DMRS processing by determining the interpolation window for channel estimation across frequency-domain resource elements.

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