Glossary

Active Constellation Extension (ACE)

A distortionless peak-to-average power ratio (PAPR) reduction method that intelligently extends outer constellation points outward within acceptable error vector magnitude (EVM) margins to create peak-canceling signals without introducing spectral regrowth.

Get in touch Learn more

Engineer reviewing vector database search results on laptop, embeddings visualization on screen, home office coding session.

PAPR Reduction Technique

What is Active Constellation Extension (ACE)?

Active Constellation Extension (ACE) is a distortionless peak-to-average power ratio (PAPR) reduction technique that intelligently shifts outer constellation points outward within acceptable error vector magnitude (EVM) tolerances. Unlike clipping-based methods, ACE exploits the decision region margins of higher-order quadrature amplitude modulation (QAM) schemes to generate a peak-canceling signal without increasing bit error rate (BER) or introducing in-band distortion.

The algorithm operates iteratively in the time domain, projecting clipped signal peaks back onto allowable constellation extension regions defined by the modulation format's outer boundaries. Because ACE only modifies symbols in directions that increase their distance from decision thresholds, it achieves crest factor reduction while preserving spectral mask compliance and avoiding the spectral regrowth typically associated with nonlinear distortion techniques.

Distortionless PAPR Reduction

Key Characteristics of ACE

Active Constellation Extension (ACE) is a crest factor reduction technique that intelligently manipulates outer constellation points within acceptable EVM limits to generate peak-canceling signals without introducing spectral regrowth.

Distortionless Operation

Unlike clipping-based CFR methods, ACE operates without introducing in-band distortion. It exploits the decision boundaries of the modulation constellation, moving outer points outward into regions that do not increase symbol error probability. This preserves the integrity of the transmitted data while reducing the peak-to-average power ratio.

Constellation Extension Regions

ACE defines extension regions for each outer constellation point—areas in the complex plane where a point can be moved without crossing a decision boundary. For QPSK, these are the quadrants beyond the nominal point. For higher-order QAM, only corner and edge points have valid extension regions. Points are projected onto these regions iteratively to construct a peak-canceling signal.

Iterative Projection Algorithm

The core ACE algorithm operates through iterative clipping and projection:

Clip signal peaks exceeding a target threshold
Project clipped samples back onto valid constellation extension regions
Filter to remove out-of-band components
Repeat until PAPR target is met or convergence is achieved

This projection-onto-convex-sets (POCS) approach guarantees convergence to a feasible solution.

Zero Spectral Regrowth

Because ACE only modifies in-band constellation points and applies frequency-domain filtering after each iteration, it introduces no out-of-band spectral components. This makes it ideal for systems operating under strict spectral masks where adjacent channel leakage must be minimized. The trade-off is a controlled increase in EVM, bounded by the constellation extension margins.

EVM-Constrained Optimization

ACE operates within a predefined EVM budget. The maximum extension distance for each constellation point is bounded by the acceptable error vector magnitude. This creates a convex feasible region that guarantees the transmitted signal remains demodulable. System designers can tune the EVM constraint to balance PAPR reduction against modulation fidelity requirements.

OFDM System Integration

ACE is particularly effective in OFDM systems where high PAPR is inherent. It can be applied per-symbol or across multiple symbols. Modern implementations combine ACE with tone reservation (TR) for additional PAPR reduction: ACE handles outer constellation points while reserved tones carry supplementary peak-canceling signals, maximizing reduction without exceeding EVM limits.

ACTIVE CONSTELLATION EXTENSION

Frequently Asked Questions

Clear answers to common questions about Active Constellation Extension (ACE), a distortionless PAPR reduction technique that manipulates outer constellation points to create peak-canceling signals without introducing spectral regrowth.

Active Constellation Extension (ACE) is a distortionless peak-to-average power ratio (PAPR) reduction technique that intelligently extends outer constellation points outward within acceptable error vector magnitude (EVM) margins to generate peak-canceling signals. Unlike clipping-based methods, ACE exploits the fact that outer constellation points can be moved away from decision boundaries without increasing symbol error rate. The algorithm iteratively identifies time-domain signal peaks exceeding a target threshold, then projects the corresponding frequency-domain symbols outward in the constellation space—effectively adding a corrective signal that reduces peaks while preserving the original data. Because the modifications remain within the valid decision region for each symbol, ACE achieves PAPR reduction without introducing in-band distortion or spectral regrowth, making it particularly valuable for OFDM systems where regulatory spectral masks are stringent.

PAPR REDUCTION METHOD COMPARISON

ACE vs. Other PAPR Reduction Techniques

Comparative analysis of Active Constellation Extension against alternative peak-to-average power ratio reduction methods for OFDM systems, evaluating distortion characteristics, spectral regrowth impact, and implementation complexity.

Feature	Active Constellation Extension (ACE)	Clipping & Filtering	Tone Reservation (TR)	Companding
Distortion Type	Distortionless (outer constellation points only)	In-band clipping distortion	Distortionless (reserved tones)	Non-uniform distortion across all amplitudes
Spectral Regrowth	None (no out-of-band components generated)	Significant (requires iterative filtering)	None (reserved tones outside data band)	Moderate (companding noise spreads spectrally)
EVM Impact	Controlled (within acceptable margin)	Degraded (clipping threshold dependent)	None (data subcarriers unaffected)	Degraded (expansion noise at receiver)
PAPR Reduction Capability	3-5 dB	5-8 dB	4-6 dB	4-7 dB
Computational Complexity	Moderate (convex optimization per symbol)	Low (amplitude thresholding)	High (peak-canceling signal optimization)	Low (lookup table or simple function)
Compatibility with DPD	Excellent (no distortion to linearize)	Moderate (clipping complicates DPD training)	Good (linear operation maintained)	Poor (companding nonlinearity conflicts with DPD)
Side Information Required
Power Efficiency Gain	15-25%	20-30%	15-25%	18-28%

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.

Enterprise searchRAGPermissions

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.

AI agentsWorkflow automationGovernance

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.

AI integrationDecision supportModel routing

SPECTRAL REGROWTH MITIGATION

Related Terms

Explore the key concepts and complementary techniques that work alongside Active Constellation Extension to manage signal peaks and ensure spectral compliance.

Crest Factor Reduction (CFR)

A foundational signal conditioning technique that reduces the Peak-to-Average Power Ratio (PAPR) before amplification. Unlike ACE, which operates on constellation points, CFR typically applies peak windowing or clipping and filtering directly to the time-domain waveform. This allows power amplifiers to operate at higher average power without inducing severe spectral regrowth, though it introduces in-band distortion that must be managed.

3-6 dB

Typical PAPR Reduction

Tone Reservation (TR)

A distortionless PAPR reduction method that reserves a subset of OFDM subcarriers specifically for peak-canceling signals. Like ACE, TR avoids in-band data distortion. However, instead of moving outer constellation points, it dedicates reserved tones to generate a corrective time-domain pulse. This technique is highly effective but comes at the cost of reduced data throughput due to the sacrificed subcarriers.

Error Vector Magnitude (EVM)

The critical modulation quality metric that defines the operational boundary for ACE. EVM measures the vector difference between ideal and actual symbol locations. ACE intelligently exploits the EVM margin—the acceptable error budget—by extending outer constellation points outward. This increases peak-canceling capability without violating the transmitter's modulation accuracy requirements.

< 3.5%

5G NR 256-QAM EVM Limit

Clipping Distortion

The primary nonlinear impairment that ACE seeks to avoid. Hard clipping abruptly truncates waveform peaks, generating severe intermodulation products and spectral regrowth. ACE provides an intelligent alternative by pre-distorting the signal in the constellation domain, effectively creating a 'soft' peak reduction that maintains spectral containment without the harsh out-of-band emissions associated with direct amplitude clipping.

Adjacent Channel Leakage Ratio (ACLR)

The ultimate regulatory compliance metric that ACE is designed to protect. ACLR quantifies the power leaking into adjacent channels due to nonlinear amplification. By reducing PAPR without introducing spectral regrowth, ACE directly improves ACLR performance. This allows transmitters to operate closer to the 1dB compression point of the power amplifier, maximizing efficiency while passing stringent emission mask tests.

-45 dBc

Typical 3GPP ACLR Limit

Iterative Clipping and Filtering (ICF)

A repeated signal conditioning process that alternates between time-domain clipping and frequency-domain filtering. While effective at reducing PAPR, ICF can suffer from peak regrowth after filtering. ACE is often compared against ICF as a more elegant, single-shot solution that operates directly on the modulation constellation, avoiding the iterative convergence delays and potential for spectral leakage inherent in repeated clipping cycles.

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.

Limited slotsGet a Free AI Consultation

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.