RADE: A Neural Codec for Transmitting Speech over HF Radio Channels

2025-05-10

AI Summary

High-frequency (HF) wireless voice transmission faces severe challenges including low signal-to-noise ratio (SNR), multipath fading, and high peak-to-average power ratio (PAPR). Method: This paper proposes the first end-to-end neural speech codec jointly optimized for HF channels. Departing from conventional cascaded pipelines of speech coding, channel coding, and modulation, our architecture directly maps acoustic features to low-PAPR (<1 dB) continuous QAM symbols and employs a customized OFDM waveform for efficient over-the-air transmission. The method integrates neural autoencoders, vocoder-based feature modeling, differentiable modulation, and channel-aware joint training. Results: Evaluated on both simulated and real-world HF channels, the proposed system significantly improves speech intelligibility—especially under ultra-low SNR (<0 dB)—and outperforms state-of-the-art analog and digital radio systems across all metrics. It achieves, for the first time, simultaneous optimization of robustness, fidelity, and power efficiency in HF communications.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Natural Language Processing: Speech
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Speech compression is commonly used to send voice over radio channels in applications such as mobile telephony and two-way push-to-talk (PTT) radio. In classical systems, the speech codec is combined with forward error correction, modulation and radio hardware. In this paper we describe an autoencoder that replaces many of the traditional signal processing elements with a neural network. The encoder takes a vocoder feature set (short term spectrum, pitch, voicing), and produces discrete time, but continuously valued quadrature amplitude modulation (QAM) symbols. We use orthogonal frequency domain multiplexing (OFDM) to send and receive these symbols over high frequency (HF) radio channels. The decoder converts received QAM symbols to vocoder features suitable for synthesis. The autoencoder has been trained to be robust to additive Gaussian noise and multipath channel impairments while simultaneously maintaining a Peak To Average Power Ratio (PAPR) of less than 1~dB. Over simulated and real world HF radio channels we have achieved output speech intelligibility that clearly surpasses existing analog and digital radio systems over a range of SNRs.

Problem

• Develops neural autoencoder for HF radio speech transmission
• Replaces traditional signal processing with neural network
• Enhances speech intelligibility in noisy HF channels

Innovation

• Autoencoder replaces traditional signal processing elements
• Uses OFDM for QAM symbol transmission over HF radio
• Trained for robustness to noise and multipath impairments