FreeDV RADE: Digital Voice for HF with Enhanced Speech Quality

Demonstrates FreeDV RADE, an open-source digital voice mode for HF, which combines machine learning with classical Digital Signal Processing (DSP) to achieve improved speech quality and spectral efficiency. Released in June 2025 by ARDC grantee FreeDV, RADE operates at Signal-to-Noise Ratios (SNRs) as low as -2 dB and occupies approximately 1.5 kHz of bandwidth, half that of a typical SSB signal, while transmitting about 8 kHz of voice audio. The mode also features an optimized Peak-to-Average Power Ratio (PAPR) for efficient power amplifier use, reducing heat and consumption.

Users, such as VK3TPM, report significantly improved listening ease compared to analog modes, particularly in high HF noise environments, making it easier to conduct nets across various Australian states. The mode's integration into popular radios like the FlexRadio 6000 series allows direct operation without external computers. A C library for RADE is available, facilitating broader integration into other radios and SDR programs, with development continuing on FreeDV RADE v2 to further narrow bandwidth below 1 kHz.