FreeDV and RADE: Open-Source HF Digital Voice Explained

FreeDV is an open-source digital voice system for HF radio. It lets you talk with another amateur using your own voice, converted by software into a narrow digital signal that your ordinary SSB transceiver transmits and receives. Instead of the hiss, fading and interference that make weak SSB tiring to copy, a FreeDV signal either decodes cleanly or does not decode at all, and when it does, the speech can sound remarkably clear. Better still, it requires no new radio: any SSB transceiver, a computer and a free program are enough.

The project has changed a great deal in recent years, with a new flagship mode called RADE that uses machine learning. This guide explains what FreeDV is, how its main modes work, what you need to get on the air, how to set it up, how it compares with other digital voice systems, and how to find someone to talk to. The details come from the project’s own pages, so check them for the latest version and instructions.

What FreeDV Is

The FreeDV project describes itself as a suite of digital voice modes for HF radio. Its flagship mode is the Radio Autoencoder, or RADE, and it runs in a free GUI application for Windows, Linux and macOS that allows any SSB radio to be used for high-quality digital voice. The software is developed by an international team of radio amateurs, with David Rowe, VK5DGR, leading the signal processing side, and development is currently funded by a grant from Amateur Radio Digital Communications (ARDC), with the Software Freedom Conservancy as financial sponsor.

What makes FreeDV unusual is that everything is open. The project states that all of its software is open source, including the speech codec, which is released under licences that allow anyone to inspect, modify and experiment with it. That is a direct contrast with many other amateur digital voice systems, which rely on proprietary vocoders. For a hobby built on experimentation, that openness matters.

How It Works: Codec 2, Modems and RADE

Every digital voice system has three parts: a speech coder that turns your voice into a small number of bits, a modem that turns those bits into an audio signal your radio can transmit, and the matching receiving chain. FreeDV uses these building blocks:

• Codec 2. This is the open-source low-bit-rate speech codec developed for the project. It was designed so that voice can be carried in a very narrow bandwidth and without the patent restrictions of commercial vocoders.
• Legacy FreeDV modes. The original FreeDV 1600 mode pairs a frequency division multiplex modem with Codec 2. Its published specification describes 14 DQPSK carriers plus one DBPSK carrier, a total bandwidth of 1.125 kHz and a data rate of 1450 bits per second. According to that specification, path simulation and on-air testing showed that decoding is possible at a signal-to-noise ratio of 4 dB.
• RADE, the Radio Autoencoder. The newer flagship mode combines machine learning with classical digital signal processing. According to the project’s news, RADE can operate at signal-to-noise ratios as low as -2 dB, and controlled testing suggests it compares favourably with SSB on both high and low SNR channels. Its design also addresses spectrum use, speech quality and power efficiency.

The practical result is that RADE is aimed at working the same paths as SSB, but with speech that stays clear when ordinary SSB would be buried in noise. It is still a developing technology, with new versions in testing, so expect the details to evolve.

What You Need to Get on the Air

The getting started page lists the requirements. In short, you need an SSB receiver or transceiver and the FreeDV GUI software. The software is officially supported on Windows 10 and later, macOS Big Sur (11.0) and later on both ARM and Intel, and recent versions of Ubuntu and Fedora. The download page provides installers for each platform.

The radio connection is similar to other PC-based digital modes. If your radio has no built-in sound card, you can use its normal audio input and output, with the same cables and interfaces that work for other digital modes. FreeDV also needs a second sound interface for your microphone and speaker. The project suggests that a USB headset of the kind used by gamers is all you need. If you already run FT8 or similar modes with rig control, our guide to rig control with Hamlib covers the control side of the setup.

Easy Setup Step by Step

When you start FreeDV for the first time, an Easy Setup dialog appears. It divides configuration into three parts, and you can return to it later from the Tools menu.

1. Sound card configuration. Select the sound device associated with your radio, then choose the microphone and speaker devices you will use to hear decoded audio and to transmit your own voice. If you only want to receive, you can choose none for the transmit audio device.
2. CAT and PTT control. Easy Setup supports three methods: no radio control, for example with a VOX interface; Hamlib CAT control; and serial port PTT control. Choose the one that matches your station and fill in the radio type and port or the serial port and RTS or DTR pin.
3. Reporting. Reporting is not required, but it is recommended so that others can see who is receiving you. Both sides of a contact must have it enabled for contacts to be reported. You enter your call sign and your current grid square.
4. Set your levels. The Test button emits a constant carrier so that you can adjust your radio sound levels. A TX Level slider in the program lets you attenuate the transmit signal, and anything below 0 dB reduces it.

Take the audio levels seriously. As with other digital modes, overdriving the transmit audio creates distortion and splatter, which hurts your signal and your neighbours on the band. Start low and increase the level only until the radio’s ALC barely moves.

FreeDV Compared With Other Voice Modes

The table places FreeDV beside the voice modes it is most often compared with. The comparison is general and summarises the main differences.

If you want to compare these systems in more depth, our article on DMR, D-STAR and System Fusion explains the VHF and UHF digital voice systems, and the guide to digital modes beyond FT8 covers the text-based HF modes.

Where and When to Operate

Digital voice does not belong everywhere on the band. Band plans and national rules decide where voice and digital modes may be used, and these differ between regions. Before you transmit, check the band plan for your region and the rules that apply to your license. The IARU band plan guide explains how the regions differ.

There are several ways to find activity:

• FreeDV Reporter. The project’s reporting site shows who is currently using FreeDV and where, so you can see whether anyone is on the air before you call.
• Activity days. The project lists a worldwide FreeDV activity day on the third weekend of every month, running from Saturday to Monday around 0600 or 0700 UTC depending on US daylight saving time.
• Regular nets. The project’s home page lists regional nets, for example weekly sessions on 7 MHz in several countries and a UK session on 3.6 MHz. These schedules change, so check the page for current details.
• Chat channels. The project has a mailing list, a Discord server and an IRC channel for coordinating contacts and getting help.

If you have no HF radio, the project even links to a way of listening to FreeDV signals through a web receiver, which is a good way to hear what the mode sounds like.

Practical Tips for Good Results

• Keep your audio clean. Disable speech processing and compression on your radio for digital voice, because the digital waveform expects an undistorted audio signal.
• Do not overdrive the transmitter. The output of a digital voice signal should look clean on a monitor. If the ALC moves noticeably, reduce the drive.
• Close other audio programs. Other software that grabs your sound devices can break the audio setup, so close it before you start.
• Use a noise-free antenna and location. Digital voice does not cure local noise. If your receiver is swamped by interference, work on that first; our guide to finding and killing RF noise in a modern home explains how.
• Be patient with new versions. RADE is under active development, with test releases and new versions appearing regularly. Use the release recommended on the download page.

Why FreeDV Matters for the Hobby

The project’s own argument is that amateur radio is moving from analog to digital voice, as it moved from AM to SSB in the 1950s and 1960s, and that this transition should not leave the hobby dependent on one or two companies owning the patents. Whether or not you share that view, there is a practical benefit: because every part of the system is public, anyone can study how it works, improve it or build new equipment around it. The project also supports a hardware adaptor called ezDV for operators who prefer a self-contained device, and it offers a Flex waveform file for operators of certain commercial radios.

For an experimenter, this is exactly the kind of mode that makes amateur radio interesting. For an operator who just wants clearer contacts on a noisy band, it offers a new option without buying a new radio.

A Typical First Session

A sensible first session starts with listening, not transmitting. Install the program, run Easy Setup, and tune your radio to a frequency where FreeDV activity is expected, for example during a monthly activity day. Watch the program’s displays as signals arrive: you will see the received signal, the sync indicator and the decoded audio level. When a station decodes cleanly, you hear speech through the computer rather than through the radio’s speaker.

Once you have received a few stations, make a short test transmission on a quiet frequency, with the transmit level set conservatively, and ask a nearby station to report. Adjust the level based on what they tell you, then call CQ or answer a call. Keep your first contacts short, ask for a report on your audio, and note which settings worked. Because the mode is new to many operators, most are glad to help a newcomer sort out levels and software.

The Bottom Line

FreeDV is an open-source way to send digital voice over ordinary HF SSB equipment. Its flagship RADE mode uses machine learning to keep speech intelligible at low signal-to-noise ratios, and the older Codec 2 based modes remain available. You need an SSB transceiver, the free FreeDV program, a second sound interface and careful audio levels, and the Easy Setup dialog guides you through the rest. Enable reporting, check the band plan for your region, and join an activity day or net to find your first contact. As the mode keeps developing, follow the project’s pages for current versions and instructions.

Frequently Asked Questions

What is FreeDV?

FreeDV is a suite of open-source digital voice modes for HF radio. A free program on your computer encodes your voice, and an ordinary SSB radio transmits and receives the digital signal.

Do I need a special radio for FreeDV?

No. Any SSB receiver or transceiver can be used. You also need a computer running the FreeDV GUI and a second sound interface for your microphone and speaker, which can be a simple USB headset.

What is RADE?

RADE stands for Radio Autoencoder. It is the flagship FreeDV mode, combining machine learning with classical signal processing, and it is designed to work at low signal-to-noise ratios.

Which operating systems are supported?

The project officially supports Windows 10 and later, macOS Big Sur and later on ARM and Intel, and recent Ubuntu and Fedora releases. Other platforms may work but are not guaranteed.

Is FreeDV really free and open source?

Yes. The project says all of its software is open source, including the speech codec, under licences that include the GNU Lesser General Public License and a two-clause BSD license.

How do I find other FreeDV stations?

Use FreeDV Reporter to see who is active, join the monthly worldwide activity day, or try the regional nets listed on the project’s home page. Check the current schedules because they change.