FreeDV RADE: An Open-Source Digital Voice Mode for HF that Beats SSB at Low SNR

Thank you to Rebecca Key (KO4KVG) for writing in and sharing with us an article she's written over on the ARDC website about a relatively new amateur open-source radio digital voice mode called FreeDV RADE.

ARDC is a private foundation that issues grants that support amateur radio and digital communications technologies. FreeDV is an ARDC grantee, and it has been using the funds to develop "a digital voice mode that could compete with SSB across varying signal-to-noise conditions on HF while providing more natural-sounding speech".

Last year, they released the FreeDV Radio Autoencoder (FreeDV RADE), which is their flagship digital voice mode. FreeDV RADE incorporates both machine learning and classical DSP, resulting in working audio at SNRs as low as -2 dB, with only half the bandwidth of a typical 3 kHz SSB voice signal. The article goes on to describe how the audio quality, power, and bandwidth efficiency are superior to other digital voice modes. 

The article also mentions how FreeDV RADE is seeing increased adoption by hardware manufacturers, with FlexRadio now supporting the mode in several radios. It also mentions upcoming improvements, including a goal of reducing bandwidth use to under 1 kHz.

FreeDV Software
FreeDV Software
Tweet
Share
Reddit
Vote
Email

Related posts:

  1. Using a PlutoSDR to compare FreeDV Digital Voice with other Modes
  2. Listening to FreeDV Digital Speech with an RTL-SDR
  3. Listening to D-STAR Digital Voice with DSD 1.7
  4. Creating an Open Source DMR Transceiver with a LimeSDR Mini
  5. Decoding Amateur Radio Digital Voice with an RTL-SDR and the QRadioLink Android App
Subscribe
Notify of
guest

3 Comments
snn47
#275103
  1. The 1.5 kHz used for comparisson to a SSB-signal is only the bandwidth at 0dB, but does not identify how much spectrum is occupied by the RF-digital signal, that can interfere with other analogue (SSB) or digital signals. This information is required for frequency sharing between FreeDV digital signal and SSB-Signals. The wider a digital signal is and the higher the noisefloor is the more seperation between a digital and an SSB signal is required. In consequence the benfit of a “narrow” digital signal at 0 dB may offset the lower benefit compared to a SSB signal.
  2. Any digital signal is much wider than the specified audio cutoff frequency at 0 dB, and mostly the generated noisefloor is ommitted which is from those systems I know at best suppressed by ~ -60 dB.
  3. I am not aware that any SSB transmitter ever required more than ~2.5 kHz, since a DSB suppressed carriere is generated, and afilter used to suppress the 2nd undesired sideband. The same filter is used during reception, allowing close separation of SSB signal. I remember 2.4 kHz and 2.1 kHz XTAL filter if you desired to be narrower.
  4. There is also one method to reduce the required audio bandwith to 1.8 kHz which is descibed in the ARRL-HB.
  5. Even VHF-COM for ATC is specified as 2.5 kHz, even though in Germany some ground transmitter were operated for decades with an audio bandwidth of 2.1kHz.

FreeDV is quite an achievement, but the claimed benefits should be stated/compared to real SSB specifications and not fictional values.

0
Reply
KF9M
#274869

It looks and sounds very interesting, I do wish there would be more work toward an option of maximizing audio quality within a 2.5kHz bandwidth instead of always seeming to be chasing reducing bandwidth with a minimum acceptable audio. Basically a digital version of the ESSB trend.

1
Reply
KC1WZQ
#274773

Bit late but never late than never haha. Nice

0
Reply