OpenSync FM brings precise GPS-assisted audio/MPX synchronization to the FM radio broadcasting community.
It is specifically designed for creating well-synchronised single-frequency-network (SFN) FM radio broadcasts, but can be used by anyone who wants to deliver well-synchronized audio or MPX signals to multiple locations.
Note: SFN's require precise audio synchronisation of broadcast signals to the sub-millisecond level. Typical computer architectures and audio player engines are not designed with precision and repeatability in mind.
OpenSync FM is a docker app having two modes of operation:
This is the main precision playback engine. It plays an OpenSync-style segmented audio stream - typically from a CDN. This is designed to run on Raspberry Pi Model 4B devices. Precision GPS synchronisation was designed for use with the "HiFiBerry DAC2 ADC Pro" sound card.
Converts an Icecast Stream to an OpenSyncFM-style segmented audio stream - streaming it to an S3-compatible CDN bucket. This should be compatible with a broad range of Linux systems (both cloud and physical hardware), but has primarily been developed with Ubuntu and Raspberry Pi OS targets in mind.
# Create user/group with matching UID/GID for the container
sudo groupadd -g 901 opensyncfm
sudo useradd -u 2001 -g opensyncfm opensyncfm
sudo usermod -a -G audio opensyncfm # ensure opensyncfm user is in audio group
sudo usermod -a -G opensyncfm $USER # ensure current user can access opensyncfm storage folder
# (restart your shell session before continuing)
# Create storage folders
sudo mkdir -p /opensyncfm/logs
sudo mkdir -p /opensyncfm/debug
sudo chown -R opensyncfm:opensyncfm /opensyncfm # ownership
sudo chmod -R g+w,g+s /opensyncfm # group write permissions, setgid ensures new files inherit group
# Configure firewall ports (IF UFW INSTALLED)
sudo ufw allow 4000
# Install docker
curl -fsSL https://get.docker.com -o get-docker.sh
sudo sh get-docker.sh
sudo usermod -aG docker $USER
# install 'osfm' helper script
sudo curl -fsSL https://raw.githubusercontent.com/faithfm/opensync-fm/master/osfm -o /usr/local/bin/osfm
sudo chmod 755 /usr/local/bin/osfm
# optionally configure a system tag to identify this system
echo '{ "system-tag": "my tag" }' > /opensyncfm/meta.json
# use the helper script to pull and start the OpenSync docker app
osfm pull
osfm restart
OpenSync FM should now be running and accessible on port 4000 - ie:
Our main Faith FM OpenSync stream can be used to test if the basics are working. Apply the following in Broadcast > Stream Source > URL template:
http://streams.faithfm.com.au/%t[1-3]/%t[1-5]/%t[1-7]/FaithFM.ogg-opus-%t.opus
Save configuration then toggle "Stream Source Active", then confirm that audio segments are being downloaded.
Adjust some precision player settings:
- Broadcast > Precision Player > Audio Device = "bcm2835 Headphones"
- Broadcast > Precision Player > Channel 1 (left) = "STREAM"
- Broadcast > Precision Player > Channel 2 (right) = "PPS"
Save configuration then toggle "Precision Player Active". If everything is working you should hear the Faith FM audio stream in the left channel, and a 1 PPS click in the right channel.
Note: the "HiFiBerry DAC+ADC Pro HiFi multicodec-0" device is required for precision GPS synchronised MPX signals. Instructions not yet documented.
OpenSync Studio will not normally be used on the same hardware as your broadcast device(s) - ie: a cloud server or hardware in a central location with a good internet connection.
Note: You'll need write access to an S3-compatible block storage bucket as an output for OpenSync Studio.
Our main Faith FM icecast stream can be used to test if the basics are working:
- Studio > Stream Recorder > Stream Sources > + Add Stream
- Studio > Stream Recorder > Stream Sources > Static Streams > Stream URL = "https://stream2.faithfm.com.au/FaithFM.mp3"
- Studio > Stream Recorder > Stream Sources > Static Streams > Name = "FaithFM"
- Studio > Stream Recorder > Stream Sources > Dynamic Streams > Icecast Server =
Save configuration then toggle "Sream Recorder Active". A log of encoding attempts should start, although you'll get a output + upload errors because no output bucket has been configured.
Now stop the "Stream Recorder Active" and configure output settings using your S3-compatible bucket. I've included a screenshot of a configuration we tested using a Backblaze B2 bucket, but you'll need to use your own details here.
Note: We have found that 64kbps Opus streams provide sufficient quality for our purposes at Faith FM.
Note that the Studio > Stream Recorder > Output Settings > S3 Path Template that was used in this sample:
s3://streams.faithfm.com.au/deleteme/%t[1-3]/%t[1-5]/%t[1-7]/%m-{format}-%t.{format}
Coresponds to the following Broadcast > Stream Source > URL template that would be used in your OpenSync Broadcast units:
https://s3.us-west-001.backblazeb2.com/streams-faithfm-com-au/deleteme/%t[1-3]/%t[1-5]/%t[1-7]/FaithFM-opus-%t.opus
Note: When testing this on a Ubuntu 24.04 Virtual Machine running on UTM, docker was not passing a valid nameserver to containers, and all DNS lookups were failing. Applying this fix to the VM worked:
sudo mkdir -p /etc/docker
cat << 'EOF' | sudo tee /etc/docker/daemon.json
{
"dns": ["8.8.8.8", "1.1.1.1"]
}
EOF
sudo systemctl restart docker