Simple audio HLS streaming server.

The goal of this program is to receive an audio stream with SRT and broadcast it with HLS. It can serve as a CDN origin server or as a set-and-forget streaming server.

srt2hls use Liquidsoap to receive a stream and encode it in HLS, then it use [nginx] (https://www.nginx.com/) to serve HLS content.

The Liquidsoap container, by default, run radio/live.liq script. It's a fully functionnal example that will :

1. Receive 2 SRT inputs on ports 10000 and 10001
2. Create a production stream with basic logic between the 2 inputs
3. Encode the stream in aac with 3 quality
4. Segment it in HLS format in /hls directory

The Nginx container come with a specific configuration to serve HLS content with proper Content-Type, CORS and Cache-Control headers. It need read only access to /hls directory to serve HLS segments and playlists.

mkdir hls
sudo chown -R 10000:10001 hls
sudo docker-compose up

• Liquidsoap 2.0.0+
• ffmpeg

By default the encoder send blank HLS segments, that mean if it's started, you can already listen the blank stream.

ffplay http://localhost:8080/live.m3u8
vlc http://localhost:8080/live.m3u8

The default script allow to switch the output between inputs.

# Check available input to switch to
curl http://localhost:8080/api/list?livesource
# Switch the livesource
curl http://localhost:8080/api/set?livesource=srt2
curl http://localhost:8080/api/set?livesource=srt1
# check the current live source
curl http://localhost:8080/api/get?livesource

Requirement : ffmpeg compiled with srt support (https://johnvansickle.com/ffmpeg/ for example)

# live stream or static file
ffmpeg -re -i $AUDIOFILE -vn -f wav -codec:a pcm_s16le srt://127.0.0.1:10000

http://localhost:3000/

http://localhost:9090/

sudo docker-compose build

Folks of Liquidsoap

Monitoring stack is mostly inspired by https://github.com/stefanprodan/dockprom