FFMpeg can be used to send an audio-only stream. An example of settings that work to stream a file to our systems:
ffmpeg -re -i test.flac -vn -c:a aac -b:a 96k -f flv rtmp://user-origin.secdn.net/user-origin/live/stream
Any source can be used, as long as you ensure the settings you are sending with are correct. Make sure the audio is AAC or MP3, and the container format is FLV.
If you are sending from a pre-recorded file, the -re flag makes it send
the stream in 'real time', rather than as fast as it can. Do not use it if your source is already
a stream.
ScaleEngine does live transcoding that will output various lower bitrates and frame sizes, as well as an audio only feed. We can literally remove video and output 32k audio, or simply strip out the video and pass through the original audio track for listeners. Generally speaking this is the best option if you are running a podcast and want a lower bandwidth option for your audience, who may be driving and not interested in the pictures.
It's important to remember that an audio option needs to be independent of the video player. Video players on any device do not understand "missing video" and will not "go back" to playing video if audio only is present and selected by the audience member on their device.
ScaleEngine specializes in transcoding "odd" audio codecs to the industry standard AAC, which is especially useful in the case of installed cameras, which use GPP and MP2 audio containers. These have to be transcoded to AAC to be usable on the modern device. There is also WebRTC, which uses OPUS - completely incompatible with an HLS feed (unless transcoded to AAC).