Implement a simple MPEG-TS muxer using ffmpeg-lib

Question

I have an application that records raw audio data in LPCM stored in a buffer. I would like to encapsulate the data in a transport stream and send that transport stream through UDP to a stream segmenter (according to HTTP Live Streaming specifications) on another host.

FFmpeg provides a command-line utility to do so but with a file as input ffmpeg -re -i output.aac -acodec copy -f mpegts udp://127.0.0.1:5555 .

My first thought was to use FFmpeg API, especially the libavformat library. Does libavformat provide a muxer that I could use to encapsulate my audio in LPCM into a transport stream or do I have to implement it from scratch?

I have found this source code https://github.com/FFmpeg/FFmpeg/blob/master/libavformat/mpegts.c but I am not sure if it actually does what I'm looking for.

Thanks for your help,

Answer 1

So based on your comment about not needing it to necessarily be LPCM in the TS you will need to:

1. Decode your audio / read the frames
2. Encode it as a as something suitable for sending in a Transport Stream eg mp3 or AAC I believe this is the list of options: https://github.com/FFmpeg/FFmpeg/blob/master/libavformat/mpegts.h#L45-L64
3. Package it in a TS suitable for your network conditions eg packet sizing etc
4. Send it via UDP

There is a reasonable example of all this here: https://github.com/rvs/ffmpeg/blob/master/libavformat/output-example.c

As mentioned in the prior answer from szatmary you could also just pipe this to ffmpeg which may be simplest

Answer 2

You can use the ts muxer directly via libavformat. However you can also pipe the audio to ffmpeg using -i -