How to setup an L16 PCM Track?

[George-Madeley]

I have taken the "streamer" example and removed support for video leaving only OPUS. My frontend can successfully receive the OPUS audio packets and play them without complaint. I'm now adding support for L16 PCM so the user can select what format they want using a CLI argument. I have refactored the addAudio() function from the example to be addPcm():

rtc_client_track_data rtc_sender::addPcm(const std::shared_ptr<rtc::PeerConnection> pc, const uint8_t payload_type,
    const uint32_t ssrc, const std::string& cname, const std::string& msid, const std::function<void(void)> on_open)
{
  const unsigned int clock_rate = 16000;
  auto audio = rtc::Description::Audio(cname);
  audio.addAudioCodec(payload_type, "L16/" + std::to_string(clock_rate) + "/1");
  audio.addSSRC(ssrc, cname, msid, cname);
  auto track = pc->addTrack(audio);
  // create RTP configuration
  auto rtp_config = std::make_shared<rtc::RtpPacketizationConfig>(ssrc, cname, payload_type, clock_rate);
  // create packetizer
  auto packetizer = std::make_shared<rtc::AudioRtpPacketizer<16000>>(rtp_config);
  // add RTCP SR handler
  auto sr_reporter = std::make_shared<rtc::RtcpSrReporter>(rtp_config);
  packetizer->addToChain(sr_reporter);
  // add RTCP NACK handler
  auto nack_responder = std::make_shared<rtc::RtcpNackResponder>();
  packetizer->addToChain(nack_responder);
  // set handler
  track->setMediaHandler(packetizer);
  track->onOpen(on_open);
  auto track_data = std::make_shared<ClientTrackData>(track, sr_reporter);
  return track_data;
}

I believe I have set this up to support streaming L16 PCM audio to my frontend however, I get the following error on my frontend TypeScript code when I set the local description:

Uncaught (in promise) OperationError: Failed to execute 'setLocalDescription' on 'RTCPeerConnection': Failed to set local answer sdp: Failed to apply the description for m= section with mid='0': Failed to setup RTCP mux.

I don't get this error when using OPUS and as I've only changed the addAudio() function to addPcm(), I believe it is to do with my newly created method. I know I can convert the PCM audio to OPUS then stream to the frontend but what the format the frontend receives must match that of the CLI argument (i.e., it needs to be PCM if the user set PCM). I am aware the library has support for PCMU and PCMA but even when using those packetizes, I get the same error.

If anyone has any direction or examples on how to stream L16 PCM audio using Web RTC that would much appreciated!

[paullouisageneau]

You can't receive uncompressed PCM because it is not supported for WebRTC in browsers. You should use PCMA or PCMU instead.