Files
rtphone/src/engine/media/MT_AudioStream.cpp
T
Dmytro Bogovych 03c58c4572 - G.722: decode as 16 kHz audio, keep the 8 kHz RTP clock separate
G722Codec::info() reported an 8 kHz sample rate for its 16 kHz output (640
bytes of PCM per 20 ms frame, read as 40 ms at 8 kHz). AudioReceiver therefore
fetched packets at half the rate they arrived, and the high-water trim
discarded about 44% of every G.722 stream: 109 of 250 packets per 5 s, with the
buffer at 360 ms instead of 100 ms.

The 8 kHz figure is right for one thing: RFC 3551 keeps G.722's RTP clock at
8000. Codec::Info already had mTimestampUnit for that. G.722 now reports
mSamplerate 16000 and mTimestampUnit 1/8000, and Codec::rtpClockRate() gives
the clock rate. The uses of samplerate() that mean RTP time now use it:
- the jitter statistics in AudioReceiver::add;
- the timestamp-gap-to-milliseconds conversion in AudioReceiver;
- AudioStream's transmit timestamp unit and increment (numerically unchanged
  for every codec, including G.722).
PCM uses stay on samplerate(), so AudioStream now also feeds the G.722
encoder 16 kHz audio instead of 8 kHz. The factory still advertises G722/8000
in SDP, as it must; only its backwards comment is fixed.

AudioReceiver::infoFor() also applied a "G729A silence" rule to payload type
9, which is G.722; G.729 is 18. It now applies to PT 18. Other fixed-frame
payloads are timed proportionally, as AudioReceiver::add already does, so a
10 ms G.722 packet (half a 160-byte frame) counts as 10 ms, not 20.

Measured in vq-core with 16 replayed calls (NISQA speech; G.722 from ffmpeg
and pjmedia, G.711 controls of the same speech; 3 runs each, all identical):
- G.722 packet drops: 44% -> 0; buffer 100 ms like G.711; no underruns;
- Sevana MOS of every G.722 call now equals its G.711 control, normal and loud
  (with the previous code three calls read 4.45 against 4.25);
- DeadAir-01 0.68-0.89 -> 0.01-0.27 (G.711: 0.00-0.14); SilentCall within 0.01
  of G.711 for the ffmpeg-encoded calls (was 0.15-0.17 higher);
- AmpClipping on loud calls 0.004/0.001, as G.711; the 0.02-0.03 seen after
  the codec swap alone came from the dropped packets;
- G.711 results, packet counts, loss, jitter and audio durations unchanged.
The transmit path (AudioStream) builds but was not exercised; vq-core only
receives.

Needs the vq_net change that computes jitter with rtpClockRate().

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-21 18:03:58 +02:00

495 lines
16 KiB
C++

/* Copyright(C) 2007-2017 VoIPobjects (voipobjects.com)
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "MT_AudioStream.h"
#include "MT_Dtmf.h"
#include "../helper/HL_StreamState.h"
#include "../helper/HL_Log.h"
#include "../audio/Audio_Resampler.h"
#include "../audio/Audio_Interface.h"
#include "jrtplib/src/rtpipv4address.h"
#include "jrtplib/src/rtpipv6address.h"
#include "jrtplib/src/rtppacket.h"
#include "jrtplib/src/rtptransmitter.h"
#include "jrtplib/src/rtpsessionparams.h"
#define LOG_SUBSYSTEM "media"
// #define DUMP_SENDING_AUDIO
using namespace MT;
AudioStream::AudioStream(const CodecList::Settings& settings)
: mPacketTime(0), mEncodedTime(0), mCodecSettings(settings), mRemoteTelephoneCodec(0), mRtpSession(),
mTransmittingPayloadType(-1), mRtpSender(mStat)
{
mOutputBuffer.setCapacity(16384);
mCapturedAudio.setCapacity(16384);
mCaptureResampler8.start(AUDIO_CHANNELS, AUDIO_SAMPLERATE, 8000);
mCaptureResampler16.start(AUDIO_CHANNELS, AUDIO_SAMPLERATE, 16000);
mCaptureResampler32.start(AUDIO_CHANNELS, AUDIO_SAMPLERATE, 32000);
mCaptureResampler48.start(AUDIO_CHANNELS, AUDIO_SAMPLERATE, 48000);
// Configure transmitter
jrtplib::RTPExternalTransmissionParams params(&mRtpSender, 0);
jrtplib::RTPSessionParams sessionParams;
sessionParams.SetAcceptOwnPackets(true);
sessionParams.SetMaximumPacketSize(MT_MAXRTPPACKET);
sessionParams.SetResolveLocalHostname(false);
sessionParams.SetUsePollThread(false);
sessionParams.SetOwnTimestampUnit(1 / 8000.0);
mRtpSession.Create(sessionParams, &params, jrtplib::RTPTransmitter::ExternalProto);
mRtpDtmfSession.Create(sessionParams, &params, jrtplib::RTPTransmitter::ExternalProto);
// Attach srtp session to sender
mRtpSender.setSrtpSession(&mSrtpSession);
// mRtpDump = new RtpDump("d:\\outgoing.rtp");
// mRtpSender.setDumpWriter(mRtpDump);
#if defined(DUMP_SENDING_AUDIO)
mSendingDump = std::make_shared<WavFileWriter>();
mSendingDump->open("sending_audio.wav", 8000, 1);
#endif
}
AudioStream::~AudioStream()
{
ICELogInfo(<< "Delete AudioStream instance");
if (mSendingDump)
{
mSendingDump->close();
mSendingDump.reset();
}
// Delete used rtp streams
for (AudioStreamMap::iterator streamIter = mStreamMap.begin(); streamIter != mStreamMap.end(); ++streamIter)
delete streamIter->second;
mStreamMap.clear();
if (mRtpDtmfSession.IsActive())
mRtpDtmfSession.Destroy();
if (mRtpSession.IsActive())
mRtpSession.Destroy();
#if defined(USE_RTPDUMP)
if (mRtpDump)
{
mRtpDump->flush();
delete mRtpDump;
}
#endif
mCaptureResampler8.stop();
mCaptureResampler16.stop();
mCaptureResampler32.stop();
mCaptureResampler48.stop();
ICELogInfo(<< "Encoded " << mEncodedTime << " milliseconds of audio");
if (mDumpStreams.mStreamForRecordingIncoming)
mDumpStreams.mStreamForRecordingIncoming->close();
if (mDumpStreams.mStreamForReadingOutgoing)
mDumpStreams.mStreamForReadingOutgoing->close();
if (mFinalStatistics)
*mFinalStatistics = mStat;
ICELogInfo(<< mStat.toString());
}
void AudioStream::setDestination(const RtpPair<InternetAddress>& dest)
{
Lock l(mMutex);
Stream::setDestination(dest);
mRtpSender.setDestination(dest);
}
void AudioStream::setTransmittingCodec(Codec::Factory& factory, int payloadType)
{
ICELogInfo(<< "Selected codec " << factory.name() << "/" << factory.samplerate() << " for transmitting");
Lock l(mMutex);
mTransmittingCodec = factory.create();
mTransmittingPayloadType = payloadType;
if (mRtpSession.IsActive())
mRtpSession.SetTimestampUnit(mTransmittingCodec->timestampUnit());
}
PCodec AudioStream::transmittingCodec()
{
Lock l(mMutex);
return mTransmittingCodec;
}
void AudioStream::addData(const void* buffer, int bytes)
{
assert(bytes == AUDIO_MIC_BUFFER_SIZE);
// Read predefined audio if configured
if (mDumpStreams.mStreamForReadingOutgoing)
{
if (mDumpStreams.mStreamForReadingOutgoing->isOpened())
mDumpStreams.mStreamForReadingOutgoing->read(const_cast<void*>(buffer), bytes);
}
// Read mirrored audio if needed
if (mMirror && mMirrorPrebuffered)
mMirrorBuffer.read(const_cast<void*>(buffer), bytes);
if (mMediaObserver)
mMediaObserver->onMedia(buffer, bytes, MT::Stream::MediaDirection::Outgoing, this, mMediaObserverTag);
Codec* codec = nullptr;
{
Lock l(mMutex);
codec = mTransmittingCodec.get();
if (nullptr == codec)
{
// ICELogDebug(<< "No transmitting codec selected.");
return;
}
}
// Resample
unsigned dstlen = unsigned(float(codec->samplerate() / float(AUDIO_SAMPLERATE)) * bytes);
Audio::Resampler* r = nullptr;
switch (codec->samplerate())
{
case 8000:
r = &mCaptureResampler8;
break;
case 16000:
r = &mCaptureResampler16;
break;
case 32000:
r = &mCaptureResampler32;
break;
case 48000:
r = &mCaptureResampler48;
break;
default:
assert(0);
}
size_t processedInput = 0;
dstlen = r->processBuffer(buffer, bytes, processedInput, mResampleBuffer, dstlen);
// ProcessedInput output value is ignored - because sample rate of input is always 8/16/32/48K - so all buffer is
// processed
// See if we need stereo <-> mono conversions
unsigned stereolen = 0;
if (codec->channels() != AUDIO_CHANNELS)
{
if (codec->channels() == 2)
stereolen = Audio::ChannelConverter::monoToStereo(mResampleBuffer, dstlen, mStereoBuffer, dstlen * 2);
else
dstlen = Audio::ChannelConverter::stereoToMono(mResampleBuffer, dstlen, mResampleBuffer, dstlen / 2);
}
// See if inband dtmf audio should be sent instead
ByteBuffer dtmf;
if (mDtmfContext.type() == DtmfContext::Dtmf_Inband &&
mDtmfContext.getInband(AUDIO_MIC_BUFFER_LENGTH, codec->samplerate(), dtmf))
mCapturedAudio.add(dtmf.data(), dtmf.size());
else
mCapturedAudio.add(stereolen ? mStereoBuffer : mResampleBuffer, stereolen ? stereolen : dstlen);
// See if it is time to send RFC2833 tone
ByteBuffer rfc2833, stopPacket;
if (mDtmfContext.type() == DtmfContext::Dtmf_Rfc2833 &&
mDtmfContext.getRfc2833(AUDIO_MIC_BUFFER_LENGTH, rfc2833, stopPacket))
{
if (rfc2833.size())
mRtpDtmfSession.SendPacket(rfc2833.data(), rfc2833.size(), mRemoteTelephoneCodec, true,
AUDIO_MIC_BUFFER_LENGTH * 8);
if (stopPacket.size())
{
for (int i = 0; i < 3; i++)
mRtpDtmfSession.SendPacket(stopPacket.data(), stopPacket.size(), mRemoteTelephoneCodec, true,
AUDIO_MIC_BUFFER_LENGTH * 8);
}
}
int processed = 0;
int encodedTime = 0;
int packetTime = mPacketTime ? mPacketTime : codec->frameTime();
// Make stereo version if required
for (int i = 0; i < mCapturedAudio.filled() / codec->pcmLength(); i++)
{
if (mSendingDump)
mSendingDump->write((const char*)mCapturedAudio.data() + codec->pcmLength() * i, codec->pcmLength());
auto r =
codec->encode({(const uint8_t*)mCapturedAudio.data() + codec->pcmLength() * i, (size_t)codec->pcmLength()},
{(uint8_t*)mFrameBuffer, MT_MAXAUDIOFRAME});
// Counter of processed input bytes of raw pcm data from microphone
processed += codec->pcmLength();
encodedTime += codec->frameTime();
mEncodedTime += codec->frameTime();
if (r.mEncoded)
{
mEncodedAudio.appendBuffer(mFrameBuffer, r.mEncoded);
if (packetTime <= encodedTime)
{
// Time to send packet
ICELogMedia(<< "Sending RTP packet pt = " << mTransmittingPayloadType
<< ", plength = " << (int)mEncodedAudio.size() << " to ");
mRtpSession.SendPacketEx(mEncodedAudio.data(), mEncodedAudio.size(), mTransmittingPayloadType, false,
packetTime * codec->rtpClockRate() / 1000, 0, nullptr, 0);
mEncodedAudio.clear();
encodedTime = 0;
}
}
}
if (processed > 0)
mCapturedAudio.erase(processed);
}
void AudioStream::copyDataTo(Audio::Mixer& mixer, int needed)
{
// mStreamMap is also mutated from the network thread (dataArrived)
Lock l(mMutex);
// Local audio mixer - used to send audio to media observer
Audio::Mixer localMixer;
Audio::DataWindow forObserver;
// Iterate
for (auto& streamIter : mStreamMap)
{
Audio::DataWindow w;
w.setCapacity(32768);
SingleAudioStream* sas = streamIter.second;
if (sas)
{
sas->copyPcmTo(w, needed);
// Provide mirroring if needed
if (mMirror)
{
mMirrorBuffer.add(w.data(), w.filled());
if (!mMirrorPrebuffered)
mMirrorPrebuffered = mMirrorBuffer.filled() >= MT_MIRROR_PREBUFFER;
}
if (!(state() & (int)StreamState::Receiving))
w.zero(needed);
// Check if we do not need input from this stream
if (w.filled())
{
if (mDumpStreams.mStreamForRecordingIncoming)
{
if (mDumpStreams.mStreamForRecordingIncoming->isOpened())
mDumpStreams.mStreamForRecordingIncoming->write(w.data(), w.filled());
}
mixer.addPcm(this, streamIter.first, w, AUDIO_SAMPLERATE, false);
if (mMediaObserver)
localMixer.addPcm(this, streamIter.first, w, AUDIO_SAMPLERATE, false);
}
}
}
if (mMediaObserver)
{
int mixedBytes = localMixer.mixAndGetPcm(forObserver);
if (mixedBytes > 0)
mMediaObserver->onMedia(forObserver.data(), mixedBytes, MT::Stream::MediaDirection::Incoming, this,
mMediaObserverTag);
}
}
void AudioStream::dataArrived(PDatagramSocket s, const void* buffer, int length, InternetAddress& source)
{
// Protects mStreamMap (also iterated by copyDataTo on the audio thread)
// and the receive/decrypt buffers.
Lock l(mMutex);
jrtplib::RTPIPv6Address addr6;
jrtplib::RTPIPv4Address addr4;
jrtplib::RTPExternalTransmissionInfo* info =
dynamic_cast<jrtplib::RTPExternalTransmissionInfo*>(mRtpSession.GetTransmissionInfo());
assert(info);
// Drop RTP packets if stream is not receiving now; let RTCP go
if (!(state() & (int)StreamState::Receiving) && RtpHelper::isRtp(buffer, length))
{
ICELogMedia(<< "Stream is not allowed to receive RTP stream. Ignore the RTP packet");
return;
}
// Copy incoming data to temp buffer to perform possible srtp unprotect
int receiveLength = length;
memcpy(mReceiveBuffer, buffer, length);
if (mSrtpSession.active())
{
bool srtpResult;
size_t srcLength = length;
size_t dstLength = sizeof mSrtpDecodeBuffer;
if (RtpHelper::isRtp(mReceiveBuffer, receiveLength))
srtpResult = mSrtpSession.unprotectRtp(mReceiveBuffer, srcLength, mSrtpDecodeBuffer, &dstLength);
else
srtpResult = mSrtpSession.unprotectRtcp(mReceiveBuffer, srcLength, mSrtpDecodeBuffer, &dstLength);
if (!srtpResult)
{
ICELogError(<< "Cannot decrypt SRTP packet.");
return;
}
memcpy(mReceiveBuffer, mSrtpDecodeBuffer, dstLength);
receiveLength = dstLength;
}
switch (source.family())
{
case AF_INET:
addr4.SetIP(source.sockaddr4()->sin_addr.s_addr);
addr4.SetPort(source.port());
ICELogMedia(<< "Injecting RTP/RTCP packet into jrtplib");
info->GetPacketInjector()->InjectRTPorRTCP(mReceiveBuffer, receiveLength, addr4);
break;
case AF_INET6:
addr6.SetIP(source.sockaddr6()->sin6_addr);
addr6.SetPort(source.port());
ICELogMedia(<< "Injecting RTP/RTCP packet into jrtplib");
info->GetPacketInjector()->InjectRTPorRTCP(mReceiveBuffer, receiveLength, addr6);
break;
default:
assert(0);
}
mStat.mReceived += length;
auto& perDst = mStat.mPerDestination[source];
perDst.mReceivedBytes += length;
if (RtpHelper::isRtp(mReceiveBuffer, receiveLength))
{
if (!mStat.mFirstRtpTime)
mStat.mFirstRtpTime = std::chrono::steady_clock::now();
mStat.mReceivedRtp++;
mStat.mReceivedRtpBytes += length;
perDst.mReceivedRtp++;
}
else
{
mStat.mReceivedRtcp++;
perDst.mReceivedRtcp++;
}
mRtpSession.Poll(); // maybe it is extra with external transmitter
bool hasData = mRtpSession.GotoFirstSourceWithData();
while (hasData)
{
std::shared_ptr<jrtplib::RTPPacket> packet(mRtpSession.GetNextPacket());
if (packet)
{
ICELogMedia(<< "jrtplib returned packet");
// Find right handler for rtp stream
SingleAudioStream* rtpStream = nullptr;
auto streamIter = mStreamMap.find(packet->GetSSRC());
if (streamIter == mStreamMap.end())
{
rtpStream = new SingleAudioStream(mCodecSettings, mStat);
mStreamMap.insert({packet->GetSSRC(), rtpStream});
}
else
rtpStream = streamIter->second;
// Process incoming data packet
rtpStream->process(packet);
// RTT sanity filter: jrtplib's INF_GetRoundtripTime() does the
// RFC 3550 §6.4.1 math but omits clock-skew / outlier guards;
// without these, a skewed or buggy peer can poison mRttDelay
// (and therefore the Id term in MOS).
double rtt = mRtpSession.GetCurrentSourceInfo()->INF_GetRoundtripTime().GetDouble();
if (rtt > 0 && rtt < 30.0) // reject "RTT not making any sense" (>30s)
{
// Once an average is established, cap a new sample at 3x mean
// so a single outlier can't skew the running RTT.
constexpr double kRttNormalizeFactor = 3.0;
const double meanRtt = mStat.mRttDelay.average();
if (mStat.mRttDelay.is_initialized() && meanRtt > 0.0 && rtt > meanRtt * kRttNormalizeFactor)
{
rtt = meanRtt * kRttNormalizeFactor;
}
mStat.mRttDelay.process(rtt);
}
}
hasData = mRtpSession.GotoNextSourceWithData();
}
}
void AudioStream::setState(unsigned state)
{
Stream::setState(state);
}
void AudioStream::setTelephoneCodec(int payloadType)
{
mRemoteTelephoneCodec = payloadType;
}
void AudioStream::setSocket(const RtpPair<PDatagramSocket>& socket)
{
Stream::setSocket(socket);
mRtpSender.setSocket(socket);
}
DtmfContext& AudioStream::queueOfDtmf()
{
return mDtmfContext;
}
void AudioStream::readFile(const Audio::PWavFileReader& stream, MediaDirection direction)
{
switch (direction)
{
case MediaDirection::Outgoing:
mDumpStreams.mStreamForReadingOutgoing = stream;
break;
case MediaDirection::Incoming:
mDumpStreams.mStreamForReadingIncoming = stream;
break;
}
}
void AudioStream::writeFile(const Audio::PWavFileWriter& writer, MediaDirection direction)
{
switch (direction)
{
case MediaDirection::Outgoing:
mDumpStreams.mStreamForRecordingOutgoing = writer;
break;
case MediaDirection::Incoming:
mDumpStreams.mStreamForRecordingIncoming = writer;
break;
}
}
void AudioStream::setupMirror(bool enable)
{
if (!mMirror && enable)
{
mMirrorBuffer.setCapacity(MT_MIRROR_CAPACITY);
mMirrorPrebuffered = false;
}
mMirror = enable;
}
void AudioStream::setFinalStatisticsOutput(Statistics* stats)
{
mFinalStatistics = stats;
}