Files
platform-external-webrtc/audio/audio_receive_stream.cc
Tommi 90738ddb4e Split VideoReceiveStream2 init into worker / network steps.
This is in preparation for actually doing this initialization
differently in the Call class. This CL takes the registration
steps that are inherently network thread associated and makes
them separate from the ctor/dtor.

Inject Call* instead of worker_thread(), which will simplify upcoming
work that needs to access the network_thread() as well.

This is related to:
https://webrtc-review.googlesource.com/c/src/+/220608
https://webrtc-review.googlesource.com/c/src/+/220609

Bug: webrtc:11993
Change-Id: I72769fd61de84967d9a645750c40d01660a2716b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/220764
Reviewed-by: Markus Handell <handellm@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34172}
2021-05-31 17:10:23 +00:00

430 lines
16 KiB
C++

/*
* Copyright (c) 2015 The WebRTC project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include "audio/audio_receive_stream.h"
#include <string>
#include <utility>
#include "absl/memory/memory.h"
#include "api/array_view.h"
#include "api/audio_codecs/audio_format.h"
#include "api/call/audio_sink.h"
#include "api/rtp_parameters.h"
#include "api/sequence_checker.h"
#include "audio/audio_send_stream.h"
#include "audio/audio_state.h"
#include "audio/channel_receive.h"
#include "audio/conversion.h"
#include "call/rtp_config.h"
#include "call/rtp_stream_receiver_controller_interface.h"
#include "modules/rtp_rtcp/source/rtp_packet_received.h"
#include "rtc_base/checks.h"
#include "rtc_base/logging.h"
#include "rtc_base/strings/string_builder.h"
#include "rtc_base/time_utils.h"
namespace webrtc {
std::string AudioReceiveStream::Config::Rtp::ToString() const {
char ss_buf[1024];
rtc::SimpleStringBuilder ss(ss_buf);
ss << "{remote_ssrc: " << remote_ssrc;
ss << ", local_ssrc: " << local_ssrc;
ss << ", transport_cc: " << (transport_cc ? "on" : "off");
ss << ", nack: " << nack.ToString();
ss << ", extensions: [";
for (size_t i = 0; i < extensions.size(); ++i) {
ss << extensions[i].ToString();
if (i != extensions.size() - 1) {
ss << ", ";
}
}
ss << ']';
ss << '}';
return ss.str();
}
std::string AudioReceiveStream::Config::ToString() const {
char ss_buf[1024];
rtc::SimpleStringBuilder ss(ss_buf);
ss << "{rtp: " << rtp.ToString();
ss << ", rtcp_send_transport: "
<< (rtcp_send_transport ? "(Transport)" : "null");
if (!sync_group.empty()) {
ss << ", sync_group: " << sync_group;
}
ss << '}';
return ss.str();
}
namespace internal {
namespace {
std::unique_ptr<voe::ChannelReceiveInterface> CreateChannelReceive(
Clock* clock,
webrtc::AudioState* audio_state,
ProcessThread* module_process_thread,
NetEqFactory* neteq_factory,
const webrtc::AudioReceiveStream::Config& config,
RtcEventLog* event_log) {
RTC_DCHECK(audio_state);
internal::AudioState* internal_audio_state =
static_cast<internal::AudioState*>(audio_state);
return voe::CreateChannelReceive(
clock, module_process_thread, neteq_factory,
internal_audio_state->audio_device_module(), config.rtcp_send_transport,
event_log, config.rtp.local_ssrc, config.rtp.remote_ssrc,
config.jitter_buffer_max_packets, config.jitter_buffer_fast_accelerate,
config.jitter_buffer_min_delay_ms,
config.jitter_buffer_enable_rtx_handling, config.decoder_factory,
config.codec_pair_id, config.frame_decryptor, config.crypto_options,
std::move(config.frame_transformer));
}
} // namespace
AudioReceiveStream::AudioReceiveStream(
Clock* clock,
PacketRouter* packet_router,
ProcessThread* module_process_thread,
NetEqFactory* neteq_factory,
const webrtc::AudioReceiveStream::Config& config,
const rtc::scoped_refptr<webrtc::AudioState>& audio_state,
webrtc::RtcEventLog* event_log)
: AudioReceiveStream(clock,
packet_router,
config,
audio_state,
event_log,
CreateChannelReceive(clock,
audio_state.get(),
module_process_thread,
neteq_factory,
config,
event_log)) {}
AudioReceiveStream::AudioReceiveStream(
Clock* clock,
PacketRouter* packet_router,
const webrtc::AudioReceiveStream::Config& config,
const rtc::scoped_refptr<webrtc::AudioState>& audio_state,
webrtc::RtcEventLog* event_log,
std::unique_ptr<voe::ChannelReceiveInterface> channel_receive)
: config_(config),
audio_state_(audio_state),
source_tracker_(clock),
channel_receive_(std::move(channel_receive)) {
RTC_LOG(LS_INFO) << "AudioReceiveStream: " << config.rtp.remote_ssrc;
RTC_DCHECK(config.decoder_factory);
RTC_DCHECK(config.rtcp_send_transport);
RTC_DCHECK(audio_state_);
RTC_DCHECK(channel_receive_);
packet_sequence_checker_.Detach();
RTC_DCHECK(packet_router);
// Configure bandwidth estimation.
channel_receive_->RegisterReceiverCongestionControlObjects(packet_router);
// When output is muted, ChannelReceive will directly notify the source
// tracker of "delivered" frames, so RtpReceiver information will continue to
// be updated.
channel_receive_->SetSourceTracker(&source_tracker_);
// Complete configuration.
// TODO(solenberg): Config NACK history window (which is a packet count),
// using the actual packet size for the configured codec.
channel_receive_->SetNACKStatus(config.rtp.nack.rtp_history_ms != 0,
config.rtp.nack.rtp_history_ms / 20);
channel_receive_->SetReceiveCodecs(config.decoder_map);
channel_receive_->SetDepacketizerToDecoderFrameTransformer(
config.frame_transformer);
}
AudioReceiveStream::~AudioReceiveStream() {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
RTC_LOG(LS_INFO) << "~AudioReceiveStream: " << config_.rtp.remote_ssrc;
Stop();
channel_receive_->SetAssociatedSendChannel(nullptr);
channel_receive_->ResetReceiverCongestionControlObjects();
}
void AudioReceiveStream::RegisterWithTransport(
RtpStreamReceiverControllerInterface* receiver_controller) {
RTC_DCHECK_RUN_ON(&packet_sequence_checker_);
RTC_DCHECK(!rtp_stream_receiver_);
rtp_stream_receiver_ = receiver_controller->CreateReceiver(
config_.rtp.remote_ssrc, channel_receive_.get());
}
void AudioReceiveStream::UnregisterFromTransport() {
RTC_DCHECK_RUN_ON(&packet_sequence_checker_);
rtp_stream_receiver_.reset();
}
void AudioReceiveStream::Reconfigure(
const webrtc::AudioReceiveStream::Config& config) {
RTC_DCHECK(worker_thread_checker_.IsCurrent());
// Configuration parameters which cannot be changed.
RTC_DCHECK(config_.rtp.remote_ssrc == config.rtp.remote_ssrc);
RTC_DCHECK(config_.rtcp_send_transport == config.rtcp_send_transport);
// Decoder factory cannot be changed because it is configured at
// voe::Channel construction time.
RTC_DCHECK(config_.decoder_factory == config.decoder_factory);
// SSRC can't be changed mid-stream.
RTC_DCHECK_EQ(config_.rtp.local_ssrc, config.rtp.local_ssrc);
RTC_DCHECK_EQ(config_.rtp.remote_ssrc, config.rtp.remote_ssrc);
// TODO(solenberg): Config NACK history window (which is a packet count),
// using the actual packet size for the configured codec.
if (config_.rtp.nack.rtp_history_ms != config.rtp.nack.rtp_history_ms) {
channel_receive_->SetNACKStatus(config.rtp.nack.rtp_history_ms != 0,
config.rtp.nack.rtp_history_ms / 20);
}
if (config_.decoder_map != config.decoder_map) {
channel_receive_->SetReceiveCodecs(config.decoder_map);
}
if (config_.frame_transformer != config.frame_transformer) {
channel_receive_->SetDepacketizerToDecoderFrameTransformer(
config.frame_transformer);
}
config_ = config;
}
void AudioReceiveStream::Start() {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
if (playing_) {
return;
}
channel_receive_->StartPlayout();
playing_ = true;
audio_state()->AddReceivingStream(this);
}
void AudioReceiveStream::Stop() {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
if (!playing_) {
return;
}
channel_receive_->StopPlayout();
playing_ = false;
audio_state()->RemoveReceivingStream(this);
}
bool AudioReceiveStream::IsRunning() const {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return playing_;
}
webrtc::AudioReceiveStream::Stats AudioReceiveStream::GetStats(
bool get_and_clear_legacy_stats) const {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
webrtc::AudioReceiveStream::Stats stats;
stats.remote_ssrc = config_.rtp.remote_ssrc;
webrtc::CallReceiveStatistics call_stats =
channel_receive_->GetRTCPStatistics();
// TODO(solenberg): Don't return here if we can't get the codec - return the
// stats we *can* get.
auto receive_codec = channel_receive_->GetReceiveCodec();
if (!receive_codec) {
return stats;
}
stats.payload_bytes_rcvd = call_stats.payload_bytes_rcvd;
stats.header_and_padding_bytes_rcvd =
call_stats.header_and_padding_bytes_rcvd;
stats.packets_rcvd = call_stats.packetsReceived;
stats.packets_lost = call_stats.cumulativeLost;
stats.capture_start_ntp_time_ms = call_stats.capture_start_ntp_time_ms_;
stats.last_packet_received_timestamp_ms =
call_stats.last_packet_received_timestamp_ms;
stats.codec_name = receive_codec->second.name;
stats.codec_payload_type = receive_codec->first;
int clockrate_khz = receive_codec->second.clockrate_hz / 1000;
if (clockrate_khz > 0) {
stats.jitter_ms = call_stats.jitterSamples / clockrate_khz;
}
stats.delay_estimate_ms = channel_receive_->GetDelayEstimate();
stats.audio_level = channel_receive_->GetSpeechOutputLevelFullRange();
stats.total_output_energy = channel_receive_->GetTotalOutputEnergy();
stats.total_output_duration = channel_receive_->GetTotalOutputDuration();
stats.estimated_playout_ntp_timestamp_ms =
channel_receive_->GetCurrentEstimatedPlayoutNtpTimestampMs(
rtc::TimeMillis());
// Get jitter buffer and total delay (alg + jitter + playout) stats.
auto ns = channel_receive_->GetNetworkStatistics(get_and_clear_legacy_stats);
stats.fec_packets_received = ns.fecPacketsReceived;
stats.fec_packets_discarded = ns.fecPacketsDiscarded;
stats.jitter_buffer_ms = ns.currentBufferSize;
stats.jitter_buffer_preferred_ms = ns.preferredBufferSize;
stats.total_samples_received = ns.totalSamplesReceived;
stats.concealed_samples = ns.concealedSamples;
stats.silent_concealed_samples = ns.silentConcealedSamples;
stats.concealment_events = ns.concealmentEvents;
stats.jitter_buffer_delay_seconds =
static_cast<double>(ns.jitterBufferDelayMs) /
static_cast<double>(rtc::kNumMillisecsPerSec);
stats.jitter_buffer_emitted_count = ns.jitterBufferEmittedCount;
stats.jitter_buffer_target_delay_seconds =
static_cast<double>(ns.jitterBufferTargetDelayMs) /
static_cast<double>(rtc::kNumMillisecsPerSec);
stats.inserted_samples_for_deceleration = ns.insertedSamplesForDeceleration;
stats.removed_samples_for_acceleration = ns.removedSamplesForAcceleration;
stats.expand_rate = Q14ToFloat(ns.currentExpandRate);
stats.speech_expand_rate = Q14ToFloat(ns.currentSpeechExpandRate);
stats.secondary_decoded_rate = Q14ToFloat(ns.currentSecondaryDecodedRate);
stats.secondary_discarded_rate = Q14ToFloat(ns.currentSecondaryDiscardedRate);
stats.accelerate_rate = Q14ToFloat(ns.currentAccelerateRate);
stats.preemptive_expand_rate = Q14ToFloat(ns.currentPreemptiveRate);
stats.jitter_buffer_flushes = ns.packetBufferFlushes;
stats.delayed_packet_outage_samples = ns.delayedPacketOutageSamples;
stats.relative_packet_arrival_delay_seconds =
static_cast<double>(ns.relativePacketArrivalDelayMs) /
static_cast<double>(rtc::kNumMillisecsPerSec);
stats.interruption_count = ns.interruptionCount;
stats.total_interruption_duration_ms = ns.totalInterruptionDurationMs;
auto ds = channel_receive_->GetDecodingCallStatistics();
stats.decoding_calls_to_silence_generator = ds.calls_to_silence_generator;
stats.decoding_calls_to_neteq = ds.calls_to_neteq;
stats.decoding_normal = ds.decoded_normal;
stats.decoding_plc = ds.decoded_neteq_plc;
stats.decoding_codec_plc = ds.decoded_codec_plc;
stats.decoding_cng = ds.decoded_cng;
stats.decoding_plc_cng = ds.decoded_plc_cng;
stats.decoding_muted_output = ds.decoded_muted_output;
stats.last_sender_report_timestamp_ms =
call_stats.last_sender_report_timestamp_ms;
stats.last_sender_report_remote_timestamp_ms =
call_stats.last_sender_report_remote_timestamp_ms;
stats.sender_reports_packets_sent = call_stats.sender_reports_packets_sent;
stats.sender_reports_bytes_sent = call_stats.sender_reports_bytes_sent;
stats.sender_reports_reports_count = call_stats.sender_reports_reports_count;
return stats;
}
void AudioReceiveStream::SetSink(AudioSinkInterface* sink) {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
channel_receive_->SetSink(sink);
}
void AudioReceiveStream::SetGain(float gain) {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
channel_receive_->SetChannelOutputVolumeScaling(gain);
}
bool AudioReceiveStream::SetBaseMinimumPlayoutDelayMs(int delay_ms) {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return channel_receive_->SetBaseMinimumPlayoutDelayMs(delay_ms);
}
int AudioReceiveStream::GetBaseMinimumPlayoutDelayMs() const {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return channel_receive_->GetBaseMinimumPlayoutDelayMs();
}
std::vector<RtpSource> AudioReceiveStream::GetSources() const {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return source_tracker_.GetSources();
}
AudioMixer::Source::AudioFrameInfo AudioReceiveStream::GetAudioFrameWithInfo(
int sample_rate_hz,
AudioFrame* audio_frame) {
AudioMixer::Source::AudioFrameInfo audio_frame_info =
channel_receive_->GetAudioFrameWithInfo(sample_rate_hz, audio_frame);
if (audio_frame_info != AudioMixer::Source::AudioFrameInfo::kError) {
source_tracker_.OnFrameDelivered(audio_frame->packet_infos_);
}
return audio_frame_info;
}
int AudioReceiveStream::Ssrc() const {
return config_.rtp.remote_ssrc;
}
int AudioReceiveStream::PreferredSampleRate() const {
return channel_receive_->PreferredSampleRate();
}
uint32_t AudioReceiveStream::id() const {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return config_.rtp.remote_ssrc;
}
absl::optional<Syncable::Info> AudioReceiveStream::GetInfo() const {
// TODO(bugs.webrtc.org/11993): This is called via RtpStreamsSynchronizer,
// expect to be called on the network thread.
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return channel_receive_->GetSyncInfo();
}
bool AudioReceiveStream::GetPlayoutRtpTimestamp(uint32_t* rtp_timestamp,
int64_t* time_ms) const {
// Called on video capture thread.
return channel_receive_->GetPlayoutRtpTimestamp(rtp_timestamp, time_ms);
}
void AudioReceiveStream::SetEstimatedPlayoutNtpTimestampMs(
int64_t ntp_timestamp_ms,
int64_t time_ms) {
// Called on video capture thread.
channel_receive_->SetEstimatedPlayoutNtpTimestampMs(ntp_timestamp_ms,
time_ms);
}
bool AudioReceiveStream::SetMinimumPlayoutDelay(int delay_ms) {
// TODO(bugs.webrtc.org/11993): This is called via RtpStreamsSynchronizer,
// expect to be called on the network thread.
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return channel_receive_->SetMinimumPlayoutDelay(delay_ms);
}
void AudioReceiveStream::AssociateSendStream(AudioSendStream* send_stream) {
RTC_DCHECK_RUN_ON(&packet_sequence_checker_);
channel_receive_->SetAssociatedSendChannel(
send_stream ? send_stream->GetChannel() : nullptr);
associated_send_stream_ = send_stream;
}
void AudioReceiveStream::DeliverRtcp(const uint8_t* packet, size_t length) {
// TODO(solenberg): Tests call this function on a network thread, libjingle
// calls on the worker thread. We should move towards always using a network
// thread. Then this check can be enabled.
// RTC_DCHECK(!thread_checker_.IsCurrent());
channel_receive_->ReceivedRTCPPacket(packet, length);
}
const webrtc::AudioReceiveStream::Config& AudioReceiveStream::config() const {
RTC_DCHECK_RUN_ON(&worker_thread_checker_);
return config_;
}
const AudioSendStream* AudioReceiveStream::GetAssociatedSendStreamForTesting()
const {
RTC_DCHECK_RUN_ON(&packet_sequence_checker_);
return associated_send_stream_;
}
internal::AudioState* AudioReceiveStream::audio_state() const {
auto* audio_state = static_cast<internal::AudioState*>(audio_state_.get());
RTC_DCHECK(audio_state);
return audio_state;
}
} // namespace internal
} // namespace webrtc