| /* |
| * 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_state.h" |
| |
| #include <algorithm> |
| #include <memory> |
| #include <utility> |
| #include <vector> |
| |
| #include "api/sequence_checker.h" |
| #include "api/task_queue/task_queue_base.h" |
| #include "api/units/time_delta.h" |
| #include "audio/audio_receive_stream.h" |
| #include "audio/audio_send_stream.h" |
| #include "modules/audio_device/include/audio_device.h" |
| #include "rtc_base/checks.h" |
| #include "rtc_base/logging.h" |
| |
| namespace webrtc { |
| namespace internal { |
| |
| AudioState::AudioState(const AudioState::Config& config) |
| : config_(config), |
| audio_transport_(config_.audio_mixer.get(), |
| config_.audio_processing.get(), |
| config_.async_audio_processing_factory.get()) { |
| process_thread_checker_.Detach(); |
| RTC_DCHECK(config_.audio_mixer); |
| RTC_DCHECK(config_.audio_device_module); |
| } |
| |
| AudioState::~AudioState() { |
| RTC_DCHECK_RUN_ON(&thread_checker_); |
| RTC_DCHECK(receiving_streams_.empty()); |
| RTC_DCHECK(sending_streams_.empty()); |
| RTC_DCHECK(!null_audio_poller_.Running()); |
| } |
| |
| AudioProcessing* AudioState::audio_processing() { |
| return config_.audio_processing.get(); |
| } |
| |
| AudioTransport* AudioState::audio_transport() { |
| return &audio_transport_; |
| } |
| |
| void AudioState::AddReceivingStream( |
| webrtc::AudioReceiveStreamInterface* stream) { |
| RTC_DCHECK_RUN_ON(&thread_checker_); |
| RTC_DCHECK_EQ(0, receiving_streams_.count(stream)); |
| receiving_streams_.insert(stream); |
| if (!config_.audio_mixer->AddSource( |
| static_cast<AudioReceiveStreamImpl*>(stream))) { |
| RTC_DLOG(LS_ERROR) << "Failed to add source to mixer."; |
| } |
| |
| // Make sure playback is initialized; start playing if enabled. |
| UpdateNullAudioPollerState(); |
| auto* adm = config_.audio_device_module.get(); |
| if (!adm->Playing()) { |
| if (adm->InitPlayout() == 0) { |
| if (playout_enabled_) { |
| adm->StartPlayout(); |
| } |
| } else { |
| RTC_DLOG_F(LS_ERROR) << "Failed to initialize playout."; |
| } |
| } |
| } |
| |
| void AudioState::RemoveReceivingStream( |
| webrtc::AudioReceiveStreamInterface* stream) { |
| RTC_DCHECK_RUN_ON(&thread_checker_); |
| auto count = receiving_streams_.erase(stream); |
| RTC_DCHECK_EQ(1, count); |
| config_.audio_mixer->RemoveSource( |
| static_cast<AudioReceiveStreamImpl*>(stream)); |
| UpdateNullAudioPollerState(); |
| if (receiving_streams_.empty()) { |
| config_.audio_device_module->StopPlayout(); |
| } |
| } |
| |
| void AudioState::AddSendingStream(webrtc::AudioSendStream* stream, |
| int sample_rate_hz, |
| size_t num_channels) { |
| RTC_DCHECK_RUN_ON(&thread_checker_); |
| auto& properties = sending_streams_[stream]; |
| properties.sample_rate_hz = sample_rate_hz; |
| properties.num_channels = num_channels; |
| UpdateAudioTransportWithSendingStreams(); |
| |
| // Make sure recording is initialized; start recording if enabled. |
| auto* adm = config_.audio_device_module.get(); |
| if (!adm->Recording()) { |
| if (adm->InitRecording() == 0) { |
| if (recording_enabled_) { |
| adm->StartRecording(); |
| } |
| } else { |
| RTC_DLOG_F(LS_ERROR) << "Failed to initialize recording."; |
| } |
| } |
| } |
| |
| void AudioState::RemoveSendingStream(webrtc::AudioSendStream* stream) { |
| RTC_DCHECK_RUN_ON(&thread_checker_); |
| auto count = sending_streams_.erase(stream); |
| RTC_DCHECK_EQ(1, count); |
| UpdateAudioTransportWithSendingStreams(); |
| if (sending_streams_.empty()) { |
| config_.audio_device_module->StopRecording(); |
| } |
| } |
| |
| void AudioState::SetPlayout(bool enabled) { |
| RTC_LOG(LS_INFO) << "SetPlayout(" << enabled << ")"; |
| RTC_DCHECK_RUN_ON(&thread_checker_); |
| if (playout_enabled_ != enabled) { |
| playout_enabled_ = enabled; |
| if (enabled) { |
| UpdateNullAudioPollerState(); |
| if (!receiving_streams_.empty()) { |
| config_.audio_device_module->StartPlayout(); |
| } |
| } else { |
| config_.audio_device_module->StopPlayout(); |
| UpdateNullAudioPollerState(); |
| } |
| } |
| } |
| |
| void AudioState::SetRecording(bool enabled) { |
| RTC_LOG(LS_INFO) << "SetRecording(" << enabled << ")"; |
| RTC_DCHECK_RUN_ON(&thread_checker_); |
| if (recording_enabled_ != enabled) { |
| recording_enabled_ = enabled; |
| if (enabled) { |
| if (!sending_streams_.empty()) { |
| config_.audio_device_module->StartRecording(); |
| } |
| } else { |
| config_.audio_device_module->StopRecording(); |
| } |
| } |
| } |
| |
| void AudioState::SetStereoChannelSwapping(bool enable) { |
| RTC_DCHECK(thread_checker_.IsCurrent()); |
| audio_transport_.SetStereoChannelSwapping(enable); |
| } |
| |
| void AudioState::UpdateAudioTransportWithSendingStreams() { |
| RTC_DCHECK(thread_checker_.IsCurrent()); |
| std::vector<AudioSender*> audio_senders; |
| int max_sample_rate_hz = 8000; |
| size_t max_num_channels = 1; |
| for (const auto& kv : sending_streams_) { |
| audio_senders.push_back(kv.first); |
| max_sample_rate_hz = std::max(max_sample_rate_hz, kv.second.sample_rate_hz); |
| max_num_channels = std::max(max_num_channels, kv.second.num_channels); |
| } |
| audio_transport_.UpdateAudioSenders(std::move(audio_senders), |
| max_sample_rate_hz, max_num_channels); |
| } |
| |
| void AudioState::UpdateNullAudioPollerState() { |
| // Run NullAudioPoller when there are receiving streams and playout is |
| // disabled. |
| if (!receiving_streams_.empty() && !playout_enabled_) { |
| if (!null_audio_poller_.Running()) { |
| AudioTransport* audio_transport = &audio_transport_; |
| null_audio_poller_ = RepeatingTaskHandle::Start( |
| TaskQueueBase::Current(), [audio_transport] { |
| static constexpr size_t kNumChannels = 1; |
| static constexpr uint32_t kSamplesPerSecond = 48'000; |
| // 10ms of samples |
| static constexpr size_t kNumSamples = kSamplesPerSecond / 100; |
| |
| // Buffer to hold the audio samples. |
| int16_t buffer[kNumSamples * kNumChannels]; |
| |
| // Output variables from `NeedMorePlayData`. |
| size_t n_samples; |
| int64_t elapsed_time_ms; |
| int64_t ntp_time_ms; |
| audio_transport->NeedMorePlayData( |
| kNumSamples, sizeof(int16_t), kNumChannels, kSamplesPerSecond, |
| buffer, n_samples, &elapsed_time_ms, &ntp_time_ms); |
| |
| // Reschedule the next poll iteration. |
| return TimeDelta::Millis(10); |
| }); |
| } |
| } else { |
| null_audio_poller_.Stop(); |
| } |
| } |
| } // namespace internal |
| |
| rtc::scoped_refptr<AudioState> AudioState::Create( |
| const AudioState::Config& config) { |
| return rtc::make_ref_counted<internal::AudioState>(config); |
| } |
| } // namespace webrtc |