2011-07-07 08:21:25 +00:00
|
|
|
/*
|
2012-01-30 20:51:15 +00:00
|
|
|
* Copyright (c) 2012 The WebRTC project authors. All Rights Reserved.
|
2011-07-07 08:21:25 +00:00
|
|
|
*
|
|
|
|
|
* 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.
|
|
|
|
|
*/
|
|
|
|
|
|
2013-03-01 16:36:19 +00:00
|
|
|
#include "webrtc/modules/audio_processing/audio_processing_impl.h"
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2011-08-03 21:08:51 +00:00
|
|
|
#include <assert.h>
|
2015-07-23 11:41:39 -07:00
|
|
|
#include <algorithm>
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-06-29 14:57:29 +02:00
|
|
|
#include "webrtc/base/checks.h"
|
2014-10-10 08:36:56 +00:00
|
|
|
#include "webrtc/base/platform_file.h"
|
2015-08-14 10:35:55 -07:00
|
|
|
#include "webrtc/common_audio/audio_converter.h"
|
2015-03-25 16:37:27 -07:00
|
|
|
#include "webrtc/common_audio/channel_buffer.h"
|
2015-08-14 10:35:55 -07:00
|
|
|
#include "webrtc/common_audio/include/audio_util.h"
|
2014-01-07 17:45:09 +00:00
|
|
|
#include "webrtc/common_audio/signal_processing/include/signal_processing_library.h"
|
2015-06-29 14:57:29 +02:00
|
|
|
extern "C" {
|
|
|
|
|
#include "webrtc/modules/audio_processing/aec/aec_core.h"
|
|
|
|
|
}
|
2014-12-15 09:41:24 +00:00
|
|
|
#include "webrtc/modules/audio_processing/agc/agc_manager_direct.h"
|
2013-03-01 16:36:19 +00:00
|
|
|
#include "webrtc/modules/audio_processing/audio_buffer.h"
|
2015-03-13 00:13:32 +00:00
|
|
|
#include "webrtc/modules/audio_processing/beamformer/nonlinear_beamformer.h"
|
2014-04-22 21:00:04 +00:00
|
|
|
#include "webrtc/modules/audio_processing/common.h"
|
2014-02-27 22:23:17 +00:00
|
|
|
#include "webrtc/modules/audio_processing/echo_cancellation_impl.h"
|
2013-03-01 16:36:19 +00:00
|
|
|
#include "webrtc/modules/audio_processing/echo_control_mobile_impl.h"
|
|
|
|
|
#include "webrtc/modules/audio_processing/gain_control_impl.h"
|
|
|
|
|
#include "webrtc/modules/audio_processing/high_pass_filter_impl.h"
|
2015-08-14 10:35:55 -07:00
|
|
|
#include "webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.h"
|
2013-03-01 16:36:19 +00:00
|
|
|
#include "webrtc/modules/audio_processing/level_estimator_impl.h"
|
|
|
|
|
#include "webrtc/modules/audio_processing/noise_suppression_impl.h"
|
|
|
|
|
#include "webrtc/modules/audio_processing/processing_component.h"
|
2014-12-19 19:57:34 +00:00
|
|
|
#include "webrtc/modules/audio_processing/transient/transient_suppressor.h"
|
2013-03-01 16:36:19 +00:00
|
|
|
#include "webrtc/modules/audio_processing/voice_detection_impl.h"
|
2015-11-04 08:31:52 +01:00
|
|
|
#include "webrtc/modules/include/module_common_types.h"
|
2015-10-28 18:17:40 +01:00
|
|
|
#include "webrtc/system_wrappers/include/file_wrapper.h"
|
|
|
|
|
#include "webrtc/system_wrappers/include/logging.h"
|
|
|
|
|
#include "webrtc/system_wrappers/include/metrics.h"
|
2011-12-03 00:03:31 +00:00
|
|
|
|
|
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
|
|
|
|
// Files generated at build-time by the protobuf compiler.
|
2012-03-16 21:36:00 +00:00
|
|
|
#ifdef WEBRTC_ANDROID_PLATFORM_BUILD
|
2012-10-22 21:21:52 +00:00
|
|
|
#include "external/webrtc/webrtc/modules/audio_processing/debug.pb.h"
|
2011-08-03 23:34:31 +00:00
|
|
|
#else
|
2011-08-03 21:08:51 +00:00
|
|
|
#include "webrtc/audio_processing/debug.pb.h"
|
2011-08-03 23:34:31 +00:00
|
|
|
#endif
|
2011-12-03 00:03:31 +00:00
|
|
|
#endif // WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-07-23 11:41:39 -07:00
|
|
|
#define RETURN_ON_ERR(expr) \
|
|
|
|
|
do { \
|
|
|
|
|
int err = (expr); \
|
|
|
|
|
if (err != kNoError) { \
|
|
|
|
|
return err; \
|
|
|
|
|
} \
|
2014-01-07 17:45:09 +00:00
|
|
|
} while (0)
|
|
|
|
|
|
2011-07-07 08:21:25 +00:00
|
|
|
namespace webrtc {
|
2015-07-23 11:41:39 -07:00
|
|
|
namespace {
|
|
|
|
|
|
|
|
|
|
static bool LayoutHasKeyboard(AudioProcessing::ChannelLayout layout) {
|
|
|
|
|
switch (layout) {
|
|
|
|
|
case AudioProcessing::kMono:
|
|
|
|
|
case AudioProcessing::kStereo:
|
|
|
|
|
return false;
|
|
|
|
|
case AudioProcessing::kMonoAndKeyboard:
|
|
|
|
|
case AudioProcessing::kStereoAndKeyboard:
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
assert(false);
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
} // namespace
|
2014-01-07 17:45:09 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
struct AudioProcessingImpl::ApmPublicSubmodules {
|
|
|
|
|
ApmPublicSubmodules()
|
|
|
|
|
: echo_cancellation(nullptr),
|
|
|
|
|
echo_control_mobile(nullptr),
|
|
|
|
|
gain_control(nullptr),
|
|
|
|
|
high_pass_filter(nullptr),
|
|
|
|
|
level_estimator(nullptr),
|
|
|
|
|
noise_suppression(nullptr),
|
|
|
|
|
voice_detection(nullptr) {}
|
|
|
|
|
// Accessed externally of APM without any lock acquired.
|
|
|
|
|
EchoCancellationImpl* echo_cancellation;
|
|
|
|
|
EchoControlMobileImpl* echo_control_mobile;
|
|
|
|
|
GainControlImpl* gain_control;
|
|
|
|
|
HighPassFilterImpl* high_pass_filter;
|
|
|
|
|
LevelEstimatorImpl* level_estimator;
|
|
|
|
|
NoiseSuppressionImpl* noise_suppression;
|
|
|
|
|
VoiceDetectionImpl* voice_detection;
|
|
|
|
|
rtc::scoped_ptr<GainControlForNewAgc> gain_control_for_new_agc;
|
|
|
|
|
|
|
|
|
|
// Accessed internally from both render and capture.
|
|
|
|
|
rtc::scoped_ptr<TransientSuppressor> transient_suppressor;
|
|
|
|
|
rtc::scoped_ptr<IntelligibilityEnhancer> intelligibility_enhancer;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
struct AudioProcessingImpl::ApmPrivateSubmodules {
|
|
|
|
|
explicit ApmPrivateSubmodules(Beamformer<float>* beamformer)
|
|
|
|
|
: beamformer(beamformer) {}
|
|
|
|
|
// Accessed internally from capture or during initialization
|
|
|
|
|
std::list<ProcessingComponent*> component_list;
|
|
|
|
|
rtc::scoped_ptr<Beamformer<float>> beamformer;
|
|
|
|
|
rtc::scoped_ptr<AgcManagerDirect> agc_manager;
|
|
|
|
|
};
|
|
|
|
|
|
2014-01-07 17:45:09 +00:00
|
|
|
// Throughout webrtc, it's assumed that success is represented by zero.
|
2015-01-14 10:51:54 +00:00
|
|
|
static_assert(AudioProcessing::kNoError == 0, "kNoError must be zero");
|
2014-01-07 17:45:09 +00:00
|
|
|
|
2014-12-15 09:41:24 +00:00
|
|
|
// This class has two main functionalities:
|
|
|
|
|
//
|
|
|
|
|
// 1) It is returned instead of the real GainControl after the new AGC has been
|
|
|
|
|
// enabled in order to prevent an outside user from overriding compression
|
|
|
|
|
// settings. It doesn't do anything in its implementation, except for
|
|
|
|
|
// delegating the const methods and Enable calls to the real GainControl, so
|
|
|
|
|
// AGC can still be disabled.
|
|
|
|
|
//
|
|
|
|
|
// 2) It is injected into AgcManagerDirect and implements volume callbacks for
|
|
|
|
|
// getting and setting the volume level. It just caches this value to be used
|
|
|
|
|
// in VoiceEngine later.
|
|
|
|
|
class GainControlForNewAgc : public GainControl, public VolumeCallbacks {
|
|
|
|
|
public:
|
|
|
|
|
explicit GainControlForNewAgc(GainControlImpl* gain_control)
|
2015-07-23 11:41:39 -07:00
|
|
|
: real_gain_control_(gain_control), volume_(0) {}
|
2014-12-15 09:41:24 +00:00
|
|
|
|
|
|
|
|
// GainControl implementation.
|
2015-03-04 12:58:35 +00:00
|
|
|
int Enable(bool enable) override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return real_gain_control_->Enable(enable);
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
bool is_enabled() const override { return real_gain_control_->is_enabled(); }
|
|
|
|
|
int set_stream_analog_level(int level) override {
|
2014-12-15 09:41:24 +00:00
|
|
|
volume_ = level;
|
|
|
|
|
return AudioProcessing::kNoError;
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int stream_analog_level() override { return volume_; }
|
|
|
|
|
int set_mode(Mode mode) override { return AudioProcessing::kNoError; }
|
|
|
|
|
Mode mode() const override { return GainControl::kAdaptiveAnalog; }
|
|
|
|
|
int set_target_level_dbfs(int level) override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return AudioProcessing::kNoError;
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int target_level_dbfs() const override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return real_gain_control_->target_level_dbfs();
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int set_compression_gain_db(int gain) override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return AudioProcessing::kNoError;
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int compression_gain_db() const override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return real_gain_control_->compression_gain_db();
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int enable_limiter(bool enable) override { return AudioProcessing::kNoError; }
|
|
|
|
|
bool is_limiter_enabled() const override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return real_gain_control_->is_limiter_enabled();
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int set_analog_level_limits(int minimum, int maximum) override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return AudioProcessing::kNoError;
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int analog_level_minimum() const override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return real_gain_control_->analog_level_minimum();
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
int analog_level_maximum() const override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return real_gain_control_->analog_level_maximum();
|
|
|
|
|
}
|
2015-03-04 12:58:35 +00:00
|
|
|
bool stream_is_saturated() const override {
|
2014-12-15 09:41:24 +00:00
|
|
|
return real_gain_control_->stream_is_saturated();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// VolumeCallbacks implementation.
|
2015-03-04 12:58:35 +00:00
|
|
|
void SetMicVolume(int volume) override { volume_ = volume; }
|
|
|
|
|
int GetMicVolume() override { return volume_; }
|
2014-12-15 09:41:24 +00:00
|
|
|
|
|
|
|
|
private:
|
|
|
|
|
GainControl* real_gain_control_;
|
|
|
|
|
int volume_;
|
|
|
|
|
};
|
|
|
|
|
|
2015-09-23 12:49:12 -07:00
|
|
|
const int AudioProcessing::kNativeSampleRatesHz[] = {
|
|
|
|
|
AudioProcessing::kSampleRate8kHz,
|
|
|
|
|
AudioProcessing::kSampleRate16kHz,
|
|
|
|
|
AudioProcessing::kSampleRate32kHz,
|
|
|
|
|
AudioProcessing::kSampleRate48kHz};
|
|
|
|
|
const size_t AudioProcessing::kNumNativeSampleRates =
|
|
|
|
|
arraysize(AudioProcessing::kNativeSampleRatesHz);
|
|
|
|
|
const int AudioProcessing::kMaxNativeSampleRateHz = AudioProcessing::
|
|
|
|
|
kNativeSampleRatesHz[AudioProcessing::kNumNativeSampleRates - 1];
|
|
|
|
|
const int AudioProcessing::kMaxAECMSampleRateHz = kSampleRate16kHz;
|
|
|
|
|
|
2014-01-25 02:09:06 +00:00
|
|
|
AudioProcessing* AudioProcessing::Create() {
|
|
|
|
|
Config config;
|
2015-01-15 18:07:21 +00:00
|
|
|
return Create(config, nullptr);
|
2014-01-25 02:09:06 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AudioProcessing* AudioProcessing::Create(const Config& config) {
|
2015-01-15 18:07:21 +00:00
|
|
|
return Create(config, nullptr);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AudioProcessing* AudioProcessing::Create(const Config& config,
|
2015-03-25 16:37:27 -07:00
|
|
|
Beamformer<float>* beamformer) {
|
2015-01-15 18:07:21 +00:00
|
|
|
AudioProcessingImpl* apm = new AudioProcessingImpl(config, beamformer);
|
2011-07-07 08:21:25 +00:00
|
|
|
if (apm->Initialize() != kNoError) {
|
|
|
|
|
delete apm;
|
2015-11-28 12:35:15 -08:00
|
|
|
apm = nullptr;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return apm;
|
|
|
|
|
}
|
|
|
|
|
|
2014-01-25 02:09:06 +00:00
|
|
|
AudioProcessingImpl::AudioProcessingImpl(const Config& config)
|
2015-01-15 18:07:21 +00:00
|
|
|
: AudioProcessingImpl(config, nullptr) {}
|
|
|
|
|
|
|
|
|
|
AudioProcessingImpl::AudioProcessingImpl(const Config& config,
|
2015-03-25 16:37:27 -07:00
|
|
|
Beamformer<float>* beamformer)
|
2015-11-28 12:35:15 -08:00
|
|
|
: public_submodules_(new ApmPublicSubmodules()),
|
|
|
|
|
private_submodules_(new ApmPrivateSubmodules(beamformer)),
|
|
|
|
|
constants_(config.Get<ExperimentalAgc>().startup_min_volume,
|
|
|
|
|
config.Get<Beamforming>().array_geometry,
|
|
|
|
|
config.Get<Beamforming>().target_direction,
|
2014-12-15 09:41:24 +00:00
|
|
|
#if defined(WEBRTC_ANDROID) || defined(WEBRTC_IOS)
|
2015-11-28 12:35:15 -08:00
|
|
|
false,
|
2014-12-15 09:41:24 +00:00
|
|
|
#else
|
2015-11-28 12:35:15 -08:00
|
|
|
config.Get<ExperimentalAgc>().enabled,
|
2014-12-15 09:41:24 +00:00
|
|
|
#endif
|
2015-11-28 12:35:15 -08:00
|
|
|
config.Get<Intelligibility>().enabled,
|
|
|
|
|
config.Get<Beamforming>().enabled),
|
|
|
|
|
|
2015-06-24 18:14:14 -07:00
|
|
|
#if defined(WEBRTC_ANDROID) || defined(WEBRTC_IOS)
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_(false)
|
2015-06-24 18:14:14 -07:00
|
|
|
#else
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_(config.Get<ExperimentalNs>().enabled)
|
2015-06-24 18:14:14 -07:00
|
|
|
#endif
|
2015-11-28 12:35:15 -08:00
|
|
|
{
|
|
|
|
|
{
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
|
|
|
|
|
|
|
|
|
public_submodules_->echo_cancellation =
|
|
|
|
|
new EchoCancellationImpl(this, &crit_render_, &crit_capture_);
|
|
|
|
|
public_submodules_->echo_control_mobile =
|
|
|
|
|
new EchoControlMobileImpl(this, &crit_render_, &crit_capture_);
|
|
|
|
|
public_submodules_->gain_control =
|
|
|
|
|
new GainControlImpl(this, &crit_capture_, &crit_capture_);
|
|
|
|
|
public_submodules_->high_pass_filter =
|
|
|
|
|
new HighPassFilterImpl(this, &crit_capture_);
|
|
|
|
|
public_submodules_->level_estimator =
|
|
|
|
|
new LevelEstimatorImpl(this, &crit_capture_);
|
|
|
|
|
public_submodules_->noise_suppression =
|
|
|
|
|
new NoiseSuppressionImpl(this, &crit_capture_);
|
|
|
|
|
public_submodules_->voice_detection =
|
|
|
|
|
new VoiceDetectionImpl(this, &crit_capture_);
|
|
|
|
|
public_submodules_->gain_control_for_new_agc.reset(
|
|
|
|
|
new GainControlForNewAgc(public_submodules_->gain_control));
|
|
|
|
|
|
|
|
|
|
private_submodules_->component_list.push_back(
|
|
|
|
|
public_submodules_->echo_cancellation);
|
|
|
|
|
private_submodules_->component_list.push_back(
|
|
|
|
|
public_submodules_->echo_control_mobile);
|
|
|
|
|
private_submodules_->component_list.push_back(
|
|
|
|
|
public_submodules_->gain_control);
|
|
|
|
|
private_submodules_->component_list.push_back(
|
|
|
|
|
public_submodules_->high_pass_filter);
|
|
|
|
|
private_submodules_->component_list.push_back(
|
|
|
|
|
public_submodules_->level_estimator);
|
|
|
|
|
private_submodules_->component_list.push_back(
|
|
|
|
|
public_submodules_->noise_suppression);
|
|
|
|
|
private_submodules_->component_list.push_back(
|
|
|
|
|
public_submodules_->voice_detection);
|
|
|
|
|
}
|
2014-12-15 09:41:24 +00:00
|
|
|
|
2014-01-25 02:09:06 +00:00
|
|
|
SetExtraOptions(config);
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AudioProcessingImpl::~AudioProcessingImpl() {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Depends on gain_control_ and
|
|
|
|
|
// public_submodules_->gain_control_for_new_agc.
|
|
|
|
|
private_submodules_->agc_manager.reset();
|
|
|
|
|
// Depends on gain_control_.
|
|
|
|
|
public_submodules_->gain_control_for_new_agc.reset();
|
|
|
|
|
while (!private_submodules_->component_list.empty()) {
|
|
|
|
|
ProcessingComponent* component =
|
|
|
|
|
private_submodules_->component_list.front();
|
|
|
|
|
component->Destroy();
|
|
|
|
|
delete component;
|
|
|
|
|
private_submodules_->component_list.pop_front();
|
|
|
|
|
}
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
debug_dump_.debug_file->CloseFile();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
#endif
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::Initialize() {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner during initialization.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2011-07-07 08:21:25 +00:00
|
|
|
return InitializeLocked();
|
|
|
|
|
}
|
|
|
|
|
|
2014-04-22 21:00:04 +00:00
|
|
|
int AudioProcessingImpl::Initialize(int input_sample_rate_hz,
|
|
|
|
|
int output_sample_rate_hz,
|
2014-03-10 22:26:12 +00:00
|
|
|
int reverse_sample_rate_hz,
|
2014-04-22 21:00:04 +00:00
|
|
|
ChannelLayout input_layout,
|
|
|
|
|
ChannelLayout output_layout,
|
|
|
|
|
ChannelLayout reverse_layout) {
|
2015-07-23 11:41:39 -07:00
|
|
|
const ProcessingConfig processing_config = {
|
2015-08-14 10:35:55 -07:00
|
|
|
{{input_sample_rate_hz,
|
|
|
|
|
ChannelsFromLayout(input_layout),
|
2015-07-23 11:41:39 -07:00
|
|
|
LayoutHasKeyboard(input_layout)},
|
2015-08-14 10:35:55 -07:00
|
|
|
{output_sample_rate_hz,
|
|
|
|
|
ChannelsFromLayout(output_layout),
|
2015-07-23 11:41:39 -07:00
|
|
|
LayoutHasKeyboard(output_layout)},
|
2015-08-14 10:35:55 -07:00
|
|
|
{reverse_sample_rate_hz,
|
|
|
|
|
ChannelsFromLayout(reverse_layout),
|
|
|
|
|
LayoutHasKeyboard(reverse_layout)},
|
|
|
|
|
{reverse_sample_rate_hz,
|
|
|
|
|
ChannelsFromLayout(reverse_layout),
|
2015-07-23 11:41:39 -07:00
|
|
|
LayoutHasKeyboard(reverse_layout)}}};
|
|
|
|
|
|
|
|
|
|
return Initialize(processing_config);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::Initialize(const ProcessingConfig& processing_config) {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner during initialization.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2015-07-23 11:41:39 -07:00
|
|
|
return InitializeLocked(processing_config);
|
2014-03-10 22:26:12 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
int AudioProcessingImpl::MaybeInitializeRender(
|
2015-11-27 02:47:28 -08:00
|
|
|
const ProcessingConfig& processing_config) {
|
2015-11-28 12:35:15 -08:00
|
|
|
return MaybeInitialize(processing_config);
|
2015-11-27 02:47:28 -08:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
int AudioProcessingImpl::MaybeInitializeCapture(
|
2015-11-27 02:47:28 -08:00
|
|
|
const ProcessingConfig& processing_config) {
|
2015-11-28 12:35:15 -08:00
|
|
|
return MaybeInitialize(processing_config);
|
2015-11-27 02:47:28 -08:00
|
|
|
}
|
|
|
|
|
|
2015-11-17 02:16:45 -08:00
|
|
|
// Calls InitializeLocked() if any of the audio parameters have changed from
|
2015-11-28 12:35:15 -08:00
|
|
|
// their current values (needs to be called while holding the crit_render_lock).
|
|
|
|
|
int AudioProcessingImpl::MaybeInitialize(
|
2015-11-17 02:16:45 -08:00
|
|
|
const ProcessingConfig& processing_config) {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Called from both threads. Thread check is therefore not possible.
|
|
|
|
|
if (processing_config == formats_.api_format) {
|
2015-11-17 02:16:45 -08:00
|
|
|
return kNoError;
|
|
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2015-11-17 02:16:45 -08:00
|
|
|
return InitializeLocked(processing_config);
|
|
|
|
|
}
|
|
|
|
|
|
2011-07-07 08:21:25 +00:00
|
|
|
int AudioProcessingImpl::InitializeLocked() {
|
2015-07-23 11:41:39 -07:00
|
|
|
const int fwd_audio_buffer_channels =
|
2015-11-28 12:35:15 -08:00
|
|
|
constants_.beamformer_enabled
|
|
|
|
|
? formats_.api_format.input_stream().num_channels()
|
|
|
|
|
: formats_.api_format.output_stream().num_channels();
|
2015-08-14 10:35:55 -07:00
|
|
|
const int rev_audio_buffer_out_num_frames =
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.reverse_output_stream().num_frames() == 0
|
|
|
|
|
? formats_.rev_proc_format.num_frames()
|
|
|
|
|
: formats_.api_format.reverse_output_stream().num_frames();
|
|
|
|
|
if (formats_.api_format.reverse_input_stream().num_channels() > 0) {
|
|
|
|
|
render_.render_audio.reset(new AudioBuffer(
|
|
|
|
|
formats_.api_format.reverse_input_stream().num_frames(),
|
|
|
|
|
formats_.api_format.reverse_input_stream().num_channels(),
|
|
|
|
|
formats_.rev_proc_format.num_frames(),
|
|
|
|
|
formats_.rev_proc_format.num_channels(),
|
2015-08-14 10:35:55 -07:00
|
|
|
rev_audio_buffer_out_num_frames));
|
|
|
|
|
if (rev_conversion_needed()) {
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_converter = AudioConverter::Create(
|
|
|
|
|
formats_.api_format.reverse_input_stream().num_channels(),
|
|
|
|
|
formats_.api_format.reverse_input_stream().num_frames(),
|
|
|
|
|
formats_.api_format.reverse_output_stream().num_channels(),
|
|
|
|
|
formats_.api_format.reverse_output_stream().num_frames());
|
2015-08-14 10:35:55 -07:00
|
|
|
} else {
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_converter.reset(nullptr);
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
2015-07-23 11:41:39 -07:00
|
|
|
} else {
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_audio.reset(nullptr);
|
|
|
|
|
render_.render_converter.reset(nullptr);
|
2015-07-23 11:41:39 -07:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.capture_audio.reset(
|
|
|
|
|
new AudioBuffer(formats_.api_format.input_stream().num_frames(),
|
|
|
|
|
formats_.api_format.input_stream().num_channels(),
|
|
|
|
|
capture_nonlocked_.fwd_proc_format.num_frames(),
|
|
|
|
|
fwd_audio_buffer_channels,
|
|
|
|
|
formats_.api_format.output_stream().num_frames()));
|
2011-07-07 08:21:25 +00:00
|
|
|
|
|
|
|
|
// Initialize all components.
|
2015-11-28 12:35:15 -08:00
|
|
|
for (auto item : private_submodules_->component_list) {
|
2015-03-12 23:23:38 +00:00
|
|
|
int err = item->Initialize();
|
2011-07-07 08:21:25 +00:00
|
|
|
if (err != kNoError) {
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-15 11:42:40 +02:00
|
|
|
InitializeExperimentalAgc();
|
2014-12-15 09:41:24 +00:00
|
|
|
|
2015-04-15 11:42:40 +02:00
|
|
|
InitializeTransient();
|
2014-12-15 09:41:24 +00:00
|
|
|
|
2014-12-19 19:57:34 +00:00
|
|
|
InitializeBeamformer();
|
|
|
|
|
|
2015-08-14 10:35:55 -07:00
|
|
|
InitializeIntelligibility();
|
|
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
2011-08-03 21:08:51 +00:00
|
|
|
int err = WriteInitMessage();
|
|
|
|
|
if (err != kNoError) {
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
}
|
2011-12-03 00:03:31 +00:00
|
|
|
#endif
|
2011-08-03 21:08:51 +00:00
|
|
|
|
2011-07-07 08:21:25 +00:00
|
|
|
return kNoError;
|
|
|
|
|
}
|
|
|
|
|
|
2015-07-23 11:41:39 -07:00
|
|
|
int AudioProcessingImpl::InitializeLocked(const ProcessingConfig& config) {
|
|
|
|
|
for (const auto& stream : config.streams) {
|
|
|
|
|
if (stream.num_channels() < 0) {
|
|
|
|
|
return kBadNumberChannelsError;
|
|
|
|
|
}
|
|
|
|
|
if (stream.num_channels() > 0 && stream.sample_rate_hz() <= 0) {
|
|
|
|
|
return kBadSampleRateError;
|
|
|
|
|
}
|
Revert of Allow more than 2 input channels in AudioProcessing. (patchset #13 id:240001 of https://codereview.webrtc.org/1226093007/)
Reason for revert:
Breaks Chromium FYI content_browsertest on all platforms. The testcase that fails is WebRtcAecDumpBrowserTest.CallWithAecDump.
https://build.chromium.org/p/chromium.webrtc.fyi/builders/Linux/builds/19388
Sample output:
[ RUN ] WebRtcAecDumpBrowserTest.CallWithAecDump
Xlib: extension "RANDR" missing on display ":9".
[4:14:0722/211548:1282124453:WARNING:webrtcvoiceengine.cc(472)] Unexpected codec: ISAC/48000/1 (105)
[4:14:0722/211548:1282124593:WARNING:webrtcvoiceengine.cc(472)] Unexpected codec: PCMU/8000/2 (110)
[4:14:0722/211548:1282124700:WARNING:webrtcvoiceengine.cc(472)] Unexpected codec: PCMA/8000/2 (118)
[4:14:0722/211548:1282124815:WARNING:webrtcvoiceengine.cc(472)] Unexpected codec: G722/8000/2 (119)
[19745:19745:0722/211548:1282133667:INFO:CONSOLE(64)] "Looking at video in element remote-view-1", source: http://127.0.0.1:48819/media/webrtc_test_utilities.js (64)
[19745:19745:0722/211548:1282136892:INFO:CONSOLE(64)] "Looking at video in element remote-view-2", source: http://127.0.0.1:48819/media/webrtc_test_utilities.js (64)
../../content/test/webrtc_content_browsertest_base.cc:62: Failure
Value of: ExecuteScriptAndExtractString( shell()->web_contents(), javascript, &result)
Actual: false
Expected: true
Failed to execute javascript call({video: true, audio: true});.
From javascript: (nothing)
When executing 'call({video: true, audio: true});'
../../content/test/webrtc_content_browsertest_base.cc:75: Failure
Failed
../../content/browser/media/webrtc_aecdump_browsertest.cc:26: Failure
Expected: (base::kNullProcessId) != (*id), actual: 0 vs 0
../../content/browser/media/webrtc_aecdump_browsertest.cc:95: Failure
Value of: GetRenderProcessHostId(&render_process_id)
Actual: false
Expected: true
../../content/browser/media/webrtc_aecdump_browsertest.cc:99: Failure
Value of: base::PathExists(dump_file)
Actual: false
Expected: true
../../content/browser/media/webrtc_aecdump_browsertest.cc:101: Failure
Value of: base::GetFileSize(dump_file, &file_size)
Actual: false
Expected: true
../../content/browser/media/webrtc_aecdump_browsertest.cc:102: Failure
Expected: (file_size) > (0), actual: 0 vs 0
[ FAILED ] WebRtcAecDumpBrowserTest.CallWithAecDump, where TypeParam = and GetParam() = (361 ms)
Original issue's description:
> Allow more than 2 input channels in AudioProcessing.
>
> The number of output channels is constrained to be equal to either 1 or the
> number of input channels.
>
> R=aluebs@webrtc.org, andrew@webrtc.org, pbos@webrtc.org
>
> Committed: https://chromium.googlesource.com/external/webrtc/+/c204754b7a0cc801c70e8ce6c689f57f6ce00b3b
TBR=andrew@webrtc.org,aluebs@webrtc.org,ajm@chromium.org,pbos@chromium.org,pbos@webrtc.org,mgraczyk@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.webrtc.org/1253573005
Cr-Commit-Position: refs/heads/master@{#9621}
2015-07-23 04:30:06 -07:00
|
|
|
}
|
2015-07-23 11:41:39 -07:00
|
|
|
|
|
|
|
|
const int num_in_channels = config.input_stream().num_channels();
|
|
|
|
|
const int num_out_channels = config.output_stream().num_channels();
|
|
|
|
|
|
|
|
|
|
// Need at least one input channel.
|
|
|
|
|
// Need either one output channel or as many outputs as there are inputs.
|
|
|
|
|
if (num_in_channels == 0 ||
|
|
|
|
|
!(num_out_channels == 1 || num_out_channels == num_in_channels)) {
|
2014-03-10 22:26:12 +00:00
|
|
|
return kBadNumberChannelsError;
|
|
|
|
|
}
|
2015-07-23 11:41:39 -07:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.beamformer_enabled && (static_cast<size_t>(num_in_channels) !=
|
|
|
|
|
constants_.array_geometry.size() ||
|
|
|
|
|
num_out_channels > 1)) {
|
2015-01-15 18:07:21 +00:00
|
|
|
return kBadNumberChannelsError;
|
|
|
|
|
}
|
2014-04-22 21:00:04 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format = config;
|
2014-04-22 21:00:04 +00:00
|
|
|
|
|
|
|
|
// We process at the closest native rate >= min(input rate, output rate)...
|
2015-07-23 11:41:39 -07:00
|
|
|
const int min_proc_rate =
|
2015-11-28 12:35:15 -08:00
|
|
|
std::min(formats_.api_format.input_stream().sample_rate_hz(),
|
|
|
|
|
formats_.api_format.output_stream().sample_rate_hz());
|
2014-04-22 21:00:04 +00:00
|
|
|
int fwd_proc_rate;
|
2015-09-23 12:49:12 -07:00
|
|
|
for (size_t i = 0; i < kNumNativeSampleRates; ++i) {
|
|
|
|
|
fwd_proc_rate = kNativeSampleRatesHz[i];
|
|
|
|
|
if (fwd_proc_rate >= min_proc_rate) {
|
|
|
|
|
break;
|
|
|
|
|
}
|
2014-04-22 21:00:04 +00:00
|
|
|
}
|
|
|
|
|
// ...with one exception.
|
2015-11-28 12:35:15 -08:00
|
|
|
if (public_submodules_->echo_control_mobile->is_enabled() &&
|
2015-09-23 12:49:12 -07:00
|
|
|
min_proc_rate > kMaxAECMSampleRateHz) {
|
|
|
|
|
fwd_proc_rate = kMaxAECMSampleRateHz;
|
2014-04-22 21:00:04 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_nonlocked_.fwd_proc_format = StreamConfig(fwd_proc_rate);
|
2014-04-22 21:00:04 +00:00
|
|
|
|
|
|
|
|
// We normally process the reverse stream at 16 kHz. Unless...
|
|
|
|
|
int rev_proc_rate = kSampleRate16kHz;
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_nonlocked_.fwd_proc_format.sample_rate_hz() == kSampleRate8kHz) {
|
2014-04-22 21:00:04 +00:00
|
|
|
// ...the forward stream is at 8 kHz.
|
|
|
|
|
rev_proc_rate = kSampleRate8kHz;
|
|
|
|
|
} else {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (formats_.api_format.reverse_input_stream().sample_rate_hz() ==
|
2015-08-14 10:35:55 -07:00
|
|
|
kSampleRate32kHz) {
|
2014-04-22 21:00:04 +00:00
|
|
|
// ...or the input is at 32 kHz, in which case we use the splitting
|
|
|
|
|
// filter rather than the resampler.
|
|
|
|
|
rev_proc_rate = kSampleRate32kHz;
|
|
|
|
|
}
|
2014-03-10 22:26:12 +00:00
|
|
|
}
|
|
|
|
|
|
Enable render downmixing to mono in AudioProcessing.
In practice, we have been doing this since time immemorial, but have
relied on the user to do the downmixing (first voice engine then
Chromium). It's more logical for this burden to fall on AudioProcessing,
however, who can be expected to know that this is a reasonable approach
for AEC. Permitting two render channels results in running two AECs
serially.
Critically, in my recent change to have Chromium adopt the float
interface:
https://codereview.chromium.org/420603004
I removed the downmixing by Chromium, forgetting that we hadn't yet
enabled this feature in AudioProcessing. This corrects that oversight.
The change in paths hit by production users is very minor. As commented
it required adding downmixing to the int16_t path to satisfy
bit-exactness tests.
For reference, find the ApmTest.Process errors here:
https://paste.googleplex.com/6372007910309888
BUG=webrtc:3853
TESTED=listened to the files output from the Process test, and verified
that they sound as expected: higher echo while the AEC is adapting, but
afterwards very close.
R=aluebs@webrtc.org, bjornv@webrtc.org, kwiberg@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/31459004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@7292 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-09-24 20:06:23 +00:00
|
|
|
// Always downmix the reverse stream to mono for analysis. This has been
|
|
|
|
|
// demonstrated to work well for AEC in most practical scenarios.
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.rev_proc_format = StreamConfig(rev_proc_rate, 1);
|
2014-03-10 22:26:12 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_nonlocked_.fwd_proc_format.sample_rate_hz() == kSampleRate32kHz ||
|
|
|
|
|
capture_nonlocked_.fwd_proc_format.sample_rate_hz() == kSampleRate48kHz) {
|
|
|
|
|
capture_nonlocked_.split_rate = kSampleRate16kHz;
|
2014-03-10 22:26:12 +00:00
|
|
|
} else {
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_nonlocked_.split_rate =
|
|
|
|
|
capture_nonlocked_.fwd_proc_format.sample_rate_hz();
|
2014-03-10 22:26:12 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return InitializeLocked();
|
|
|
|
|
}
|
|
|
|
|
|
2013-07-25 18:28:29 +00:00
|
|
|
void AudioProcessingImpl::SetExtraOptions(const Config& config) {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner when setting the extra options.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
|
|
|
|
for (auto item : private_submodules_->component_list) {
|
2015-03-12 23:23:38 +00:00
|
|
|
item->SetExtraOptions(config);
|
|
|
|
|
}
|
2014-12-15 09:41:24 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.transient_suppressor_enabled !=
|
|
|
|
|
config.Get<ExperimentalNs>().enabled) {
|
|
|
|
|
capture_.transient_suppressor_enabled =
|
|
|
|
|
config.Get<ExperimentalNs>().enabled;
|
2014-12-15 09:41:24 +00:00
|
|
|
InitializeTransient();
|
|
|
|
|
}
|
2013-07-25 18:28:29 +00:00
|
|
|
}
|
|
|
|
|
|
2014-04-22 21:00:04 +00:00
|
|
|
int AudioProcessingImpl::proc_sample_rate_hz() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Used as callback from submodules, hence locking is not allowed.
|
|
|
|
|
return capture_nonlocked_.fwd_proc_format.sample_rate_hz();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2014-04-22 21:00:04 +00:00
|
|
|
int AudioProcessingImpl::proc_split_sample_rate_hz() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Used as callback from submodules, hence locking is not allowed.
|
|
|
|
|
return capture_nonlocked_.split_rate;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::num_reverse_channels() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Used as callback from submodules, hence locking is not allowed.
|
|
|
|
|
return formats_.rev_proc_format.num_channels();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::num_input_channels() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Used as callback from submodules, hence locking is not allowed.
|
|
|
|
|
return formats_.api_format.input_stream().num_channels();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::num_output_channels() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Used as callback from submodules, hence locking is not allowed.
|
|
|
|
|
return formats_.api_format.output_stream().num_channels();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-12 22:28:31 +00:00
|
|
|
void AudioProcessingImpl::set_output_will_be_muted(bool muted) {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_capture_);
|
|
|
|
|
capture_.output_will_be_muted = muted;
|
|
|
|
|
if (private_submodules_->agc_manager.get()) {
|
|
|
|
|
private_submodules_->agc_manager->SetCaptureMuted(
|
|
|
|
|
capture_.output_will_be_muted);
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
2014-02-12 22:28:31 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2014-04-22 21:00:04 +00:00
|
|
|
int AudioProcessingImpl::ProcessStream(const float* const* src,
|
Update a ton of audio code to use size_t more correctly and in general reduce
use of int16_t/uint16_t.
This is the upshot of a recommendation by henrik.lundin and kwiberg on an original small change ( https://webrtc-codereview.appspot.com/42569004/#ps1 ) to stop using int16_t just because values could fit in it, and is similar in nature to a previous "mass change to use size_t more" ( https://webrtc-codereview.appspot.com/23129004/ ) which also needed to be split up for review but to land all at once, since, like adding "const", such changes tend to cause a lot of transitive effects.
This was be reviewed and approved in pieces:
https://codereview.webrtc.org/1224093003
https://codereview.webrtc.org/1224123002
https://codereview.webrtc.org/1224163002
https://codereview.webrtc.org/1225133003
https://codereview.webrtc.org/1225173002
https://codereview.webrtc.org/1227163003
https://codereview.webrtc.org/1227203003
https://codereview.webrtc.org/1227213002
https://codereview.webrtc.org/1227893002
https://codereview.webrtc.org/1228793004
https://codereview.webrtc.org/1228803003
https://codereview.webrtc.org/1228823002
https://codereview.webrtc.org/1228823003
https://codereview.webrtc.org/1228843002
https://codereview.webrtc.org/1230693002
https://codereview.webrtc.org/1231713002
The change is being landed as TBR to all the folks who reviewed the above.
BUG=chromium:81439
TEST=none
R=andrew@webrtc.org, pbos@webrtc.org
TBR=aluebs, andrew, asapersson, henrika, hlundin, jan.skoglund, kwiberg, minyue, pbos, pthatcher
Review URL: https://codereview.webrtc.org/1230503003 .
Cr-Commit-Position: refs/heads/master@{#9768}
2015-08-24 14:52:23 -07:00
|
|
|
size_t samples_per_channel,
|
2014-04-22 21:00:04 +00:00
|
|
|
int input_sample_rate_hz,
|
2014-03-04 20:58:13 +00:00
|
|
|
ChannelLayout input_layout,
|
2014-04-22 21:00:04 +00:00
|
|
|
int output_sample_rate_hz,
|
|
|
|
|
ChannelLayout output_layout,
|
|
|
|
|
float* const* dest) {
|
2015-11-28 12:35:15 -08:00
|
|
|
StreamConfig input_stream;
|
|
|
|
|
StreamConfig output_stream;
|
|
|
|
|
{
|
|
|
|
|
// Access the formats_.api_format.input_stream beneath the capture lock.
|
|
|
|
|
// The lock must be released as it is later required in the call
|
|
|
|
|
// to ProcessStream(,,,);
|
|
|
|
|
rtc::CritScope cs(&crit_capture_);
|
|
|
|
|
input_stream = formats_.api_format.input_stream();
|
|
|
|
|
output_stream = formats_.api_format.output_stream();
|
|
|
|
|
}
|
|
|
|
|
|
2015-07-23 11:41:39 -07:00
|
|
|
input_stream.set_sample_rate_hz(input_sample_rate_hz);
|
|
|
|
|
input_stream.set_num_channels(ChannelsFromLayout(input_layout));
|
|
|
|
|
input_stream.set_has_keyboard(LayoutHasKeyboard(input_layout));
|
|
|
|
|
output_stream.set_sample_rate_hz(output_sample_rate_hz);
|
|
|
|
|
output_stream.set_num_channels(ChannelsFromLayout(output_layout));
|
|
|
|
|
output_stream.set_has_keyboard(LayoutHasKeyboard(output_layout));
|
|
|
|
|
|
|
|
|
|
if (samples_per_channel != input_stream.num_frames()) {
|
|
|
|
|
return kBadDataLengthError;
|
|
|
|
|
}
|
|
|
|
|
return ProcessStream(src, input_stream, output_stream, dest);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::ProcessStream(const float* const* src,
|
|
|
|
|
const StreamConfig& input_config,
|
|
|
|
|
const StreamConfig& output_config,
|
|
|
|
|
float* const* dest) {
|
2015-11-28 12:35:15 -08:00
|
|
|
ProcessingConfig processing_config;
|
|
|
|
|
{
|
|
|
|
|
// Acquire the capture lock in order to safely call the function
|
|
|
|
|
// that retrieves the render side data. This function accesses apm
|
|
|
|
|
// getters that need the capture lock held when being called.
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
|
|
|
|
public_submodules_->echo_cancellation->ReadQueuedRenderData();
|
|
|
|
|
public_submodules_->echo_control_mobile->ReadQueuedRenderData();
|
|
|
|
|
public_submodules_->gain_control->ReadQueuedRenderData();
|
|
|
|
|
|
|
|
|
|
if (!src || !dest) {
|
|
|
|
|
return kNullPointerError;
|
|
|
|
|
}
|
2014-01-07 17:45:09 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
processing_config = formats_.api_format;
|
|
|
|
|
}
|
2015-11-16 16:27:42 -08:00
|
|
|
|
2015-07-23 11:41:39 -07:00
|
|
|
processing_config.input_stream() = input_config;
|
|
|
|
|
processing_config.output_stream() = output_config;
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
{
|
|
|
|
|
// Do conditional reinitialization.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
RETURN_ON_ERR(MaybeInitializeCapture(processing_config));
|
|
|
|
|
}
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2015-07-23 11:41:39 -07:00
|
|
|
assert(processing_config.input_stream().num_frames() ==
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.input_stream().num_frames());
|
2014-03-04 20:58:13 +00:00
|
|
|
|
|
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
2015-10-03 00:39:14 +02:00
|
|
|
RETURN_ON_ERR(WriteConfigMessage(false));
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
debug_dump_.capture.event_msg->set_type(audioproc::Event::STREAM);
|
|
|
|
|
audioproc::Stream* msg = debug_dump_.capture.event_msg->mutable_stream();
|
2014-08-28 10:43:09 +00:00
|
|
|
const size_t channel_size =
|
2015-11-28 12:35:15 -08:00
|
|
|
sizeof(float) * formats_.api_format.input_stream().num_frames();
|
|
|
|
|
for (int i = 0; i < formats_.api_format.input_stream().num_channels(); ++i)
|
2014-04-22 21:00:04 +00:00
|
|
|
msg->add_input_channel(src[i], channel_size);
|
2014-03-04 20:58:13 +00:00
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.capture_audio->CopyFrom(src, formats_.api_format.input_stream());
|
2014-03-04 20:58:13 +00:00
|
|
|
RETURN_ON_ERR(ProcessStreamLocked());
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.capture_audio->CopyTo(formats_.api_format.output_stream(), dest);
|
2014-03-04 20:58:13 +00:00
|
|
|
|
|
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
audioproc::Stream* msg = debug_dump_.capture.event_msg->mutable_stream();
|
2014-08-28 10:43:09 +00:00
|
|
|
const size_t channel_size =
|
2015-11-28 12:35:15 -08:00
|
|
|
sizeof(float) * formats_.api_format.output_stream().num_frames();
|
|
|
|
|
for (int i = 0; i < formats_.api_format.output_stream().num_channels(); ++i)
|
2014-04-22 21:00:04 +00:00
|
|
|
msg->add_output_channel(dest[i], channel_size);
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(WriteMessageToDebugFile(debug_dump_.debug_file.get(),
|
|
|
|
|
&crit_debug_, &debug_dump_.capture));
|
2014-03-04 20:58:13 +00:00
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
return kNoError;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::ProcessStream(AudioFrame* frame) {
|
2015-11-28 12:35:15 -08:00
|
|
|
{
|
|
|
|
|
// Acquire the capture lock in order to safely call the function
|
|
|
|
|
// that retrieves the render side data. This function accesses apm
|
|
|
|
|
// getters that need the capture lock held when being called.
|
|
|
|
|
// The lock needs to be released as
|
|
|
|
|
// public_submodules_->echo_control_mobile->is_enabled() aquires this lock
|
|
|
|
|
// as well.
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
|
|
|
|
public_submodules_->echo_cancellation->ReadQueuedRenderData();
|
|
|
|
|
public_submodules_->echo_control_mobile->ReadQueuedRenderData();
|
|
|
|
|
public_submodules_->gain_control->ReadQueuedRenderData();
|
|
|
|
|
}
|
2015-11-16 16:27:42 -08:00
|
|
|
|
2014-03-04 20:58:13 +00:00
|
|
|
if (!frame) {
|
2014-01-07 17:45:09 +00:00
|
|
|
return kNullPointerError;
|
|
|
|
|
}
|
2014-04-22 21:00:04 +00:00
|
|
|
// Must be a native rate.
|
|
|
|
|
if (frame->sample_rate_hz_ != kSampleRate8kHz &&
|
|
|
|
|
frame->sample_rate_hz_ != kSampleRate16kHz &&
|
2014-11-17 23:01:23 +00:00
|
|
|
frame->sample_rate_hz_ != kSampleRate32kHz &&
|
|
|
|
|
frame->sample_rate_hz_ != kSampleRate48kHz) {
|
2014-04-22 21:00:04 +00:00
|
|
|
return kBadSampleRateError;
|
|
|
|
|
}
|
2015-11-17 02:16:45 -08:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (public_submodules_->echo_control_mobile->is_enabled() &&
|
2015-09-23 12:49:12 -07:00
|
|
|
frame->sample_rate_hz_ > kMaxAECMSampleRateHz) {
|
2014-04-22 21:00:04 +00:00
|
|
|
LOG(LS_ERROR) << "AECM only supports 16 or 8 kHz sample rates";
|
|
|
|
|
return kUnsupportedComponentError;
|
|
|
|
|
}
|
2014-03-04 20:58:13 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
ProcessingConfig processing_config;
|
|
|
|
|
{
|
|
|
|
|
// Aquire lock for the access of api_format.
|
|
|
|
|
// The lock is released immediately due to the conditional
|
|
|
|
|
// reinitialization.
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
|
|
|
|
// TODO(ajm): The input and output rates and channels are currently
|
|
|
|
|
// constrained to be identical in the int16 interface.
|
|
|
|
|
processing_config = formats_.api_format;
|
|
|
|
|
}
|
2015-07-23 11:41:39 -07:00
|
|
|
processing_config.input_stream().set_sample_rate_hz(frame->sample_rate_hz_);
|
|
|
|
|
processing_config.input_stream().set_num_channels(frame->num_channels_);
|
|
|
|
|
processing_config.output_stream().set_sample_rate_hz(frame->sample_rate_hz_);
|
|
|
|
|
processing_config.output_stream().set_num_channels(frame->num_channels_);
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
{
|
|
|
|
|
// Do conditional reinitialization.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
RETURN_ON_ERR(MaybeInitializeCapture(processing_config));
|
|
|
|
|
}
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2015-11-17 02:16:45 -08:00
|
|
|
if (frame->samples_per_channel_ !=
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.input_stream().num_frames()) {
|
2011-07-07 08:21:25 +00:00
|
|
|
return kBadDataLengthError;
|
|
|
|
|
}
|
|
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
debug_dump_.capture.event_msg->set_type(audioproc::Event::STREAM);
|
|
|
|
|
audioproc::Stream* msg = debug_dump_.capture.event_msg->mutable_stream();
|
2015-07-23 11:41:39 -07:00
|
|
|
const size_t data_size =
|
|
|
|
|
sizeof(int16_t) * frame->samples_per_channel_ * frame->num_channels_;
|
2012-05-02 23:56:37 +00:00
|
|
|
msg->set_input_data(frame->data_, data_size);
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
2011-12-03 00:03:31 +00:00
|
|
|
#endif
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.capture_audio->DeinterleaveFrom(frame);
|
2014-03-04 20:58:13 +00:00
|
|
|
RETURN_ON_ERR(ProcessStreamLocked());
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.capture_audio->InterleaveTo(frame,
|
|
|
|
|
output_copy_needed(is_data_processed()));
|
2014-03-04 20:58:13 +00:00
|
|
|
|
|
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
audioproc::Stream* msg = debug_dump_.capture.event_msg->mutable_stream();
|
2015-07-23 11:41:39 -07:00
|
|
|
const size_t data_size =
|
|
|
|
|
sizeof(int16_t) * frame->samples_per_channel_ * frame->num_channels_;
|
2014-03-04 20:58:13 +00:00
|
|
|
msg->set_output_data(frame->data_, data_size);
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(WriteMessageToDebugFile(debug_dump_.debug_file.get(),
|
|
|
|
|
&crit_debug_, &debug_dump_.capture));
|
2014-03-04 20:58:13 +00:00
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
return kNoError;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::ProcessStreamLocked() {
|
|
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
audioproc::Stream* msg = debug_dump_.capture.event_msg->mutable_stream();
|
|
|
|
|
msg->set_delay(capture_nonlocked_.stream_delay_ms);
|
|
|
|
|
msg->set_drift(
|
|
|
|
|
public_submodules_->echo_cancellation->stream_drift_samples());
|
2015-02-06 19:44:21 +00:00
|
|
|
msg->set_level(gain_control()->stream_analog_level());
|
2015-11-28 12:35:15 -08:00
|
|
|
msg->set_keypress(capture_.key_pressed);
|
2014-03-04 20:58:13 +00:00
|
|
|
}
|
|
|
|
|
#endif
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-06-29 14:57:29 +02:00
|
|
|
MaybeUpdateHistograms();
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
AudioBuffer* ca = capture_.capture_audio.get(); // For brevity.
|
2015-08-14 10:35:55 -07:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.use_new_agc &&
|
|
|
|
|
public_submodules_->gain_control->is_enabled()) {
|
|
|
|
|
private_submodules_->agc_manager->AnalyzePreProcess(
|
|
|
|
|
ca->channels()[0], ca->num_channels(),
|
|
|
|
|
capture_nonlocked_.fwd_proc_format.num_frames());
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
|
|
|
|
|
2012-04-24 18:38:03 +00:00
|
|
|
bool data_processed = is_data_processed();
|
|
|
|
|
if (analysis_needed(data_processed)) {
|
2014-11-14 22:18:10 +00:00
|
|
|
ca->SplitIntoFrequencyBands();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.intelligibility_enabled) {
|
|
|
|
|
public_submodules_->intelligibility_enhancer->AnalyzeCaptureAudio(
|
|
|
|
|
ca->split_channels_f(kBand0To8kHz), capture_nonlocked_.split_rate,
|
|
|
|
|
ca->num_channels());
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.beamformer_enabled) {
|
|
|
|
|
private_submodules_->beamformer->ProcessChunk(*ca->split_data_f(),
|
|
|
|
|
ca->split_data_f());
|
2014-12-19 19:57:34 +00:00
|
|
|
ca->set_num_channels(1);
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(public_submodules_->high_pass_filter->ProcessCaptureAudio(ca));
|
|
|
|
|
RETURN_ON_ERR(public_submodules_->gain_control->AnalyzeCaptureAudio(ca));
|
|
|
|
|
RETURN_ON_ERR(public_submodules_->noise_suppression->AnalyzeCaptureAudio(ca));
|
|
|
|
|
RETURN_ON_ERR(public_submodules_->echo_cancellation->ProcessCaptureAudio(ca));
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (public_submodules_->echo_control_mobile->is_enabled() &&
|
|
|
|
|
public_submodules_->noise_suppression->is_enabled()) {
|
2014-04-24 18:28:56 +00:00
|
|
|
ca->CopyLowPassToReference();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(public_submodules_->noise_suppression->ProcessCaptureAudio(ca));
|
|
|
|
|
RETURN_ON_ERR(
|
|
|
|
|
public_submodules_->echo_control_mobile->ProcessCaptureAudio(ca));
|
|
|
|
|
RETURN_ON_ERR(public_submodules_->voice_detection->ProcessCaptureAudio(ca));
|
2014-12-15 09:41:24 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.use_new_agc &&
|
|
|
|
|
public_submodules_->gain_control->is_enabled() &&
|
|
|
|
|
(!constants_.beamformer_enabled ||
|
|
|
|
|
private_submodules_->beamformer->is_target_present())) {
|
|
|
|
|
private_submodules_->agc_manager->Process(
|
|
|
|
|
ca->split_bands_const(0)[kBand0To8kHz], ca->num_frames_per_band(),
|
|
|
|
|
capture_nonlocked_.split_rate);
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(public_submodules_->gain_control->ProcessCaptureAudio(ca));
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2012-04-24 18:38:03 +00:00
|
|
|
if (synthesis_needed(data_processed)) {
|
2014-11-14 22:18:10 +00:00
|
|
|
ca->MergeFrequencyBands();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2014-12-15 09:41:24 +00:00
|
|
|
// TODO(aluebs): Investigate if the transient suppression placement should be
|
|
|
|
|
// before or after the AGC.
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.transient_suppressor_enabled) {
|
2014-12-15 09:41:24 +00:00
|
|
|
float voice_probability =
|
2015-11-28 12:35:15 -08:00
|
|
|
private_submodules_->agc_manager.get()
|
|
|
|
|
? private_submodules_->agc_manager->voice_probability()
|
|
|
|
|
: 1.f;
|
2014-12-15 09:41:24 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
public_submodules_->transient_suppressor->Suppress(
|
2015-07-23 11:41:39 -07:00
|
|
|
ca->channels_f()[0], ca->num_frames(), ca->num_channels(),
|
|
|
|
|
ca->split_bands_const_f(0)[kBand0To8kHz], ca->num_frames_per_band(),
|
|
|
|
|
ca->keyboard_data(), ca->num_keyboard_frames(), voice_probability,
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.key_pressed);
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
|
|
|
|
|
2011-11-15 16:57:56 +00:00
|
|
|
// The level estimator operates on the recombined data.
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(public_submodules_->level_estimator->ProcessStream(ca));
|
2014-03-04 20:58:13 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.was_stream_delay_set = false;
|
2014-03-04 20:58:13 +00:00
|
|
|
return kNoError;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::AnalyzeReverseStream(const float* const* data,
|
Update a ton of audio code to use size_t more correctly and in general reduce
use of int16_t/uint16_t.
This is the upshot of a recommendation by henrik.lundin and kwiberg on an original small change ( https://webrtc-codereview.appspot.com/42569004/#ps1 ) to stop using int16_t just because values could fit in it, and is similar in nature to a previous "mass change to use size_t more" ( https://webrtc-codereview.appspot.com/23129004/ ) which also needed to be split up for review but to land all at once, since, like adding "const", such changes tend to cause a lot of transitive effects.
This was be reviewed and approved in pieces:
https://codereview.webrtc.org/1224093003
https://codereview.webrtc.org/1224123002
https://codereview.webrtc.org/1224163002
https://codereview.webrtc.org/1225133003
https://codereview.webrtc.org/1225173002
https://codereview.webrtc.org/1227163003
https://codereview.webrtc.org/1227203003
https://codereview.webrtc.org/1227213002
https://codereview.webrtc.org/1227893002
https://codereview.webrtc.org/1228793004
https://codereview.webrtc.org/1228803003
https://codereview.webrtc.org/1228823002
https://codereview.webrtc.org/1228823003
https://codereview.webrtc.org/1228843002
https://codereview.webrtc.org/1230693002
https://codereview.webrtc.org/1231713002
The change is being landed as TBR to all the folks who reviewed the above.
BUG=chromium:81439
TEST=none
R=andrew@webrtc.org, pbos@webrtc.org
TBR=aluebs, andrew, asapersson, henrika, hlundin, jan.skoglund, kwiberg, minyue, pbos, pthatcher
Review URL: https://codereview.webrtc.org/1230503003 .
Cr-Commit-Position: refs/heads/master@{#9768}
2015-08-24 14:52:23 -07:00
|
|
|
size_t samples_per_channel,
|
2015-08-14 10:35:55 -07:00
|
|
|
int rev_sample_rate_hz,
|
2014-03-04 20:58:13 +00:00
|
|
|
ChannelLayout layout) {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_render_);
|
2015-07-23 11:41:39 -07:00
|
|
|
const StreamConfig reverse_config = {
|
2015-08-14 10:35:55 -07:00
|
|
|
rev_sample_rate_hz, ChannelsFromLayout(layout), LayoutHasKeyboard(layout),
|
2015-07-23 11:41:39 -07:00
|
|
|
};
|
|
|
|
|
if (samples_per_channel != reverse_config.num_frames()) {
|
|
|
|
|
return kBadDataLengthError;
|
|
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
return AnalyzeReverseStreamLocked(data, reverse_config, reverse_config);
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::ProcessReverseStream(
|
|
|
|
|
const float* const* src,
|
|
|
|
|
const StreamConfig& reverse_input_config,
|
|
|
|
|
const StreamConfig& reverse_output_config,
|
|
|
|
|
float* const* dest) {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_render_);
|
|
|
|
|
RETURN_ON_ERR(AnalyzeReverseStreamLocked(src, reverse_input_config,
|
|
|
|
|
reverse_output_config));
|
2015-08-14 10:35:55 -07:00
|
|
|
if (is_rev_processed()) {
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_audio->CopyTo(formats_.api_format.reverse_output_stream(),
|
|
|
|
|
dest);
|
2015-11-27 02:47:28 -08:00
|
|
|
} else if (render_check_rev_conversion_needed()) {
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_converter->Convert(src, reverse_input_config.num_samples(),
|
|
|
|
|
dest,
|
|
|
|
|
reverse_output_config.num_samples());
|
2015-08-14 10:35:55 -07:00
|
|
|
} else {
|
|
|
|
|
CopyAudioIfNeeded(src, reverse_input_config.num_frames(),
|
|
|
|
|
reverse_input_config.num_channels(), dest);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return kNoError;
|
2015-07-23 11:41:39 -07:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
int AudioProcessingImpl::AnalyzeReverseStreamLocked(
|
2015-08-14 10:35:55 -07:00
|
|
|
const float* const* src,
|
|
|
|
|
const StreamConfig& reverse_input_config,
|
|
|
|
|
const StreamConfig& reverse_output_config) {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (src == nullptr) {
|
2014-03-04 20:58:13 +00:00
|
|
|
return kNullPointerError;
|
|
|
|
|
}
|
2011-11-15 16:57:56 +00:00
|
|
|
|
2015-08-14 10:35:55 -07:00
|
|
|
if (reverse_input_config.num_channels() <= 0) {
|
2015-07-23 11:41:39 -07:00
|
|
|
return kBadNumberChannelsError;
|
2014-03-04 20:58:13 +00:00
|
|
|
}
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
ProcessingConfig processing_config = formats_.api_format;
|
2015-08-14 10:35:55 -07:00
|
|
|
processing_config.reverse_input_stream() = reverse_input_config;
|
|
|
|
|
processing_config.reverse_output_stream() = reverse_output_config;
|
2015-07-23 11:41:39 -07:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(MaybeInitializeRender(processing_config));
|
2015-08-14 10:35:55 -07:00
|
|
|
assert(reverse_input_config.num_frames() ==
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.reverse_input_stream().num_frames());
|
2015-07-23 11:41:39 -07:00
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
debug_dump_.render.event_msg->set_type(audioproc::Event::REVERSE_STREAM);
|
|
|
|
|
audioproc::ReverseStream* msg =
|
|
|
|
|
debug_dump_.render.event_msg->mutable_reverse_stream();
|
2014-08-28 10:43:09 +00:00
|
|
|
const size_t channel_size =
|
2015-11-28 12:35:15 -08:00
|
|
|
sizeof(float) * formats_.api_format.reverse_input_stream().num_frames();
|
2015-11-17 02:16:45 -08:00
|
|
|
for (int i = 0;
|
2015-11-28 12:35:15 -08:00
|
|
|
i < formats_.api_format.reverse_input_stream().num_channels(); ++i)
|
2015-08-14 10:35:55 -07:00
|
|
|
msg->add_channel(src[i], channel_size);
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(WriteMessageToDebugFile(debug_dump_.debug_file.get(),
|
|
|
|
|
&crit_debug_, &debug_dump_.render));
|
2011-08-03 21:08:51 +00:00
|
|
|
}
|
2011-12-03 00:03:31 +00:00
|
|
|
#endif
|
2011-08-03 21:08:51 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_audio->CopyFrom(src,
|
|
|
|
|
formats_.api_format.reverse_input_stream());
|
2015-08-14 10:35:55 -07:00
|
|
|
return ProcessReverseStreamLocked();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::ProcessReverseStream(AudioFrame* frame) {
|
|
|
|
|
RETURN_ON_ERR(AnalyzeReverseStream(frame));
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_render_);
|
2015-08-14 10:35:55 -07:00
|
|
|
if (is_rev_processed()) {
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_audio->InterleaveTo(frame, true);
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return kNoError;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::AnalyzeReverseStream(AudioFrame* frame) {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_render_);
|
|
|
|
|
if (frame == nullptr) {
|
2011-07-07 08:21:25 +00:00
|
|
|
return kNullPointerError;
|
|
|
|
|
}
|
2014-04-22 21:00:04 +00:00
|
|
|
// Must be a native rate.
|
|
|
|
|
if (frame->sample_rate_hz_ != kSampleRate8kHz &&
|
|
|
|
|
frame->sample_rate_hz_ != kSampleRate16kHz &&
|
2014-11-17 23:01:23 +00:00
|
|
|
frame->sample_rate_hz_ != kSampleRate32kHz &&
|
|
|
|
|
frame->sample_rate_hz_ != kSampleRate48kHz) {
|
2014-04-22 21:00:04 +00:00
|
|
|
return kBadSampleRateError;
|
|
|
|
|
}
|
|
|
|
|
// This interface does not tolerate different forward and reverse rates.
|
2015-11-17 02:16:45 -08:00
|
|
|
if (frame->sample_rate_hz_ !=
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.input_stream().sample_rate_hz()) {
|
2011-07-07 08:21:25 +00:00
|
|
|
return kBadSampleRateError;
|
|
|
|
|
}
|
2014-03-10 22:26:12 +00:00
|
|
|
|
2015-07-23 11:41:39 -07:00
|
|
|
if (frame->num_channels_ <= 0) {
|
|
|
|
|
return kBadNumberChannelsError;
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
ProcessingConfig processing_config = formats_.api_format;
|
2015-08-14 10:35:55 -07:00
|
|
|
processing_config.reverse_input_stream().set_sample_rate_hz(
|
|
|
|
|
frame->sample_rate_hz_);
|
|
|
|
|
processing_config.reverse_input_stream().set_num_channels(
|
|
|
|
|
frame->num_channels_);
|
|
|
|
|
processing_config.reverse_output_stream().set_sample_rate_hz(
|
|
|
|
|
frame->sample_rate_hz_);
|
|
|
|
|
processing_config.reverse_output_stream().set_num_channels(
|
|
|
|
|
frame->num_channels_);
|
2015-07-23 11:41:39 -07:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(MaybeInitializeRender(processing_config));
|
2015-07-23 11:41:39 -07:00
|
|
|
if (frame->samples_per_channel_ !=
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.reverse_input_stream().num_frames()) {
|
2014-03-04 20:58:13 +00:00
|
|
|
return kBadDataLengthError;
|
|
|
|
|
}
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
debug_dump_.render.event_msg->set_type(audioproc::Event::REVERSE_STREAM);
|
|
|
|
|
audioproc::ReverseStream* msg =
|
|
|
|
|
debug_dump_.render.event_msg->mutable_reverse_stream();
|
2015-07-23 11:41:39 -07:00
|
|
|
const size_t data_size =
|
|
|
|
|
sizeof(int16_t) * frame->samples_per_channel_ * frame->num_channels_;
|
2012-05-02 23:56:37 +00:00
|
|
|
msg->set_data(frame->data_, data_size);
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(WriteMessageToDebugFile(debug_dump_.debug_file.get(),
|
|
|
|
|
&crit_debug_, &debug_dump_.render));
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
2011-12-03 00:03:31 +00:00
|
|
|
#endif
|
2015-11-28 12:35:15 -08:00
|
|
|
render_.render_audio->DeinterleaveFrom(frame);
|
2015-08-14 10:35:55 -07:00
|
|
|
return ProcessReverseStreamLocked();
|
2014-03-04 20:58:13 +00:00
|
|
|
}
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-08-14 10:35:55 -07:00
|
|
|
int AudioProcessingImpl::ProcessReverseStreamLocked() {
|
2015-11-28 12:35:15 -08:00
|
|
|
AudioBuffer* ra = render_.render_audio.get(); // For brevity.
|
|
|
|
|
if (formats_.rev_proc_format.sample_rate_hz() == kSampleRate32kHz) {
|
2014-11-14 22:18:10 +00:00
|
|
|
ra->SplitIntoFrequencyBands();
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.intelligibility_enabled) {
|
|
|
|
|
// Currently run in single-threaded mode when the intelligibility
|
|
|
|
|
// enhancer is activated.
|
|
|
|
|
// TODO(peah): Fix to be properly multi-threaded.
|
|
|
|
|
rtc::CritScope cs(&crit_capture_);
|
|
|
|
|
public_submodules_->intelligibility_enhancer->ProcessRenderAudio(
|
|
|
|
|
ra->split_channels_f(kBand0To8kHz), capture_nonlocked_.split_rate,
|
|
|
|
|
ra->num_channels());
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(public_submodules_->echo_cancellation->ProcessRenderAudio(ra));
|
|
|
|
|
RETURN_ON_ERR(
|
|
|
|
|
public_submodules_->echo_control_mobile->ProcessRenderAudio(ra));
|
|
|
|
|
if (!constants_.use_new_agc) {
|
|
|
|
|
RETURN_ON_ERR(public_submodules_->gain_control->ProcessRenderAudio(ra));
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (formats_.rev_proc_format.sample_rate_hz() == kSampleRate32kHz &&
|
2015-08-14 10:35:55 -07:00
|
|
|
is_rev_processed()) {
|
|
|
|
|
ra->MergeFrequencyBands();
|
|
|
|
|
}
|
|
|
|
|
|
2014-03-04 20:58:13 +00:00
|
|
|
return kNoError;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::set_stream_delay_ms(int delay) {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_capture_);
|
2012-05-29 21:14:06 +00:00
|
|
|
Error retval = kNoError;
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.was_stream_delay_set = true;
|
|
|
|
|
delay += capture_.delay_offset_ms;
|
2012-03-06 19:03:39 +00:00
|
|
|
|
2011-07-07 08:21:25 +00:00
|
|
|
if (delay < 0) {
|
2012-05-29 21:14:06 +00:00
|
|
|
delay = 0;
|
|
|
|
|
retval = kBadStreamParameterWarning;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// TODO(ajm): the max is rather arbitrarily chosen; investigate.
|
|
|
|
|
if (delay > 500) {
|
2012-05-29 21:14:06 +00:00
|
|
|
delay = 500;
|
|
|
|
|
retval = kBadStreamParameterWarning;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_nonlocked_.stream_delay_ms = delay;
|
2012-05-29 21:14:06 +00:00
|
|
|
return retval;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::stream_delay_ms() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Used as callback from submodules, hence locking is not allowed.
|
|
|
|
|
return capture_nonlocked_.stream_delay_ms;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bool AudioProcessingImpl::was_stream_delay_set() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Used as callback from submodules, hence locking is not allowed.
|
|
|
|
|
return capture_.was_stream_delay_set;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2014-03-04 20:58:13 +00:00
|
|
|
void AudioProcessingImpl::set_stream_key_pressed(bool key_pressed) {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_capture_);
|
|
|
|
|
capture_.key_pressed = key_pressed;
|
2014-03-04 20:58:13 +00:00
|
|
|
}
|
|
|
|
|
|
2012-03-06 19:03:39 +00:00
|
|
|
void AudioProcessingImpl::set_delay_offset_ms(int offset) {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_capture_);
|
|
|
|
|
capture_.delay_offset_ms = offset;
|
2012-03-06 19:03:39 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::delay_offset_ms() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
rtc::CritScope cs(&crit_capture_);
|
|
|
|
|
return capture_.delay_offset_ms;
|
2012-03-06 19:03:39 +00:00
|
|
|
}
|
|
|
|
|
|
2011-07-07 08:21:25 +00:00
|
|
|
int AudioProcessingImpl::StartDebugRecording(
|
|
|
|
|
const char filename[AudioProcessing::kMaxFilenameSize]) {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2015-05-20 11:11:07 +02:00
|
|
|
static_assert(kMaxFilenameSize == FileWrapper::kMaxFileNameSize, "");
|
2011-07-07 08:21:25 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (filename == nullptr) {
|
2011-07-07 08:21:25 +00:00
|
|
|
return kNullPointerError;
|
|
|
|
|
}
|
|
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2011-07-07 08:21:25 +00:00
|
|
|
// Stop any ongoing recording.
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
if (debug_dump_.debug_file->CloseFile() == -1) {
|
2011-07-07 08:21:25 +00:00
|
|
|
return kFileError;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->OpenFile(filename, false) == -1) {
|
|
|
|
|
debug_dump_.debug_file->CloseFile();
|
2011-07-07 08:21:25 +00:00
|
|
|
return kFileError;
|
|
|
|
|
}
|
|
|
|
|
|
2015-10-03 00:39:14 +02:00
|
|
|
RETURN_ON_ERR(WriteConfigMessage(true));
|
|
|
|
|
RETURN_ON_ERR(WriteInitMessage());
|
2013-12-06 16:05:17 +00:00
|
|
|
return kNoError;
|
|
|
|
|
#else
|
|
|
|
|
return kUnsupportedFunctionError;
|
|
|
|
|
#endif // WEBRTC_AUDIOPROC_DEBUG_DUMP
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::StartDebugRecording(FILE* handle) {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2013-12-06 16:05:17 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (handle == nullptr) {
|
2013-12-06 16:05:17 +00:00
|
|
|
return kNullPointerError;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
|
|
|
|
// Stop any ongoing recording.
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
if (debug_dump_.debug_file->CloseFile() == -1) {
|
2013-12-06 16:05:17 +00:00
|
|
|
return kFileError;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->OpenFromFileHandle(handle, true, false) == -1) {
|
2013-12-06 16:05:17 +00:00
|
|
|
return kFileError;
|
|
|
|
|
}
|
|
|
|
|
|
2015-10-03 00:39:14 +02:00
|
|
|
RETURN_ON_ERR(WriteConfigMessage(true));
|
|
|
|
|
RETURN_ON_ERR(WriteInitMessage());
|
2011-07-07 08:21:25 +00:00
|
|
|
return kNoError;
|
2011-12-03 00:03:31 +00:00
|
|
|
#else
|
|
|
|
|
return kUnsupportedFunctionError;
|
|
|
|
|
#endif // WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2014-10-10 08:36:56 +00:00
|
|
|
int AudioProcessingImpl::StartDebugRecordingForPlatformFile(
|
|
|
|
|
rtc::PlatformFile handle) {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2014-10-10 08:36:56 +00:00
|
|
|
FILE* stream = rtc::FdopenPlatformFileForWriting(handle);
|
|
|
|
|
return StartDebugRecording(stream);
|
|
|
|
|
}
|
|
|
|
|
|
2011-07-07 08:21:25 +00:00
|
|
|
int AudioProcessingImpl::StopDebugRecording() {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
2011-12-03 00:03:31 +00:00
|
|
|
|
|
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2011-07-07 08:21:25 +00:00
|
|
|
// We just return if recording hasn't started.
|
2015-11-28 12:35:15 -08:00
|
|
|
if (debug_dump_.debug_file->Open()) {
|
|
|
|
|
if (debug_dump_.debug_file->CloseFile() == -1) {
|
2011-07-07 08:21:25 +00:00
|
|
|
return kFileError;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
return kNoError;
|
2011-12-03 00:03:31 +00:00
|
|
|
#else
|
|
|
|
|
return kUnsupportedFunctionError;
|
|
|
|
|
#endif // WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
EchoCancellation* AudioProcessingImpl::echo_cancellation() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Adding a lock here has no effect as it allows any access to the submodule
|
|
|
|
|
// from the returned pointer.
|
|
|
|
|
return public_submodules_->echo_cancellation;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
EchoControlMobile* AudioProcessingImpl::echo_control_mobile() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Adding a lock here has no effect as it allows any access to the submodule
|
|
|
|
|
// from the returned pointer.
|
|
|
|
|
return public_submodules_->echo_control_mobile;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
GainControl* AudioProcessingImpl::gain_control() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Adding a lock here has no effect as it allows any access to the submodule
|
|
|
|
|
// from the returned pointer.
|
|
|
|
|
if (constants_.use_new_agc) {
|
|
|
|
|
return public_submodules_->gain_control_for_new_agc.get();
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
return public_submodules_->gain_control;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
HighPassFilter* AudioProcessingImpl::high_pass_filter() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Adding a lock here has no effect as it allows any access to the submodule
|
|
|
|
|
// from the returned pointer.
|
|
|
|
|
return public_submodules_->high_pass_filter;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
LevelEstimator* AudioProcessingImpl::level_estimator() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Adding a lock here has no effect as it allows any access to the submodule
|
|
|
|
|
// from the returned pointer.
|
|
|
|
|
return public_submodules_->level_estimator;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
NoiseSuppression* AudioProcessingImpl::noise_suppression() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Adding a lock here has no effect as it allows any access to the submodule
|
|
|
|
|
// from the returned pointer.
|
|
|
|
|
return public_submodules_->noise_suppression;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
VoiceDetection* AudioProcessingImpl::voice_detection() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Adding a lock here has no effect as it allows any access to the submodule
|
|
|
|
|
// from the returned pointer.
|
|
|
|
|
return public_submodules_->voice_detection;
|
2011-07-07 08:21:25 +00:00
|
|
|
}
|
|
|
|
|
|
2012-04-24 18:38:03 +00:00
|
|
|
bool AudioProcessingImpl::is_data_processed() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.beamformer_enabled) {
|
2014-12-19 19:57:34 +00:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
int enabled_count = 0;
|
2015-11-28 12:35:15 -08:00
|
|
|
for (auto item : private_submodules_->component_list) {
|
2015-03-12 23:23:38 +00:00
|
|
|
if (item->is_component_enabled()) {
|
2011-12-03 00:03:31 +00:00
|
|
|
enabled_count++;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
// Data is unchanged if no components are enabled, or if only
|
|
|
|
|
// public_submodules_->level_estimator
|
|
|
|
|
// or public_submodules_->voice_detection is enabled.
|
2011-12-03 00:03:31 +00:00
|
|
|
if (enabled_count == 0) {
|
|
|
|
|
return false;
|
|
|
|
|
} else if (enabled_count == 1) {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (public_submodules_->level_estimator->is_enabled() ||
|
|
|
|
|
public_submodules_->voice_detection->is_enabled()) {
|
2011-12-03 00:03:31 +00:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
} else if (enabled_count == 2) {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (public_submodules_->level_estimator->is_enabled() &&
|
|
|
|
|
public_submodules_->voice_detection->is_enabled()) {
|
2011-12-03 00:03:31 +00:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2014-03-04 20:58:13 +00:00
|
|
|
bool AudioProcessingImpl::output_copy_needed(bool is_data_processed) const {
|
2012-04-24 18:38:03 +00:00
|
|
|
// Check if we've upmixed or downmixed the audio.
|
2015-11-28 12:35:15 -08:00
|
|
|
return ((formats_.api_format.output_stream().num_channels() !=
|
|
|
|
|
formats_.api_format.input_stream().num_channels()) ||
|
|
|
|
|
is_data_processed || capture_.transient_suppressor_enabled);
|
2011-12-03 00:03:31 +00:00
|
|
|
}
|
|
|
|
|
|
2012-04-24 18:38:03 +00:00
|
|
|
bool AudioProcessingImpl::synthesis_needed(bool is_data_processed) const {
|
2015-07-23 11:41:39 -07:00
|
|
|
return (is_data_processed &&
|
2015-11-28 12:35:15 -08:00
|
|
|
(capture_nonlocked_.fwd_proc_format.sample_rate_hz() ==
|
|
|
|
|
kSampleRate32kHz ||
|
|
|
|
|
capture_nonlocked_.fwd_proc_format.sample_rate_hz() ==
|
|
|
|
|
kSampleRate48kHz));
|
2012-04-24 18:38:03 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bool AudioProcessingImpl::analysis_needed(bool is_data_processed) const {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (!is_data_processed &&
|
|
|
|
|
!public_submodules_->voice_detection->is_enabled() &&
|
|
|
|
|
!capture_.transient_suppressor_enabled) {
|
|
|
|
|
// Only public_submodules_->level_estimator is enabled.
|
2011-12-03 00:03:31 +00:00
|
|
|
return false;
|
2015-11-28 12:35:15 -08:00
|
|
|
} else if (capture_nonlocked_.fwd_proc_format.sample_rate_hz() ==
|
|
|
|
|
kSampleRate32kHz ||
|
|
|
|
|
capture_nonlocked_.fwd_proc_format.sample_rate_hz() ==
|
|
|
|
|
kSampleRate48kHz) {
|
|
|
|
|
// Something besides public_submodules_->level_estimator is enabled, and we
|
|
|
|
|
// have super-wb.
|
2011-12-03 00:03:31 +00:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2015-08-14 10:35:55 -07:00
|
|
|
bool AudioProcessingImpl::is_rev_processed() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
return constants_.intelligibility_enabled &&
|
|
|
|
|
public_submodules_->intelligibility_enhancer->active();
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
|
|
|
|
|
2015-11-27 02:47:28 -08:00
|
|
|
bool AudioProcessingImpl::render_check_rev_conversion_needed() const {
|
|
|
|
|
return rev_conversion_needed();
|
|
|
|
|
}
|
|
|
|
|
|
2015-08-14 10:35:55 -07:00
|
|
|
bool AudioProcessingImpl::rev_conversion_needed() const {
|
2015-11-28 12:35:15 -08:00
|
|
|
return (formats_.api_format.reverse_input_stream() !=
|
|
|
|
|
formats_.api_format.reverse_output_stream());
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
|
|
|
|
|
2015-04-15 11:42:40 +02:00
|
|
|
void AudioProcessingImpl::InitializeExperimentalAgc() {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.use_new_agc) {
|
|
|
|
|
if (!private_submodules_->agc_manager.get()) {
|
|
|
|
|
private_submodules_->agc_manager.reset(new AgcManagerDirect(
|
|
|
|
|
public_submodules_->gain_control,
|
|
|
|
|
public_submodules_->gain_control_for_new_agc.get(),
|
|
|
|
|
constants_.agc_startup_min_volume));
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
private_submodules_->agc_manager->Initialize();
|
|
|
|
|
private_submodules_->agc_manager->SetCaptureMuted(
|
|
|
|
|
capture_.output_will_be_muted);
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-15 11:42:40 +02:00
|
|
|
void AudioProcessingImpl::InitializeTransient() {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.transient_suppressor_enabled) {
|
|
|
|
|
if (!public_submodules_->transient_suppressor.get()) {
|
|
|
|
|
public_submodules_->transient_suppressor.reset(new TransientSuppressor());
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
public_submodules_->transient_suppressor->Initialize(
|
|
|
|
|
capture_nonlocked_.fwd_proc_format.sample_rate_hz(),
|
|
|
|
|
capture_nonlocked_.split_rate,
|
|
|
|
|
formats_.api_format.output_stream().num_channels());
|
2014-12-15 09:41:24 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2014-12-19 19:57:34 +00:00
|
|
|
void AudioProcessingImpl::InitializeBeamformer() {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.beamformer_enabled) {
|
|
|
|
|
if (!private_submodules_->beamformer) {
|
|
|
|
|
private_submodules_->beamformer.reset(new NonlinearBeamformer(
|
|
|
|
|
constants_.array_geometry, constants_.target_direction));
|
2015-01-15 18:07:21 +00:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
private_submodules_->beamformer->Initialize(kChunkSizeMs,
|
|
|
|
|
capture_nonlocked_.split_rate);
|
2014-12-19 19:57:34 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-08-14 10:35:55 -07:00
|
|
|
void AudioProcessingImpl::InitializeIntelligibility() {
|
2015-11-28 12:35:15 -08:00
|
|
|
if (constants_.intelligibility_enabled) {
|
2015-08-14 10:35:55 -07:00
|
|
|
IntelligibilityEnhancer::Config config;
|
2015-11-28 12:35:15 -08:00
|
|
|
config.sample_rate_hz = capture_nonlocked_.split_rate;
|
|
|
|
|
config.num_capture_channels = capture_.capture_audio->num_channels();
|
|
|
|
|
config.num_render_channels = render_.render_audio->num_channels();
|
|
|
|
|
public_submodules_->intelligibility_enhancer.reset(
|
|
|
|
|
new IntelligibilityEnhancer(config));
|
2015-08-14 10:35:55 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-06-29 14:57:29 +02:00
|
|
|
void AudioProcessingImpl::MaybeUpdateHistograms() {
|
2015-07-05 10:46:01 +02:00
|
|
|
static const int kMinDiffDelayMs = 60;
|
2015-06-29 14:57:29 +02:00
|
|
|
|
|
|
|
|
if (echo_cancellation()->is_enabled()) {
|
2015-07-07 11:50:05 +02:00
|
|
|
// Activate delay_jumps_ counters if we know echo_cancellation is runnning.
|
|
|
|
|
// If a stream has echo we know that the echo_cancellation is in process.
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.stream_delay_jumps == -1 &&
|
|
|
|
|
echo_cancellation()->stream_has_echo()) {
|
|
|
|
|
capture_.stream_delay_jumps = 0;
|
2015-07-07 11:50:05 +02:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.aec_system_delay_jumps == -1 &&
|
2015-07-07 11:50:05 +02:00
|
|
|
echo_cancellation()->stream_has_echo()) {
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.aec_system_delay_jumps = 0;
|
2015-07-07 11:50:05 +02:00
|
|
|
}
|
|
|
|
|
|
2015-06-29 14:57:29 +02:00
|
|
|
// Detect a jump in platform reported system delay and log the difference.
|
2015-11-28 12:35:15 -08:00
|
|
|
const int diff_stream_delay_ms =
|
|
|
|
|
capture_nonlocked_.stream_delay_ms - capture_.last_stream_delay_ms;
|
|
|
|
|
if (diff_stream_delay_ms > kMinDiffDelayMs &&
|
|
|
|
|
capture_.last_stream_delay_ms != 0) {
|
2015-06-29 14:57:29 +02:00
|
|
|
RTC_HISTOGRAM_COUNTS("WebRTC.Audio.PlatformReportedStreamDelayJump",
|
|
|
|
|
diff_stream_delay_ms, kMinDiffDelayMs, 1000, 100);
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.stream_delay_jumps == -1) {
|
|
|
|
|
capture_.stream_delay_jumps = 0; // Activate counter if needed.
|
2015-07-07 11:50:05 +02:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.stream_delay_jumps++;
|
2015-06-29 14:57:29 +02:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.last_stream_delay_ms = capture_nonlocked_.stream_delay_ms;
|
2015-06-29 14:57:29 +02:00
|
|
|
|
|
|
|
|
// Detect a jump in AEC system delay and log the difference.
|
2015-11-28 12:35:15 -08:00
|
|
|
const int frames_per_ms =
|
|
|
|
|
rtc::CheckedDivExact(capture_nonlocked_.split_rate, 1000);
|
2015-06-29 14:57:29 +02:00
|
|
|
const int aec_system_delay_ms =
|
|
|
|
|
WebRtcAec_system_delay(echo_cancellation()->aec_core()) / frames_per_ms;
|
2015-07-23 11:41:39 -07:00
|
|
|
const int diff_aec_system_delay_ms =
|
2015-11-28 12:35:15 -08:00
|
|
|
aec_system_delay_ms - capture_.last_aec_system_delay_ms;
|
2015-06-29 14:57:29 +02:00
|
|
|
if (diff_aec_system_delay_ms > kMinDiffDelayMs &&
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.last_aec_system_delay_ms != 0) {
|
2015-06-29 14:57:29 +02:00
|
|
|
RTC_HISTOGRAM_COUNTS("WebRTC.Audio.AecSystemDelayJump",
|
|
|
|
|
diff_aec_system_delay_ms, kMinDiffDelayMs, 1000,
|
|
|
|
|
100);
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.aec_system_delay_jumps == -1) {
|
|
|
|
|
capture_.aec_system_delay_jumps = 0; // Activate counter if needed.
|
2015-07-07 11:50:05 +02:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.aec_system_delay_jumps++;
|
2015-06-29 14:57:29 +02:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.last_aec_system_delay_ms = aec_system_delay_ms;
|
2015-06-29 14:57:29 +02:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-07-07 11:50:05 +02:00
|
|
|
void AudioProcessingImpl::UpdateHistogramsOnCallEnd() {
|
2015-11-28 12:35:15 -08:00
|
|
|
// Run in a single-threaded manner.
|
|
|
|
|
rtc::CritScope cs_render(&crit_render_);
|
|
|
|
|
rtc::CritScope cs_capture(&crit_capture_);
|
|
|
|
|
|
|
|
|
|
if (capture_.stream_delay_jumps > -1) {
|
2015-07-07 11:50:05 +02:00
|
|
|
RTC_HISTOGRAM_ENUMERATION(
|
|
|
|
|
"WebRTC.Audio.NumOfPlatformReportedStreamDelayJumps",
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.stream_delay_jumps, 51);
|
2015-07-07 11:50:05 +02:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.stream_delay_jumps = -1;
|
|
|
|
|
capture_.last_stream_delay_ms = 0;
|
2015-07-07 11:50:05 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (capture_.aec_system_delay_jumps > -1) {
|
2015-07-07 11:50:05 +02:00
|
|
|
RTC_HISTOGRAM_ENUMERATION("WebRTC.Audio.NumOfAecSystemDelayJumps",
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.aec_system_delay_jumps, 51);
|
2015-07-07 11:50:05 +02:00
|
|
|
}
|
2015-11-28 12:35:15 -08:00
|
|
|
capture_.aec_system_delay_jumps = -1;
|
|
|
|
|
capture_.last_aec_system_delay_ms = 0;
|
2015-07-07 11:50:05 +02:00
|
|
|
}
|
|
|
|
|
|
2011-12-03 00:03:31 +00:00
|
|
|
#ifdef WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2015-11-28 12:35:15 -08:00
|
|
|
int AudioProcessingImpl::WriteMessageToDebugFile(
|
|
|
|
|
FileWrapper* debug_file,
|
|
|
|
|
rtc::CriticalSection* crit_debug,
|
|
|
|
|
ApmDebugDumpThreadState* debug_state) {
|
|
|
|
|
int32_t size = debug_state->event_msg->ByteSize();
|
2011-08-03 21:08:51 +00:00
|
|
|
if (size <= 0) {
|
|
|
|
|
return kUnspecifiedError;
|
|
|
|
|
}
|
2013-10-22 10:27:23 +00:00
|
|
|
#if defined(WEBRTC_ARCH_BIG_ENDIAN)
|
2015-07-23 11:41:39 -07:00
|
|
|
// TODO(ajm): Use little-endian "on the wire". For the moment, we can be
|
|
|
|
|
// pretty safe in assuming little-endian.
|
2011-08-03 21:08:51 +00:00
|
|
|
#endif
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
if (!debug_state->event_msg->SerializeToString(&debug_state->event_str)) {
|
2011-08-03 21:08:51 +00:00
|
|
|
return kUnspecifiedError;
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
{
|
|
|
|
|
// Ensure atomic writes of the message.
|
|
|
|
|
rtc::CritScope cs_capture(crit_debug);
|
|
|
|
|
// Write message preceded by its size.
|
|
|
|
|
if (!debug_file->Write(&size, sizeof(int32_t))) {
|
|
|
|
|
return kFileError;
|
|
|
|
|
}
|
|
|
|
|
if (!debug_file->Write(debug_state->event_str.data(),
|
|
|
|
|
debug_state->event_str.length())) {
|
|
|
|
|
return kFileError;
|
|
|
|
|
}
|
2011-08-03 21:08:51 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
debug_state->event_msg->Clear();
|
2011-08-03 21:08:51 +00:00
|
|
|
|
2014-03-04 20:58:13 +00:00
|
|
|
return kNoError;
|
2011-08-03 21:08:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::WriteInitMessage() {
|
2015-11-28 12:35:15 -08:00
|
|
|
debug_dump_.capture.event_msg->set_type(audioproc::Event::INIT);
|
|
|
|
|
audioproc::Init* msg = debug_dump_.capture.event_msg->mutable_init();
|
|
|
|
|
msg->set_sample_rate(formats_.api_format.input_stream().sample_rate_hz());
|
|
|
|
|
|
2015-11-17 02:16:45 -08:00
|
|
|
msg->set_num_input_channels(
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.input_stream().num_channels());
|
2015-11-17 02:16:45 -08:00
|
|
|
msg->set_num_output_channels(
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.output_stream().num_channels());
|
2015-08-14 10:35:55 -07:00
|
|
|
msg->set_num_reverse_channels(
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.reverse_input_stream().num_channels());
|
2015-08-14 10:35:55 -07:00
|
|
|
msg->set_reverse_sample_rate(
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.reverse_input_stream().sample_rate_hz());
|
2015-11-17 02:16:45 -08:00
|
|
|
msg->set_output_sample_rate(
|
2015-11-28 12:35:15 -08:00
|
|
|
formats_.api_format.output_stream().sample_rate_hz());
|
|
|
|
|
// TODO(ekmeyerson): Add reverse output fields to
|
|
|
|
|
// debug_dump_.capture.event_msg.
|
2011-08-03 21:08:51 +00:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(WriteMessageToDebugFile(debug_dump_.debug_file.get(),
|
|
|
|
|
&crit_debug_, &debug_dump_.capture));
|
2015-10-03 00:39:14 +02:00
|
|
|
return kNoError;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AudioProcessingImpl::WriteConfigMessage(bool forced) {
|
|
|
|
|
audioproc::Config config;
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
config.set_aec_enabled(public_submodules_->echo_cancellation->is_enabled());
|
2015-10-03 00:39:14 +02:00
|
|
|
config.set_aec_delay_agnostic_enabled(
|
2015-11-28 12:35:15 -08:00
|
|
|
public_submodules_->echo_cancellation->is_delay_agnostic_enabled());
|
2015-10-03 00:39:14 +02:00
|
|
|
config.set_aec_drift_compensation_enabled(
|
2015-11-28 12:35:15 -08:00
|
|
|
public_submodules_->echo_cancellation->is_drift_compensation_enabled());
|
2015-10-03 00:39:14 +02:00
|
|
|
config.set_aec_extended_filter_enabled(
|
2015-11-28 12:35:15 -08:00
|
|
|
public_submodules_->echo_cancellation->is_extended_filter_enabled());
|
|
|
|
|
config.set_aec_suppression_level(static_cast<int>(
|
|
|
|
|
public_submodules_->echo_cancellation->suppression_level()));
|
2015-10-03 00:39:14 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
config.set_aecm_enabled(
|
|
|
|
|
public_submodules_->echo_control_mobile->is_enabled());
|
2015-10-03 00:39:14 +02:00
|
|
|
config.set_aecm_comfort_noise_enabled(
|
2015-11-28 12:35:15 -08:00
|
|
|
public_submodules_->echo_control_mobile->is_comfort_noise_enabled());
|
|
|
|
|
config.set_aecm_routing_mode(static_cast<int>(
|
|
|
|
|
public_submodules_->echo_control_mobile->routing_mode()));
|
2015-10-03 00:39:14 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
config.set_agc_enabled(public_submodules_->gain_control->is_enabled());
|
|
|
|
|
config.set_agc_mode(
|
|
|
|
|
static_cast<int>(public_submodules_->gain_control->mode()));
|
|
|
|
|
config.set_agc_limiter_enabled(
|
|
|
|
|
public_submodules_->gain_control->is_limiter_enabled());
|
|
|
|
|
config.set_noise_robust_agc_enabled(constants_.use_new_agc);
|
2015-10-03 00:39:14 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
config.set_hpf_enabled(public_submodules_->high_pass_filter->is_enabled());
|
2015-10-03 00:39:14 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
config.set_ns_enabled(public_submodules_->noise_suppression->is_enabled());
|
|
|
|
|
config.set_ns_level(
|
|
|
|
|
static_cast<int>(public_submodules_->noise_suppression->level()));
|
2015-10-03 00:39:14 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
config.set_transient_suppression_enabled(
|
|
|
|
|
capture_.transient_suppressor_enabled);
|
2015-10-03 00:39:14 +02:00
|
|
|
|
|
|
|
|
std::string serialized_config = config.SerializeAsString();
|
2015-11-28 12:35:15 -08:00
|
|
|
if (!forced &&
|
|
|
|
|
debug_dump_.capture.last_serialized_config == serialized_config) {
|
2015-10-03 00:39:14 +02:00
|
|
|
return kNoError;
|
2011-08-03 21:08:51 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
debug_dump_.capture.last_serialized_config = serialized_config;
|
2015-10-03 00:39:14 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
debug_dump_.capture.event_msg->set_type(audioproc::Event::CONFIG);
|
|
|
|
|
debug_dump_.capture.event_msg->mutable_config()->CopyFrom(config);
|
2015-10-03 00:39:14 +02:00
|
|
|
|
2015-11-28 12:35:15 -08:00
|
|
|
RETURN_ON_ERR(WriteMessageToDebugFile(debug_dump_.debug_file.get(),
|
|
|
|
|
&crit_debug_, &debug_dump_.capture));
|
2011-08-03 21:08:51 +00:00
|
|
|
return kNoError;
|
|
|
|
|
}
|
2011-12-03 00:03:31 +00:00
|
|
|
#endif // WEBRTC_AUDIOPROC_DEBUG_DUMP
|
2014-04-22 21:00:04 +00:00
|
|
|
|
2011-07-07 08:21:25 +00:00
|
|
|
} // namespace webrtc
|