1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342
|
/*
* Copyright (c) 2014 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 "modules/audio_processing/audio_processing_impl.h"
#include <memory>
#include "absl/memory/memory.h"
#include "modules/audio_processing/include/audio_processing.h"
#include "modules/audio_processing/test/echo_control_mock.h"
#include "modules/audio_processing/test/test_utils.h"
#include "rtc_base/checks.h"
#include "rtc_base/ref_counted_object.h"
#include "rtc_base/scoped_ref_ptr.h"
#include "test/gmock.h"
#include "test/gtest.h"
namespace webrtc {
namespace {
using ::testing::Invoke;
using ::testing::NotNull;
class MockInitialize : public AudioProcessingImpl {
public:
explicit MockInitialize(const webrtc::Config& config)
: AudioProcessingImpl(config) {}
MOCK_METHOD0(InitializeLocked, int());
int RealInitializeLocked() RTC_NO_THREAD_SAFETY_ANALYSIS {
return AudioProcessingImpl::InitializeLocked();
}
MOCK_CONST_METHOD0(AddRef, void());
MOCK_CONST_METHOD0(Release, rtc::RefCountReleaseStatus());
};
// Creates MockEchoControl instances and provides a raw pointer access to
// the next created one. The raw pointer is meant to be used with gmock.
// Returning a pointer of the next created MockEchoControl instance is necessary
// for the following reasons: (i) gmock expectations must be set before any call
// occurs, (ii) APM is initialized the first time that
// AudioProcessingImpl::ProcessStream() is called and the initialization leads
// to the creation of a new EchoControl object.
class MockEchoControlFactory : public EchoControlFactory {
public:
MockEchoControlFactory() : next_mock_(absl::make_unique<MockEchoControl>()) {}
// Returns a pointer to the next MockEchoControl that this factory creates.
MockEchoControl* GetNext() const { return next_mock_.get(); }
std::unique_ptr<EchoControl> Create(int sample_rate_hz) override {
std::unique_ptr<EchoControl> mock = std::move(next_mock_);
next_mock_ = absl::make_unique<MockEchoControl>();
return mock;
}
private:
std::unique_ptr<MockEchoControl> next_mock_;
};
void InitializeAudioFrame(size_t input_rate,
size_t num_channels,
AudioFrame* frame) {
const size_t samples_per_input_channel = rtc::CheckedDivExact(
input_rate, static_cast<size_t>(rtc::CheckedDivExact(
1000, AudioProcessing::kChunkSizeMs)));
RTC_DCHECK_LE(samples_per_input_channel * num_channels,
AudioFrame::kMaxDataSizeSamples);
frame->samples_per_channel_ = samples_per_input_channel;
frame->sample_rate_hz_ = input_rate;
frame->num_channels_ = num_channels;
}
void FillFixedFrame(int16_t audio_level, AudioFrame* frame) {
const size_t num_samples = frame->samples_per_channel_ * frame->num_channels_;
for (size_t i = 0; i < num_samples; ++i) {
frame->mutable_data()[i] = audio_level;
}
}
// Mocks EchoDetector and records the first samples of the last analyzed render
// stream frame. Used to check what data is read by an EchoDetector
// implementation injected into an APM.
class TestEchoDetector : public EchoDetector {
public:
TestEchoDetector()
: analyze_render_audio_called_(false),
last_render_audio_first_sample_(0.f) {}
~TestEchoDetector() override = default;
void AnalyzeRenderAudio(rtc::ArrayView<const float> render_audio) override {
last_render_audio_first_sample_ = render_audio[0];
analyze_render_audio_called_ = true;
}
void AnalyzeCaptureAudio(rtc::ArrayView<const float> capture_audio) override {
}
void Initialize(int capture_sample_rate_hz,
int num_capture_channels,
int render_sample_rate_hz,
int num_render_channels) override {}
EchoDetector::Metrics GetMetrics() const override { return {}; }
// Returns true if AnalyzeRenderAudio() has been called at least once.
bool analyze_render_audio_called() const {
return analyze_render_audio_called_;
}
// Returns the first sample of the last analyzed render frame.
float last_render_audio_first_sample() const {
return last_render_audio_first_sample_;
}
private:
bool analyze_render_audio_called_;
float last_render_audio_first_sample_;
};
// Mocks CustomProcessing and applies ProcessSample() to all the samples.
// Meant to be injected into an APM to modify samples in a known and detectable
// way.
class TestRenderPreProcessor : public CustomProcessing {
public:
TestRenderPreProcessor() = default;
~TestRenderPreProcessor() = default;
void Initialize(int sample_rate_hz, int num_channels) override {}
void Process(AudioBuffer* audio) override {
for (size_t k = 0; k < audio->num_channels(); ++k) {
rtc::ArrayView<float> channel_view(audio->channels_f()[k],
audio->num_frames());
std::transform(channel_view.begin(), channel_view.end(),
channel_view.begin(), ProcessSample);
}
};
std::string ToString() const override { return "TestRenderPreProcessor"; }
void SetRuntimeSetting(AudioProcessing::RuntimeSetting setting) override {}
// Modifies a sample. This member is used in Process() to modify a frame and
// it is publicly visible to enable tests.
static constexpr float ProcessSample(float x) { return 2.f * x; }
};
} // namespace
TEST(AudioProcessingImplTest, AudioParameterChangeTriggersInit) {
webrtc::Config config;
MockInitialize mock(config);
ON_CALL(mock, InitializeLocked())
.WillByDefault(Invoke(&mock, &MockInitialize::RealInitializeLocked));
EXPECT_CALL(mock, InitializeLocked()).Times(1);
mock.Initialize();
AudioFrame frame;
// Call with the default parameters; there should be an init.
frame.num_channels_ = 1;
SetFrameSampleRate(&frame, 16000);
EXPECT_CALL(mock, InitializeLocked()).Times(0);
EXPECT_NOERR(mock.ProcessStream(&frame));
EXPECT_NOERR(mock.ProcessReverseStream(&frame));
// New sample rate. (Only impacts ProcessStream).
SetFrameSampleRate(&frame, 32000);
EXPECT_CALL(mock, InitializeLocked()).Times(1);
EXPECT_NOERR(mock.ProcessStream(&frame));
// New number of channels.
// TODO(peah): Investigate why this causes 2 inits.
frame.num_channels_ = 2;
EXPECT_CALL(mock, InitializeLocked()).Times(2);
EXPECT_NOERR(mock.ProcessStream(&frame));
// ProcessStream sets num_channels_ == num_output_channels.
frame.num_channels_ = 2;
EXPECT_NOERR(mock.ProcessReverseStream(&frame));
// A new sample rate passed to ProcessReverseStream should cause an init.
SetFrameSampleRate(&frame, 16000);
EXPECT_CALL(mock, InitializeLocked()).Times(1);
EXPECT_NOERR(mock.ProcessReverseStream(&frame));
}
TEST(AudioProcessingImplTest, UpdateCapturePreGainRuntimeSetting) {
std::unique_ptr<AudioProcessing> apm(AudioProcessingBuilder().Create());
webrtc::AudioProcessing::Config apm_config;
apm_config.pre_amplifier.enabled = true;
apm_config.pre_amplifier.fixed_gain_factor = 1.f;
apm->ApplyConfig(apm_config);
AudioFrame frame;
constexpr int16_t kAudioLevel = 10000;
constexpr size_t kSampleRateHz = 48000;
constexpr size_t kNumChannels = 2;
InitializeAudioFrame(kSampleRateHz, kNumChannels, &frame);
FillFixedFrame(kAudioLevel, &frame);
apm->ProcessStream(&frame);
EXPECT_EQ(frame.data()[100], kAudioLevel)
<< "With factor 1, frame shouldn't be modified.";
constexpr float kGainFactor = 2.f;
apm->SetRuntimeSetting(
AudioProcessing::RuntimeSetting::CreateCapturePreGain(kGainFactor));
// Process for two frames to have time to ramp up gain.
for (int i = 0; i < 2; ++i) {
FillFixedFrame(kAudioLevel, &frame);
apm->ProcessStream(&frame);
}
EXPECT_EQ(frame.data()[100], kGainFactor * kAudioLevel)
<< "Frame should be amplified.";
}
TEST(AudioProcessingImplTest,
EchoControllerObservesPreAmplifierEchoPathGainChange) {
// Tests that the echo controller observes an echo path gain change when the
// pre-amplifier submodule changes the gain.
auto echo_control_factory = absl::make_unique<MockEchoControlFactory>();
const auto* echo_control_factory_ptr = echo_control_factory.get();
std::unique_ptr<AudioProcessing> apm(
AudioProcessingBuilder()
.SetEchoControlFactory(std::move(echo_control_factory))
.Create());
apm->gain_control()->Enable(false); // Disable AGC.
apm->gain_control()->set_mode(GainControl::Mode::kFixedDigital);
webrtc::AudioProcessing::Config apm_config;
apm_config.gain_controller2.enabled = false;
apm_config.pre_amplifier.enabled = true;
apm_config.pre_amplifier.fixed_gain_factor = 1.f;
apm->ApplyConfig(apm_config);
AudioFrame frame;
constexpr int16_t kAudioLevel = 10000;
constexpr size_t kSampleRateHz = 48000;
constexpr size_t kNumChannels = 2;
InitializeAudioFrame(kSampleRateHz, kNumChannels, &frame);
FillFixedFrame(kAudioLevel, &frame);
MockEchoControl* echo_control_mock = echo_control_factory_ptr->GetNext();
EXPECT_CALL(*echo_control_mock, AnalyzeCapture(NotNull())).Times(1);
EXPECT_CALL(*echo_control_mock, ProcessCapture(NotNull(), false)).Times(1);
apm->ProcessStream(&frame);
EXPECT_CALL(*echo_control_mock, AnalyzeCapture(NotNull())).Times(1);
EXPECT_CALL(*echo_control_mock, ProcessCapture(NotNull(), true)).Times(1);
apm->SetRuntimeSetting(
AudioProcessing::RuntimeSetting::CreateCapturePreGain(2.f));
apm->ProcessStream(&frame);
}
TEST(AudioProcessingImplTest,
EchoControllerObservesAnalogAgc1EchoPathGainChange) {
// Tests that the echo controller observes an echo path gain change when the
// AGC1 analog adaptive submodule changes the analog gain.
auto echo_control_factory = absl::make_unique<MockEchoControlFactory>();
const auto* echo_control_factory_ptr = echo_control_factory.get();
std::unique_ptr<AudioProcessing> apm(
AudioProcessingBuilder()
.SetEchoControlFactory(std::move(echo_control_factory))
.Create());
apm->gain_control()->Enable(true); // Enable AGC.
apm->gain_control()->set_mode(GainControl::Mode::kAdaptiveAnalog);
webrtc::AudioProcessing::Config apm_config;
apm_config.gain_controller2.enabled = false;
apm_config.pre_amplifier.enabled = false;
apm->ApplyConfig(apm_config);
AudioFrame frame;
constexpr int16_t kAudioLevel = 1000;
constexpr size_t kSampleRateHz = 48000;
constexpr size_t kNumChannels = 2;
InitializeAudioFrame(kSampleRateHz, kNumChannels, &frame);
FillFixedFrame(kAudioLevel, &frame);
MockEchoControl* echo_control_mock = echo_control_factory_ptr->GetNext();
const int initial_analog_gain = apm->gain_control()->stream_analog_level();
EXPECT_CALL(*echo_control_mock, AnalyzeCapture(NotNull())).Times(1);
EXPECT_CALL(*echo_control_mock, ProcessCapture(NotNull(), false)).Times(1);
apm->ProcessStream(&frame);
// Force an analog gain change if it did not happen.
if (initial_analog_gain == apm->gain_control()->stream_analog_level()) {
apm->gain_control()->set_stream_analog_level(initial_analog_gain + 1);
}
EXPECT_CALL(*echo_control_mock, AnalyzeCapture(NotNull())).Times(1);
EXPECT_CALL(*echo_control_mock, ProcessCapture(NotNull(), true)).Times(1);
apm->ProcessStream(&frame);
}
TEST(AudioProcessingImplTest, RenderPreProcessorBeforeEchoDetector) {
// Make sure that signal changes caused by a render pre-processing sub-module
// take place before any echo detector analysis.
rtc::scoped_refptr<TestEchoDetector> test_echo_detector(
new rtc::RefCountedObject<TestEchoDetector>());
std::unique_ptr<CustomProcessing> test_render_pre_processor(
new TestRenderPreProcessor());
// Create APM injecting the test echo detector and render pre-processor.
std::unique_ptr<AudioProcessing> apm(
AudioProcessingBuilder()
.SetEchoDetector(test_echo_detector)
.SetRenderPreProcessing(std::move(test_render_pre_processor))
.Create());
webrtc::AudioProcessing::Config apm_config;
apm_config.pre_amplifier.enabled = true;
apm_config.residual_echo_detector.enabled = true;
apm->ApplyConfig(apm_config);
constexpr int16_t kAudioLevel = 1000;
constexpr int kSampleRateHz = 16000;
constexpr size_t kNumChannels = 1;
AudioFrame frame;
InitializeAudioFrame(kSampleRateHz, kNumChannels, &frame);
constexpr float kAudioLevelFloat = static_cast<float>(kAudioLevel);
constexpr float kExpectedPreprocessedAudioLevel =
TestRenderPreProcessor::ProcessSample(kAudioLevelFloat);
ASSERT_NE(kAudioLevelFloat, kExpectedPreprocessedAudioLevel);
// Analyze a render stream frame.
FillFixedFrame(kAudioLevel, &frame);
ASSERT_EQ(AudioProcessing::Error::kNoError,
apm->ProcessReverseStream(&frame));
// Trigger a call to in EchoDetector::AnalyzeRenderAudio() via
// ProcessStream().
FillFixedFrame(kAudioLevel, &frame);
ASSERT_EQ(AudioProcessing::Error::kNoError, apm->ProcessStream(&frame));
// Regardless of how the call to in EchoDetector::AnalyzeRenderAudio() is
// triggered, the line below checks that the call has occurred. If not, the
// APM implementation may have changed and this test might need to be adapted.
ASSERT_TRUE(test_echo_detector->analyze_render_audio_called());
// Check that the data read in EchoDetector::AnalyzeRenderAudio() is that
// produced by the render pre-processor.
EXPECT_EQ(kExpectedPreprocessedAudioLevel,
test_echo_detector->last_render_audio_first_sample());
}
} // namespace webrtc
|