blob: f82ef965db73fb304e3f2a8d001b8e517dcb59b9 [file] [log] [blame]
/*
* Copyright (c) 2015 The WebRTC project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include "api/audio_codecs/opus/audio_encoder_opus.h"
#include <array>
#include <memory>
#include <utility>
#include "absl/strings/string_view.h"
#include "common_audio/mocks/mock_smoothing_filter.h"
#include "modules/audio_coding/audio_network_adaptor/mock/mock_audio_network_adaptor.h"
#include "modules/audio_coding/codecs/opus/audio_encoder_opus.h"
#include "modules/audio_coding/codecs/opus/opus_interface.h"
#include "modules/audio_coding/neteq/tools/audio_loop.h"
#include "rtc_base/checks.h"
#include "rtc_base/fake_clock.h"
#include "test/field_trial.h"
#include "test/gmock.h"
#include "test/gtest.h"
#include "test/testsupport/file_utils.h"
namespace webrtc {
using ::testing::NiceMock;
using ::testing::Return;
namespace {
constexpr int kDefaultOpusPayloadType = 105;
constexpr int kDefaultOpusRate = 32000;
constexpr int kDefaultOpusPacSize = 960;
constexpr int64_t kInitialTimeUs = 12345678;
AudioEncoderOpusConfig CreateConfigWithParameters(
const CodecParameterMap& params) {
const SdpAudioFormat format("opus", 48000, 2, params);
return *AudioEncoderOpus::SdpToConfig(format);
}
struct AudioEncoderOpusStates {
MockAudioNetworkAdaptor* mock_audio_network_adaptor;
MockSmoothingFilter* mock_bitrate_smoother;
std::unique_ptr<AudioEncoderOpusImpl> encoder;
std::unique_ptr<rtc::ScopedFakeClock> fake_clock;
AudioEncoderOpusConfig config;
};
std::unique_ptr<AudioEncoderOpusStates> CreateCodec(int sample_rate_hz,
size_t num_channels) {
std::unique_ptr<AudioEncoderOpusStates> states =
std::make_unique<AudioEncoderOpusStates>();
states->mock_audio_network_adaptor = nullptr;
states->fake_clock.reset(new rtc::ScopedFakeClock());
states->fake_clock->SetTime(Timestamp::Micros(kInitialTimeUs));
MockAudioNetworkAdaptor** mock_ptr = &states->mock_audio_network_adaptor;
AudioEncoderOpusImpl::AudioNetworkAdaptorCreator creator =
[mock_ptr](absl::string_view, RtcEventLog* event_log) {
std::unique_ptr<MockAudioNetworkAdaptor> adaptor(
new NiceMock<MockAudioNetworkAdaptor>());
EXPECT_CALL(*adaptor, Die());
*mock_ptr = adaptor.get();
return adaptor;
};
AudioEncoderOpusConfig config;
config.frame_size_ms = rtc::CheckedDivExact(kDefaultOpusPacSize, 48);
config.sample_rate_hz = sample_rate_hz;
config.num_channels = num_channels;
config.bitrate_bps = kDefaultOpusRate;
config.application = num_channels == 1
? AudioEncoderOpusConfig::ApplicationMode::kVoip
: AudioEncoderOpusConfig::ApplicationMode::kAudio;
config.supported_frame_lengths_ms.push_back(config.frame_size_ms);
states->config = config;
std::unique_ptr<MockSmoothingFilter> bitrate_smoother(
new MockSmoothingFilter());
states->mock_bitrate_smoother = bitrate_smoother.get();
states->encoder.reset(
new AudioEncoderOpusImpl(states->config, kDefaultOpusPayloadType, creator,
std::move(bitrate_smoother)));
return states;
}
AudioEncoderRuntimeConfig CreateEncoderRuntimeConfig() {
constexpr int kBitrate = 40000;
constexpr int kFrameLength = 60;
constexpr bool kEnableDtx = false;
constexpr size_t kNumChannels = 1;
AudioEncoderRuntimeConfig config;
config.bitrate_bps = kBitrate;
config.frame_length_ms = kFrameLength;
config.enable_dtx = kEnableDtx;
config.num_channels = kNumChannels;
return config;
}
void CheckEncoderRuntimeConfig(const AudioEncoderOpusImpl* encoder,
const AudioEncoderRuntimeConfig& config) {
EXPECT_EQ(*config.bitrate_bps, encoder->GetTargetBitrate());
EXPECT_EQ(*config.frame_length_ms, encoder->next_frame_length_ms());
EXPECT_EQ(*config.enable_dtx, encoder->GetDtx());
EXPECT_EQ(*config.num_channels, encoder->num_channels_to_encode());
}
// Create 10ms audio data blocks for a total packet size of "packet_size_ms".
std::unique_ptr<test::AudioLoop> Create10msAudioBlocks(
const std::unique_ptr<AudioEncoderOpusImpl>& encoder,
int packet_size_ms) {
const std::string file_name =
test::ResourcePath("audio_coding/testfile32kHz", "pcm");
std::unique_ptr<test::AudioLoop> speech_data(new test::AudioLoop());
int audio_samples_per_ms =
rtc::CheckedDivExact(encoder->SampleRateHz(), 1000);
if (!speech_data->Init(
file_name,
packet_size_ms * audio_samples_per_ms *
encoder->num_channels_to_encode(),
10 * audio_samples_per_ms * encoder->num_channels_to_encode()))
return nullptr;
return speech_data;
}
} // namespace
class AudioEncoderOpusTest : public ::testing::TestWithParam<int> {
protected:
int sample_rate_hz_{GetParam()};
};
INSTANTIATE_TEST_SUITE_P(Param,
AudioEncoderOpusTest,
::testing::Values(16000, 48000));
TEST_P(AudioEncoderOpusTest, DefaultApplicationModeMono) {
auto states = CreateCodec(sample_rate_hz_, 1);
EXPECT_EQ(AudioEncoderOpusConfig::ApplicationMode::kVoip,
states->encoder->application());
}
TEST_P(AudioEncoderOpusTest, DefaultApplicationModeStereo) {
auto states = CreateCodec(sample_rate_hz_, 2);
EXPECT_EQ(AudioEncoderOpusConfig::ApplicationMode::kAudio,
states->encoder->application());
}
TEST_P(AudioEncoderOpusTest, ChangeApplicationMode) {
auto states = CreateCodec(sample_rate_hz_, 2);
EXPECT_TRUE(
states->encoder->SetApplication(AudioEncoder::Application::kSpeech));
EXPECT_EQ(AudioEncoderOpusConfig::ApplicationMode::kVoip,
states->encoder->application());
}
TEST_P(AudioEncoderOpusTest, ResetWontChangeApplicationMode) {
auto states = CreateCodec(sample_rate_hz_, 2);
// Trigger a reset.
states->encoder->Reset();
// Verify that the mode is still kAudio.
EXPECT_EQ(AudioEncoderOpusConfig::ApplicationMode::kAudio,
states->encoder->application());
// Now change to kVoip.
EXPECT_TRUE(
states->encoder->SetApplication(AudioEncoder::Application::kSpeech));
EXPECT_EQ(AudioEncoderOpusConfig::ApplicationMode::kVoip,
states->encoder->application());
// Trigger a reset again.
states->encoder->Reset();
// Verify that the mode is still kVoip.
EXPECT_EQ(AudioEncoderOpusConfig::ApplicationMode::kVoip,
states->encoder->application());
}
TEST_P(AudioEncoderOpusTest, ToggleDtx) {
auto states = CreateCodec(sample_rate_hz_, 2);
// Enable DTX
EXPECT_TRUE(states->encoder->SetDtx(true));
EXPECT_TRUE(states->encoder->GetDtx());
// Turn off DTX.
EXPECT_TRUE(states->encoder->SetDtx(false));
EXPECT_FALSE(states->encoder->GetDtx());
}
TEST_P(AudioEncoderOpusTest,
OnReceivedUplinkBandwidthWithoutAudioNetworkAdaptor) {
auto states = CreateCodec(sample_rate_hz_, 1);
// Constants are replicated from audio_states->encoderopus.cc.
const int kMinBitrateBps = 6000;
const int kMaxBitrateBps = 510000;
const int kOverheadBytesPerPacket = 64;
states->encoder->OnReceivedOverhead(kOverheadBytesPerPacket);
const int kOverheadBps = 8 * kOverheadBytesPerPacket *
rtc::CheckedDivExact(48000, kDefaultOpusPacSize);
// Set a too low bitrate.
states->encoder->OnReceivedUplinkBandwidth(kMinBitrateBps + kOverheadBps - 1,
absl::nullopt);
EXPECT_EQ(kMinBitrateBps, states->encoder->GetTargetBitrate());
// Set a too high bitrate.
states->encoder->OnReceivedUplinkBandwidth(kMaxBitrateBps + kOverheadBps + 1,
absl::nullopt);
EXPECT_EQ(kMaxBitrateBps, states->encoder->GetTargetBitrate());
// Set the minimum rate.
states->encoder->OnReceivedUplinkBandwidth(kMinBitrateBps + kOverheadBps,
absl::nullopt);
EXPECT_EQ(kMinBitrateBps, states->encoder->GetTargetBitrate());
// Set the maximum rate.
states->encoder->OnReceivedUplinkBandwidth(kMaxBitrateBps + kOverheadBps,
absl::nullopt);
EXPECT_EQ(kMaxBitrateBps, states->encoder->GetTargetBitrate());
// Set rates from kMaxBitrateBps up to 32000 bps.
for (int rate = kMinBitrateBps + kOverheadBps; rate <= 32000 + kOverheadBps;
rate += 1000) {
states->encoder->OnReceivedUplinkBandwidth(rate, absl::nullopt);
EXPECT_EQ(rate - kOverheadBps, states->encoder->GetTargetBitrate());
}
}
TEST_P(AudioEncoderOpusTest, SetReceiverFrameLengthRange) {
auto states = CreateCodec(sample_rate_hz_, 2);
// Before calling to `SetReceiverFrameLengthRange`,
// `supported_frame_lengths_ms` should contain only the frame length being
// used.
using ::testing::ElementsAre;
EXPECT_THAT(states->encoder->supported_frame_lengths_ms(),
ElementsAre(states->encoder->next_frame_length_ms()));
states->encoder->SetReceiverFrameLengthRange(0, 12345);
states->encoder->SetReceiverFrameLengthRange(21, 60);
EXPECT_THAT(states->encoder->supported_frame_lengths_ms(),
ElementsAre(40, 60));
states->encoder->SetReceiverFrameLengthRange(20, 59);
EXPECT_THAT(states->encoder->supported_frame_lengths_ms(),
ElementsAre(20, 40));
}
TEST_P(AudioEncoderOpusTest,
InvokeAudioNetworkAdaptorOnReceivedUplinkPacketLossFraction) {
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->EnableAudioNetworkAdaptor("", nullptr);
auto config = CreateEncoderRuntimeConfig();
EXPECT_CALL(*states->mock_audio_network_adaptor, GetEncoderRuntimeConfig())
.WillOnce(Return(config));
// Since using mock audio network adaptor, any packet loss fraction is fine.
constexpr float kUplinkPacketLoss = 0.1f;
EXPECT_CALL(*states->mock_audio_network_adaptor,
SetUplinkPacketLossFraction(kUplinkPacketLoss));
states->encoder->OnReceivedUplinkPacketLossFraction(kUplinkPacketLoss);
CheckEncoderRuntimeConfig(states->encoder.get(), config);
}
TEST_P(AudioEncoderOpusTest,
InvokeAudioNetworkAdaptorOnReceivedUplinkBandwidth) {
test::ScopedFieldTrials override_field_trials(
"WebRTC-Audio-StableTargetAdaptation/Disabled/");
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->EnableAudioNetworkAdaptor("", nullptr);
auto config = CreateEncoderRuntimeConfig();
EXPECT_CALL(*states->mock_audio_network_adaptor, GetEncoderRuntimeConfig())
.WillOnce(Return(config));
// Since using mock audio network adaptor, any target audio bitrate is fine.
constexpr int kTargetAudioBitrate = 30000;
constexpr int64_t kProbingIntervalMs = 3000;
EXPECT_CALL(*states->mock_audio_network_adaptor,
SetTargetAudioBitrate(kTargetAudioBitrate));
EXPECT_CALL(*states->mock_bitrate_smoother,
SetTimeConstantMs(kProbingIntervalMs * 4));
EXPECT_CALL(*states->mock_bitrate_smoother, AddSample(kTargetAudioBitrate));
states->encoder->OnReceivedUplinkBandwidth(kTargetAudioBitrate,
kProbingIntervalMs);
CheckEncoderRuntimeConfig(states->encoder.get(), config);
}
TEST_P(AudioEncoderOpusTest,
InvokeAudioNetworkAdaptorOnReceivedUplinkAllocation) {
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->EnableAudioNetworkAdaptor("", nullptr);
auto config = CreateEncoderRuntimeConfig();
EXPECT_CALL(*states->mock_audio_network_adaptor, GetEncoderRuntimeConfig())
.WillOnce(Return(config));
BitrateAllocationUpdate update;
update.target_bitrate = DataRate::BitsPerSec(30000);
update.stable_target_bitrate = DataRate::BitsPerSec(20000);
update.bwe_period = TimeDelta::Millis(200);
EXPECT_CALL(*states->mock_audio_network_adaptor,
SetTargetAudioBitrate(update.target_bitrate.bps()));
EXPECT_CALL(*states->mock_audio_network_adaptor,
SetUplinkBandwidth(update.stable_target_bitrate.bps()));
states->encoder->OnReceivedUplinkAllocation(update);
CheckEncoderRuntimeConfig(states->encoder.get(), config);
}
TEST_P(AudioEncoderOpusTest, InvokeAudioNetworkAdaptorOnReceivedRtt) {
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->EnableAudioNetworkAdaptor("", nullptr);
auto config = CreateEncoderRuntimeConfig();
EXPECT_CALL(*states->mock_audio_network_adaptor, GetEncoderRuntimeConfig())
.WillOnce(Return(config));
// Since using mock audio network adaptor, any rtt is fine.
constexpr int kRtt = 30;
EXPECT_CALL(*states->mock_audio_network_adaptor, SetRtt(kRtt));
states->encoder->OnReceivedRtt(kRtt);
CheckEncoderRuntimeConfig(states->encoder.get(), config);
}
TEST_P(AudioEncoderOpusTest, InvokeAudioNetworkAdaptorOnReceivedOverhead) {
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->EnableAudioNetworkAdaptor("", nullptr);
auto config = CreateEncoderRuntimeConfig();
EXPECT_CALL(*states->mock_audio_network_adaptor, GetEncoderRuntimeConfig())
.WillOnce(Return(config));
// Since using mock audio network adaptor, any overhead is fine.
constexpr size_t kOverhead = 64;
EXPECT_CALL(*states->mock_audio_network_adaptor, SetOverhead(kOverhead));
states->encoder->OnReceivedOverhead(kOverhead);
CheckEncoderRuntimeConfig(states->encoder.get(), config);
}
TEST_P(AudioEncoderOpusTest,
PacketLossFractionSmoothedOnSetUplinkPacketLossFraction) {
auto states = CreateCodec(sample_rate_hz_, 2);
// The values are carefully chosen so that if no smoothing is made, the test
// will fail.
constexpr float kPacketLossFraction_1 = 0.02f;
constexpr float kPacketLossFraction_2 = 0.198f;
// `kSecondSampleTimeMs` is chosen to ease the calculation since
// 0.9999 ^ 6931 = 0.5.
constexpr int64_t kSecondSampleTimeMs = 6931;
// First time, no filtering.
states->encoder->OnReceivedUplinkPacketLossFraction(kPacketLossFraction_1);
EXPECT_FLOAT_EQ(0.02f, states->encoder->packet_loss_rate());
states->fake_clock->AdvanceTime(TimeDelta::Millis(kSecondSampleTimeMs));
states->encoder->OnReceivedUplinkPacketLossFraction(kPacketLossFraction_2);
// Now the output of packet loss fraction smoother should be
// (0.02 + 0.198) / 2 = 0.109.
EXPECT_NEAR(0.109f, states->encoder->packet_loss_rate(), 0.001);
}
TEST_P(AudioEncoderOpusTest, PacketLossRateUpperBounded) {
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->OnReceivedUplinkPacketLossFraction(0.5);
EXPECT_FLOAT_EQ(0.2f, states->encoder->packet_loss_rate());
}
TEST_P(AudioEncoderOpusTest, DoNotInvokeSetTargetBitrateIfOverheadUnknown) {
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->OnReceivedUplinkBandwidth(kDefaultOpusRate * 2,
absl::nullopt);
// Since `OnReceivedOverhead` has not been called, the codec bitrate should
// not change.
EXPECT_EQ(kDefaultOpusRate, states->encoder->GetTargetBitrate());
}
// Verifies that the complexity adaptation in the config works as intended.
TEST(AudioEncoderOpusTest, ConfigComplexityAdaptation) {
AudioEncoderOpusConfig config;
config.low_rate_complexity = 8;
config.complexity = 6;
// Bitrate within hysteresis window. Expect empty output.
config.bitrate_bps = 12500;
EXPECT_EQ(absl::nullopt, AudioEncoderOpusImpl::GetNewComplexity(config));
// Bitrate below hysteresis window. Expect higher complexity.
config.bitrate_bps = 10999;
EXPECT_EQ(8, AudioEncoderOpusImpl::GetNewComplexity(config));
// Bitrate within hysteresis window. Expect empty output.
config.bitrate_bps = 12500;
EXPECT_EQ(absl::nullopt, AudioEncoderOpusImpl::GetNewComplexity(config));
// Bitrate above hysteresis window. Expect lower complexity.
config.bitrate_bps = 14001;
EXPECT_EQ(6, AudioEncoderOpusImpl::GetNewComplexity(config));
}
// Verifies that the bandwidth adaptation in the config works as intended.
TEST_P(AudioEncoderOpusTest, ConfigBandwidthAdaptation) {
AudioEncoderOpusConfig config;
const size_t opus_rate_khz = rtc::CheckedDivExact(sample_rate_hz_, 1000);
const std::vector<int16_t> silence(
opus_rate_khz * config.frame_size_ms * config.num_channels, 0);
constexpr size_t kMaxBytes = 1000;
uint8_t bitstream[kMaxBytes];
OpusEncInst* inst;
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(
&inst, config.num_channels,
config.application ==
AudioEncoderOpusConfig::ApplicationMode::kVoip
? 0
: 1,
sample_rate_hz_));
// Bitrate below minmum wideband. Expect narrowband.
config.bitrate_bps = absl::optional<int>(7999);
auto bandwidth = AudioEncoderOpusImpl::GetNewBandwidth(config, inst);
EXPECT_EQ(absl::optional<int>(OPUS_BANDWIDTH_NARROWBAND), bandwidth);
WebRtcOpus_SetBandwidth(inst, *bandwidth);
// It is necessary to encode here because Opus has some logic in the encoder
// that goes from the user-set bandwidth to the used and returned one.
WebRtcOpus_Encode(inst, silence.data(),
rtc::CheckedDivExact(silence.size(), config.num_channels),
kMaxBytes, bitstream);
// Bitrate not yet above maximum narrowband. Expect empty.
config.bitrate_bps = absl::optional<int>(9000);
bandwidth = AudioEncoderOpusImpl::GetNewBandwidth(config, inst);
EXPECT_EQ(absl::optional<int>(), bandwidth);
// Bitrate above maximum narrowband. Expect wideband.
config.bitrate_bps = absl::optional<int>(9001);
bandwidth = AudioEncoderOpusImpl::GetNewBandwidth(config, inst);
EXPECT_EQ(absl::optional<int>(OPUS_BANDWIDTH_WIDEBAND), bandwidth);
WebRtcOpus_SetBandwidth(inst, *bandwidth);
// It is necessary to encode here because Opus has some logic in the encoder
// that goes from the user-set bandwidth to the used and returned one.
WebRtcOpus_Encode(inst, silence.data(),
rtc::CheckedDivExact(silence.size(), config.num_channels),
kMaxBytes, bitstream);
// Bitrate not yet below minimum wideband. Expect empty.
config.bitrate_bps = absl::optional<int>(8000);
bandwidth = AudioEncoderOpusImpl::GetNewBandwidth(config, inst);
EXPECT_EQ(absl::optional<int>(), bandwidth);
// Bitrate above automatic threshold. Expect automatic.
config.bitrate_bps = absl::optional<int>(12001);
bandwidth = AudioEncoderOpusImpl::GetNewBandwidth(config, inst);
EXPECT_EQ(absl::optional<int>(OPUS_AUTO), bandwidth);
EXPECT_EQ(0, WebRtcOpus_EncoderFree(inst));
}
TEST_P(AudioEncoderOpusTest, EmptyConfigDoesNotAffectEncoderSettings) {
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->EnableAudioNetworkAdaptor("", nullptr);
auto config = CreateEncoderRuntimeConfig();
AudioEncoderRuntimeConfig empty_config;
EXPECT_CALL(*states->mock_audio_network_adaptor, GetEncoderRuntimeConfig())
.WillOnce(Return(config))
.WillOnce(Return(empty_config));
constexpr size_t kOverhead = 64;
EXPECT_CALL(*states->mock_audio_network_adaptor, SetOverhead(kOverhead))
.Times(2);
states->encoder->OnReceivedOverhead(kOverhead);
states->encoder->OnReceivedOverhead(kOverhead);
CheckEncoderRuntimeConfig(states->encoder.get(), config);
}
TEST_P(AudioEncoderOpusTest, UpdateUplinkBandwidthInAudioNetworkAdaptor) {
test::ScopedFieldTrials override_field_trials(
"WebRTC-Audio-StableTargetAdaptation/Disabled/");
auto states = CreateCodec(sample_rate_hz_, 2);
states->encoder->EnableAudioNetworkAdaptor("", nullptr);
const size_t opus_rate_khz = rtc::CheckedDivExact(sample_rate_hz_, 1000);
const std::vector<int16_t> audio(opus_rate_khz * 10 * 2, 0);
rtc::Buffer encoded;
EXPECT_CALL(*states->mock_bitrate_smoother, GetAverage())
.WillOnce(Return(50000));
EXPECT_CALL(*states->mock_audio_network_adaptor, SetUplinkBandwidth(50000));
states->encoder->Encode(
0, rtc::ArrayView<const int16_t>(audio.data(), audio.size()), &encoded);
// Repeat update uplink bandwidth tests.
for (int i = 0; i < 5; i++) {
// Don't update till it is time to update again.
states->fake_clock->AdvanceTime(TimeDelta::Millis(
states->config.uplink_bandwidth_update_interval_ms - 1));
states->encoder->Encode(
0, rtc::ArrayView<const int16_t>(audio.data(), audio.size()), &encoded);
// Update when it is time to update.
EXPECT_CALL(*states->mock_bitrate_smoother, GetAverage())
.WillOnce(Return(40000));
EXPECT_CALL(*states->mock_audio_network_adaptor, SetUplinkBandwidth(40000));
states->fake_clock->AdvanceTime(TimeDelta::Millis(1));
states->encoder->Encode(
0, rtc::ArrayView<const int16_t>(audio.data(), audio.size()), &encoded);
}
}
TEST_P(AudioEncoderOpusTest, EncodeAtMinBitrate) {
auto states = CreateCodec(sample_rate_hz_, 1);
constexpr int kNumPacketsToEncode = 2;
auto audio_frames =
Create10msAudioBlocks(states->encoder, kNumPacketsToEncode * 20);
ASSERT_TRUE(audio_frames) << "Create10msAudioBlocks failed";
rtc::Buffer encoded;
uint32_t rtp_timestamp = 12345; // Just a number not important to this test.
states->encoder->OnReceivedUplinkBandwidth(0, absl::nullopt);
for (int packet_index = 0; packet_index < kNumPacketsToEncode;
packet_index++) {
// Make sure we are not encoding before we have enough data for
// a 20ms packet.
for (int index = 0; index < 1; index++) {
states->encoder->Encode(rtp_timestamp, audio_frames->GetNextBlock(),
&encoded);
EXPECT_EQ(0u, encoded.size());
}
// Should encode now.
states->encoder->Encode(rtp_timestamp, audio_frames->GetNextBlock(),
&encoded);
EXPECT_GT(encoded.size(), 0u);
encoded.Clear();
}
}
TEST(AudioEncoderOpusTest, TestConfigDefaults) {
const auto config_opt = AudioEncoderOpus::SdpToConfig({"opus", 48000, 2});
ASSERT_TRUE(config_opt);
EXPECT_EQ(48000, config_opt->max_playback_rate_hz);
EXPECT_EQ(1u, config_opt->num_channels);
EXPECT_FALSE(config_opt->fec_enabled);
EXPECT_FALSE(config_opt->dtx_enabled);
EXPECT_EQ(20, config_opt->frame_size_ms);
}
TEST(AudioEncoderOpusTest, TestConfigFromParams) {
const auto config1 = CreateConfigWithParameters({{"stereo", "0"}});
EXPECT_EQ(1U, config1.num_channels);
const auto config2 = CreateConfigWithParameters({{"stereo", "1"}});
EXPECT_EQ(2U, config2.num_channels);
const auto config3 = CreateConfigWithParameters({{"useinbandfec", "0"}});
EXPECT_FALSE(config3.fec_enabled);
const auto config4 = CreateConfigWithParameters({{"useinbandfec", "1"}});
EXPECT_TRUE(config4.fec_enabled);
const auto config5 = CreateConfigWithParameters({{"usedtx", "0"}});
EXPECT_FALSE(config5.dtx_enabled);
const auto config6 = CreateConfigWithParameters({{"usedtx", "1"}});
EXPECT_TRUE(config6.dtx_enabled);
const auto config7 = CreateConfigWithParameters({{"cbr", "0"}});
EXPECT_FALSE(config7.cbr_enabled);
const auto config8 = CreateConfigWithParameters({{"cbr", "1"}});
EXPECT_TRUE(config8.cbr_enabled);
const auto config9 =
CreateConfigWithParameters({{"maxplaybackrate", "12345"}});
EXPECT_EQ(12345, config9.max_playback_rate_hz);
const auto config10 =
CreateConfigWithParameters({{"maxaveragebitrate", "96000"}});
EXPECT_EQ(96000, config10.bitrate_bps);
const auto config11 = CreateConfigWithParameters({{"maxptime", "40"}});
for (int frame_length : config11.supported_frame_lengths_ms) {
EXPECT_LE(frame_length, 40);
}
const auto config12 = CreateConfigWithParameters({{"minptime", "40"}});
for (int frame_length : config12.supported_frame_lengths_ms) {
EXPECT_GE(frame_length, 40);
}
const auto config13 = CreateConfigWithParameters({{"ptime", "40"}});
EXPECT_EQ(40, config13.frame_size_ms);
constexpr int kMinSupportedFrameLength = 10;
constexpr int kMaxSupportedFrameLength =
WEBRTC_OPUS_SUPPORT_120MS_PTIME ? 120 : 60;
const auto config14 = CreateConfigWithParameters({{"ptime", "1"}});
EXPECT_EQ(kMinSupportedFrameLength, config14.frame_size_ms);
const auto config15 = CreateConfigWithParameters({{"ptime", "2000"}});
EXPECT_EQ(kMaxSupportedFrameLength, config15.frame_size_ms);
}
TEST(AudioEncoderOpusTest, TestConfigFromInvalidParams) {
const webrtc::SdpAudioFormat format("opus", 48000, 2);
const auto default_config = *AudioEncoderOpus::SdpToConfig(format);
#if WEBRTC_OPUS_SUPPORT_120MS_PTIME
const std::vector<int> default_supported_frame_lengths_ms({20, 40, 60, 120});
#else
const std::vector<int> default_supported_frame_lengths_ms({20, 40, 60});
#endif
AudioEncoderOpusConfig config;
config = CreateConfigWithParameters({{"stereo", "invalid"}});
EXPECT_EQ(default_config.num_channels, config.num_channels);
config = CreateConfigWithParameters({{"useinbandfec", "invalid"}});
EXPECT_EQ(default_config.fec_enabled, config.fec_enabled);
config = CreateConfigWithParameters({{"usedtx", "invalid"}});
EXPECT_EQ(default_config.dtx_enabled, config.dtx_enabled);
config = CreateConfigWithParameters({{"cbr", "invalid"}});
EXPECT_EQ(default_config.dtx_enabled, config.dtx_enabled);
config = CreateConfigWithParameters({{"maxplaybackrate", "0"}});
EXPECT_EQ(default_config.max_playback_rate_hz, config.max_playback_rate_hz);
config = CreateConfigWithParameters({{"maxplaybackrate", "-23"}});
EXPECT_EQ(default_config.max_playback_rate_hz, config.max_playback_rate_hz);
config = CreateConfigWithParameters({{"maxplaybackrate", "not a number!"}});
EXPECT_EQ(default_config.max_playback_rate_hz, config.max_playback_rate_hz);
config = CreateConfigWithParameters({{"maxaveragebitrate", "0"}});
EXPECT_EQ(6000, config.bitrate_bps);
config = CreateConfigWithParameters({{"maxaveragebitrate", "-1000"}});
EXPECT_EQ(6000, config.bitrate_bps);
config = CreateConfigWithParameters({{"maxaveragebitrate", "1024000"}});
EXPECT_EQ(510000, config.bitrate_bps);
config = CreateConfigWithParameters({{"maxaveragebitrate", "not a number!"}});
EXPECT_EQ(default_config.bitrate_bps, config.bitrate_bps);
config = CreateConfigWithParameters({{"maxptime", "invalid"}});
EXPECT_EQ(default_supported_frame_lengths_ms,
config.supported_frame_lengths_ms);
config = CreateConfigWithParameters({{"minptime", "invalid"}});
EXPECT_EQ(default_supported_frame_lengths_ms,
config.supported_frame_lengths_ms);
config = CreateConfigWithParameters({{"ptime", "invalid"}});
EXPECT_EQ(default_supported_frame_lengths_ms,
config.supported_frame_lengths_ms);
}
TEST(AudioEncoderOpusTest, GetFrameLenghtRange) {
AudioEncoderOpusConfig config =
CreateConfigWithParameters({{"maxptime", "10"}, {"ptime", "10"}});
std::unique_ptr<AudioEncoder> encoder =
AudioEncoderOpus::MakeAudioEncoder(config, kDefaultOpusPayloadType);
auto ptime = webrtc::TimeDelta::Millis(10);
absl::optional<std::pair<webrtc::TimeDelta, webrtc::TimeDelta>> range = {
{ptime, ptime}};
EXPECT_EQ(encoder->GetFrameLengthRange(), range);
}
// Test that bitrate will be overridden by the "maxaveragebitrate" parameter.
// Also test that the "maxaveragebitrate" can't be set to values outside the
// range of 6000 and 510000
TEST(AudioEncoderOpusTest, SetSendCodecOpusMaxAverageBitrate) {
// Ignore if less than 6000.
const auto config1 = AudioEncoderOpus::SdpToConfig(
{"opus", 48000, 2, {{"maxaveragebitrate", "5999"}}});
EXPECT_EQ(6000, config1->bitrate_bps);
// Ignore if larger than 510000.
const auto config2 = AudioEncoderOpus::SdpToConfig(
{"opus", 48000, 2, {{"maxaveragebitrate", "510001"}}});
EXPECT_EQ(510000, config2->bitrate_bps);
const auto config3 = AudioEncoderOpus::SdpToConfig(
{"opus", 48000, 2, {{"maxaveragebitrate", "200000"}}});
EXPECT_EQ(200000, config3->bitrate_bps);
}
// Test maxplaybackrate <= 8000 triggers Opus narrow band mode.
TEST(AudioEncoderOpusTest, SetMaxPlaybackRateNb) {
auto config = CreateConfigWithParameters({{"maxplaybackrate", "8000"}});
EXPECT_EQ(8000, config.max_playback_rate_hz);
EXPECT_EQ(12000, config.bitrate_bps);
config = CreateConfigWithParameters(
{{"maxplaybackrate", "8000"}, {"stereo", "1"}});
EXPECT_EQ(8000, config.max_playback_rate_hz);
EXPECT_EQ(24000, config.bitrate_bps);
}
// Test 8000 < maxplaybackrate <= 12000 triggers Opus medium band mode.
TEST(AudioEncoderOpusTest, SetMaxPlaybackRateMb) {
auto config = CreateConfigWithParameters({{"maxplaybackrate", "8001"}});
EXPECT_EQ(8001, config.max_playback_rate_hz);
EXPECT_EQ(20000, config.bitrate_bps);
config = CreateConfigWithParameters(
{{"maxplaybackrate", "8001"}, {"stereo", "1"}});
EXPECT_EQ(8001, config.max_playback_rate_hz);
EXPECT_EQ(40000, config.bitrate_bps);
}
// Test 12000 < maxplaybackrate <= 16000 triggers Opus wide band mode.
TEST(AudioEncoderOpusTest, SetMaxPlaybackRateWb) {
auto config = CreateConfigWithParameters({{"maxplaybackrate", "12001"}});
EXPECT_EQ(12001, config.max_playback_rate_hz);
EXPECT_EQ(20000, config.bitrate_bps);
config = CreateConfigWithParameters(
{{"maxplaybackrate", "12001"}, {"stereo", "1"}});
EXPECT_EQ(12001, config.max_playback_rate_hz);
EXPECT_EQ(40000, config.bitrate_bps);
}
// Test 16000 < maxplaybackrate <= 24000 triggers Opus super wide band mode.
TEST(AudioEncoderOpusTest, SetMaxPlaybackRateSwb) {
auto config = CreateConfigWithParameters({{"maxplaybackrate", "16001"}});
EXPECT_EQ(16001, config.max_playback_rate_hz);
EXPECT_EQ(32000, config.bitrate_bps);
config = CreateConfigWithParameters(
{{"maxplaybackrate", "16001"}, {"stereo", "1"}});
EXPECT_EQ(16001, config.max_playback_rate_hz);
EXPECT_EQ(64000, config.bitrate_bps);
}
// Test 24000 < maxplaybackrate triggers Opus full band mode.
TEST(AudioEncoderOpusTest, SetMaxPlaybackRateFb) {
auto config = CreateConfigWithParameters({{"maxplaybackrate", "24001"}});
EXPECT_EQ(24001, config.max_playback_rate_hz);
EXPECT_EQ(32000, config.bitrate_bps);
config = CreateConfigWithParameters(
{{"maxplaybackrate", "24001"}, {"stereo", "1"}});
EXPECT_EQ(24001, config.max_playback_rate_hz);
EXPECT_EQ(64000, config.bitrate_bps);
}
TEST_P(AudioEncoderOpusTest, OpusFlagDtxAsNonSpeech) {
// Create encoder with DTX enabled.
AudioEncoderOpusConfig config;
config.dtx_enabled = true;
config.sample_rate_hz = sample_rate_hz_;
constexpr int payload_type = 17;
const auto encoder = AudioEncoderOpus::MakeAudioEncoder(config, payload_type);
// Open file containing speech and silence.
const std::string kInputFileName =
webrtc::test::ResourcePath("audio_coding/testfile32kHz", "pcm");
test::AudioLoop audio_loop;
// Use the file as if it were sampled at our desired input rate.
const size_t max_loop_length_samples =
sample_rate_hz_ * 10; // Max 10 second loop.
const size_t input_block_size_samples =
10 * sample_rate_hz_ / 1000; // 10 ms.
EXPECT_TRUE(audio_loop.Init(kInputFileName, max_loop_length_samples,
input_block_size_samples));
// Encode.
AudioEncoder::EncodedInfo info;
rtc::Buffer encoded(500);
int nonspeech_frames = 0;
int max_nonspeech_frames = 0;
int dtx_frames = 0;
int max_dtx_frames = 0;
uint32_t rtp_timestamp = 0u;
for (size_t i = 0; i < 500; ++i) {
encoded.Clear();
// Every second call to the encoder will generate an Opus packet.
for (int j = 0; j < 2; j++) {
info =
encoder->Encode(rtp_timestamp, audio_loop.GetNextBlock(), &encoded);
rtp_timestamp += input_block_size_samples;
}
// Bookkeeping of number of DTX frames.
if (info.encoded_bytes <= 2) {
++dtx_frames;
} else {
if (dtx_frames > max_dtx_frames)
max_dtx_frames = dtx_frames;
dtx_frames = 0;
}
// Bookkeeping of number of non-speech frames.
if (info.speech == 0) {
++nonspeech_frames;
} else {
if (nonspeech_frames > max_nonspeech_frames)
max_nonspeech_frames = nonspeech_frames;
nonspeech_frames = 0;
}
}
// Maximum number of consecutive non-speech packets should exceed 15.
EXPECT_GT(max_nonspeech_frames, 15);
}
TEST(AudioEncoderOpusTest, OpusDtxFilteringHighEnergyRefreshPackets) {
test::ScopedFieldTrials override_field_trials(
"WebRTC-Audio-OpusAvoidNoisePumpingDuringDtx/Enabled/");
const std::string kInputFileName =
webrtc::test::ResourcePath("audio_coding/testfile16kHz", "pcm");
constexpr int kSampleRateHz = 16000;
AudioEncoderOpusConfig config;
config.dtx_enabled = true;
config.sample_rate_hz = kSampleRateHz;
constexpr int payload_type = 17;
const auto encoder = AudioEncoderOpus::MakeAudioEncoder(config, payload_type);
test::AudioLoop audio_loop;
constexpr size_t kMaxLoopLengthSaples = kSampleRateHz * 11.6f;
constexpr size_t kInputBlockSizeSamples = kSampleRateHz / 100;
EXPECT_TRUE(audio_loop.Init(kInputFileName, kMaxLoopLengthSaples,
kInputBlockSizeSamples));
AudioEncoder::EncodedInfo info;
rtc::Buffer encoded(500);
// Encode the audio file and store the last part that corresponds to silence.
constexpr size_t kSilenceDurationSamples = kSampleRateHz * 0.2f;
std::array<int16_t, kSilenceDurationSamples> silence;
uint32_t rtp_timestamp = 0;
bool last_packet_dtx_frame = false;
bool opus_entered_dtx = false;
bool silence_filled = false;
size_t timestamp_start_silence = 0;
while (!silence_filled && rtp_timestamp < kMaxLoopLengthSaples) {
encoded.Clear();
// Every second call to the encoder will generate an Opus packet.
for (int j = 0; j < 2; j++) {
auto next_frame = audio_loop.GetNextBlock();
info = encoder->Encode(rtp_timestamp, next_frame, &encoded);
if (opus_entered_dtx) {
size_t silence_frame_start = rtp_timestamp - timestamp_start_silence;
silence_filled = silence_frame_start >= kSilenceDurationSamples;
if (!silence_filled) {
std::copy(next_frame.begin(), next_frame.end(),
silence.begin() + silence_frame_start);
}
}
rtp_timestamp += kInputBlockSizeSamples;
}
EXPECT_TRUE(info.encoded_bytes > 0 || last_packet_dtx_frame);
last_packet_dtx_frame = info.encoded_bytes > 0 ? info.encoded_bytes <= 2
: last_packet_dtx_frame;
if (info.encoded_bytes <= 2 && !opus_entered_dtx) {
timestamp_start_silence = rtp_timestamp;
}
opus_entered_dtx = info.encoded_bytes <= 2;
}
EXPECT_TRUE(silence_filled);
// The copied 200 ms of silence is used for creating 6 bursts that are fed to
// the encoder, the first three ones with a larger energy and the last three
// with a lower energy. This test verifies that the encoder just sends refresh
// DTX packets during the last bursts.
int number_non_empty_packets_during_increase = 0;
int number_non_empty_packets_during_decrease = 0;
for (size_t burst = 0; burst < 6; ++burst) {
uint32_t rtp_timestamp_start = rtp_timestamp;
const bool increase_noise = burst < 3;
const float gain = increase_noise ? 1.4f : 0.0f;
while (rtp_timestamp < rtp_timestamp_start + kSilenceDurationSamples) {
encoded.Clear();
// Every second call to the encoder will generate an Opus packet.
for (int j = 0; j < 2; j++) {
std::array<int16_t, kInputBlockSizeSamples> silence_frame;
size_t silence_frame_start = rtp_timestamp - rtp_timestamp_start;
std::transform(
silence.begin() + silence_frame_start,
silence.begin() + silence_frame_start + kInputBlockSizeSamples,
silence_frame.begin(), [gain](float s) { return gain * s; });
info = encoder->Encode(rtp_timestamp, silence_frame, &encoded);
rtp_timestamp += kInputBlockSizeSamples;
}
EXPECT_TRUE(info.encoded_bytes > 0 || last_packet_dtx_frame);
last_packet_dtx_frame = info.encoded_bytes > 0 ? info.encoded_bytes <= 2
: last_packet_dtx_frame;
// Tracking the number of non empty packets.
if (increase_noise && info.encoded_bytes > 2) {
number_non_empty_packets_during_increase++;
}
if (!increase_noise && info.encoded_bytes > 2) {
number_non_empty_packets_during_decrease++;
}
}
}
// Check that the refresh DTX packets are just sent during the decrease energy
// region.
EXPECT_EQ(number_non_empty_packets_during_increase, 0);
EXPECT_GT(number_non_empty_packets_during_decrease, 0);
}
} // namespace webrtc