Source code

Revision control

Copy as Markdown

Other Tools

/*
* Copyright 2018 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/audio_frame.h"
#include <stdint.h>
#include <string.h> // memcmp
#include "test/gtest.h"
namespace webrtc {
namespace {
bool AllSamplesAre(int16_t sample, InterleavedView<const int16_t> samples) {
for (const auto s : samples) {
if (s != sample) {
return false;
}
}
return true;
}
bool AllSamplesAre(int16_t sample, const AudioFrame& frame) {
return AllSamplesAre(sample, frame.data_view());
}
// Checks the values of samples in the AudioFrame buffer, regardless of whether
// they're valid or not, and disregard the `muted()` state of the frame.
// I.e. use `max_16bit_samples()` instead of `data_view().size()`
bool AllBufferSamplesAre(int16_t sample, const AudioFrame& frame) {
auto view = frame.data_view();
RTC_DCHECK(!view.empty());
const int16_t* data = &view.data()[0];
for (size_t i = 0; i < frame.max_16bit_samples(); ++i) {
if (data[i] != sample) {
return false;
}
}
return true;
}
constexpr uint32_t kTimestamp = 27;
constexpr int kSampleRateHz = 16000;
constexpr size_t kNumChannelsMono = 1;
constexpr size_t kNumChannelsStereo = 2;
constexpr size_t kNumChannels5_1 = 6;
constexpr size_t kSamplesPerChannel = kSampleRateHz / 100;
} // namespace
TEST(AudioFrameTest, FrameStartsZeroedAndMuted) {
AudioFrame frame;
EXPECT_TRUE(frame.muted());
EXPECT_TRUE(frame.data_view().empty());
EXPECT_TRUE(AllSamplesAre(0, frame));
}
// TODO: b/335805780 - Delete test when `mutable_data()` returns ArrayView.
TEST(AudioFrameTest, UnmutedFrameIsInitiallyZeroedLegacy) {
AudioFrame frame(kSampleRateHz, kNumChannelsMono, CHANNEL_LAYOUT_NONE);
frame.mutable_data();
EXPECT_FALSE(frame.muted());
EXPECT_TRUE(AllSamplesAre(0, frame));
EXPECT_TRUE(AllBufferSamplesAre(0, frame));
}
TEST(AudioFrameTest, UnmutedFrameIsInitiallyZeroed) {
AudioFrame frame;
auto data = frame.mutable_data(kSamplesPerChannel, kNumChannelsMono);
EXPECT_FALSE(frame.muted());
EXPECT_TRUE(IsMono(data));
EXPECT_EQ(frame.data_view().size(), kSamplesPerChannel);
EXPECT_EQ(SamplesPerChannel(data), kSamplesPerChannel);
EXPECT_TRUE(AllSamplesAre(0, frame));
}
TEST(AudioFrameTest, MutedFrameBufferIsZeroed) {
AudioFrame frame;
int16_t* frame_data =
frame.mutable_data(kSamplesPerChannel, kNumChannelsMono).begin();
EXPECT_FALSE(frame.muted());
// Fill the reserved buffer with non-zero data.
for (size_t i = 0; i < frame.max_16bit_samples(); i++) {
frame_data[i] = 17;
}
ASSERT_TRUE(AllSamplesAre(17, frame));
ASSERT_TRUE(AllBufferSamplesAre(17, frame));
frame.Mute();
EXPECT_TRUE(frame.muted());
EXPECT_TRUE(AllSamplesAre(0, frame));
ASSERT_TRUE(AllBufferSamplesAre(0, frame));
}
TEST(AudioFrameTest, UpdateFrameMono) {
AudioFrame frame;
int16_t samples[kNumChannelsMono * kSamplesPerChannel] = {17};
frame.UpdateFrame(kTimestamp, samples, kSamplesPerChannel, kSampleRateHz,
AudioFrame::kPLC, AudioFrame::kVadActive, kNumChannelsMono);
EXPECT_EQ(kTimestamp, frame.timestamp_);
EXPECT_EQ(kSamplesPerChannel, frame.samples_per_channel());
EXPECT_EQ(kSampleRateHz, frame.sample_rate_hz());
EXPECT_EQ(AudioFrame::kPLC, frame.speech_type_);
EXPECT_EQ(AudioFrame::kVadActive, frame.vad_activity_);
EXPECT_EQ(kNumChannelsMono, frame.num_channels());
EXPECT_EQ(CHANNEL_LAYOUT_MONO, frame.channel_layout());
EXPECT_FALSE(frame.muted());
EXPECT_EQ(0, memcmp(samples, frame.data(), sizeof(samples)));
frame.UpdateFrame(kTimestamp, nullptr /* data*/, kSamplesPerChannel,
kSampleRateHz, AudioFrame::kPLC, AudioFrame::kVadActive,
kNumChannelsMono);
EXPECT_TRUE(frame.muted());
EXPECT_TRUE(AllSamplesAre(0, frame));
}
TEST(AudioFrameTest, UpdateFrameMultiChannel) {
AudioFrame frame;
frame.UpdateFrame(kTimestamp, nullptr /* data */, kSamplesPerChannel,
kSampleRateHz, AudioFrame::kPLC, AudioFrame::kVadActive,
kNumChannelsStereo);
EXPECT_EQ(kSamplesPerChannel, frame.samples_per_channel());
EXPECT_EQ(kNumChannelsStereo, frame.num_channels());
EXPECT_EQ(CHANNEL_LAYOUT_STEREO, frame.channel_layout());
EXPECT_TRUE(frame.muted());
// Initialize the frame with valid `kNumChannels5_1` data to make sure we
// get an unmuted frame with valid samples.
int16_t samples[kSamplesPerChannel * kNumChannels5_1] = {17};
frame.UpdateFrame(kTimestamp, samples /* data */, kSamplesPerChannel,
kSampleRateHz, AudioFrame::kPLC, AudioFrame::kVadActive,
kNumChannels5_1);
EXPECT_FALSE(frame.muted());
EXPECT_EQ(kSamplesPerChannel, frame.samples_per_channel());
EXPECT_EQ(kSamplesPerChannel * kNumChannels5_1, frame.data_view().size());
EXPECT_EQ(kNumChannels5_1, frame.num_channels());
EXPECT_EQ(CHANNEL_LAYOUT_5_1, frame.channel_layout());
}
TEST(AudioFrameTest, CopyFrom) {
AudioFrame frame1;
AudioFrame frame2;
int16_t samples[kNumChannelsMono * kSamplesPerChannel] = {17};
frame2.UpdateFrame(kTimestamp, samples, kSamplesPerChannel, kSampleRateHz,
AudioFrame::kPLC, AudioFrame::kVadActive,
kNumChannelsMono);
frame1.CopyFrom(frame2);
EXPECT_EQ(frame2.timestamp_, frame1.timestamp_);
EXPECT_EQ(frame2.samples_per_channel_, frame1.samples_per_channel_);
EXPECT_EQ(frame2.sample_rate_hz_, frame1.sample_rate_hz_);
EXPECT_EQ(frame2.speech_type_, frame1.speech_type_);
EXPECT_EQ(frame2.vad_activity_, frame1.vad_activity_);
EXPECT_EQ(frame2.num_channels_, frame1.num_channels_);
EXPECT_EQ(frame2.data_view().size(), frame1.data_view().size());
EXPECT_EQ(frame2.muted(), frame1.muted());
EXPECT_EQ(0, memcmp(frame2.data(), frame1.data(), sizeof(samples)));
frame2.UpdateFrame(kTimestamp, nullptr /* data */, kSamplesPerChannel,
kSampleRateHz, AudioFrame::kPLC, AudioFrame::kVadActive,
kNumChannelsMono);
frame1.CopyFrom(frame2);
EXPECT_EQ(frame2.muted(), frame1.muted());
EXPECT_EQ(0, memcmp(frame2.data(), frame1.data(), sizeof(samples)));
}
} // namespace webrtc