gecko-dev/dom/media/gtest/TestAudioTrackEncoder.cpp

283 строки
9.5 KiB
C++

/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "gtest/gtest.h"
#include "OpusTrackEncoder.h"
#include "AudioGenerator.h"
#include "AudioSampleFormat.h"
using namespace mozilla;
class TestOpusTrackEncoder : public OpusTrackEncoder {
public:
explicit TestOpusTrackEncoder(TrackRate aTrackRate)
: OpusTrackEncoder(aTrackRate) {}
// Return true if it has successfully initialized the Opus encoder.
bool TestOpusRawCreation(int aChannels) {
if (Init(aChannels) == NS_OK) {
if (IsInitialized()) {
return true;
}
}
return false;
}
};
static bool TestOpusInit(int aChannels, TrackRate aSamplingRate) {
TestOpusTrackEncoder encoder(aSamplingRate);
return encoder.TestOpusRawCreation(aChannels);
}
TEST(OpusAudioTrackEncoder, InitRaw)
{
// Expect false with 0 or negative channels of input signal.
EXPECT_FALSE(TestOpusInit(0, 16000));
EXPECT_FALSE(TestOpusInit(-1, 16000));
// The Opus format supports up to 8 channels, and supports multitrack audio up
// to 255 channels, but the current implementation supports only mono and
// stereo, and downmixes any more than that.
// Expect false with channels of input signal exceed the max supported number.
EXPECT_FALSE(TestOpusInit(8 + 1, 16000));
// Should accept channels within valid range.
for (int i = 1; i <= 8; i++) {
EXPECT_TRUE(TestOpusInit(i, 16000));
}
// Expect false with 0 or negative sampling rate of input signal.
EXPECT_FALSE(TestOpusInit(1, 0));
EXPECT_FALSE(TestOpusInit(1, -1));
// Verify sample rate bounds checking.
EXPECT_FALSE(TestOpusInit(2, 2000));
EXPECT_FALSE(TestOpusInit(2, 4000));
EXPECT_FALSE(TestOpusInit(2, 7999));
EXPECT_TRUE(TestOpusInit(2, 8000));
EXPECT_TRUE(TestOpusInit(2, 192000));
EXPECT_FALSE(TestOpusInit(2, 192001));
EXPECT_FALSE(TestOpusInit(2, 200000));
}
TEST(OpusAudioTrackEncoder, Init)
{
{
// The encoder does not normally recieve enough info from null data to
// init. However, multiple attempts to do so, with sufficiently long
// duration segments, should result in a default-init. The first attempt
// should never do this though, even if the duration is long:
OpusTrackEncoder encoder(48000);
AudioSegment segment;
segment.AppendNullData(48000 * 100);
encoder.TryInit(segment, segment.GetDuration());
EXPECT_FALSE(encoder.IsInitialized());
// Multiple init attempts should result in best effort init:
encoder.TryInit(segment, segment.GetDuration());
EXPECT_TRUE(encoder.IsInitialized());
}
{
// For non-null segments we should init immediately
OpusTrackEncoder encoder(48000);
AudioSegment segment;
AudioGenerator<AudioDataValue> generator(2, 48000);
generator.Generate(segment, 1);
encoder.TryInit(segment, segment.GetDuration());
EXPECT_TRUE(encoder.IsInitialized());
}
{
// Test low sample rate bound
OpusTrackEncoder encoder(7999);
AudioSegment segment;
AudioGenerator<AudioDataValue> generator(2, 7999);
generator.Generate(segment, 1);
encoder.TryInit(segment, segment.GetDuration());
EXPECT_FALSE(encoder.IsInitialized());
}
{
// Test low sample rate bound
OpusTrackEncoder encoder(8000);
AudioSegment segment;
AudioGenerator<AudioDataValue> generator(2, 8000);
generator.Generate(segment, 1);
encoder.TryInit(segment, segment.GetDuration());
EXPECT_TRUE(encoder.IsInitialized());
}
{
// Test high sample rate bound
OpusTrackEncoder encoder(192001);
AudioSegment segment;
AudioGenerator<AudioDataValue> generator(2, 192001);
generator.Generate(segment, 1);
encoder.TryInit(segment, segment.GetDuration());
EXPECT_FALSE(encoder.IsInitialized());
}
{
// Test high sample rate bound
OpusTrackEncoder encoder(192000);
AudioSegment segment;
AudioGenerator<AudioDataValue> generator(2, 192000);
generator.Generate(segment, 1);
encoder.TryInit(segment, segment.GetDuration());
EXPECT_TRUE(encoder.IsInitialized());
}
{
// Test that it takes 10s to trigger default-init.
OpusTrackEncoder encoder(48000);
AudioSegment longSegment;
longSegment.AppendNullData(48000 * 10 - 1);
AudioSegment shortSegment;
shortSegment.AppendNullData(1);
encoder.TryInit(longSegment, longSegment.GetDuration());
EXPECT_FALSE(encoder.IsInitialized());
encoder.TryInit(shortSegment, shortSegment.GetDuration());
EXPECT_FALSE(encoder.IsInitialized());
encoder.TryInit(shortSegment, shortSegment.GetDuration());
EXPECT_TRUE(encoder.IsInitialized());
}
}
static int TestOpusResampler(TrackRate aSamplingRate) {
OpusTrackEncoder encoder(aSamplingRate);
return encoder.mOutputSampleRate;
}
TEST(OpusAudioTrackEncoder, Resample)
{
// Sampling rates of data to be fed to Opus encoder, should remain unchanged
// if it is one of Opus supported rates (8000, 12000, 16000, 24000 and 48000
// (kHz)) at initialization.
EXPECT_TRUE(TestOpusResampler(8000) == 8000);
EXPECT_TRUE(TestOpusResampler(12000) == 12000);
EXPECT_TRUE(TestOpusResampler(16000) == 16000);
EXPECT_TRUE(TestOpusResampler(24000) == 24000);
EXPECT_TRUE(TestOpusResampler(48000) == 48000);
// Otherwise, it should be resampled to 48kHz by resampler.
EXPECT_TRUE(TestOpusResampler(9600) == 48000);
EXPECT_TRUE(TestOpusResampler(44100) == 48000);
}
TEST(OpusAudioTrackEncoder, FetchMetadata)
{
const int32_t channels = 1;
const TrackRate sampleRate = 44100;
TestOpusTrackEncoder encoder(sampleRate);
EXPECT_TRUE(encoder.TestOpusRawCreation(channels));
RefPtr<TrackMetadataBase> metadata = encoder.GetMetadata();
ASSERT_EQ(TrackMetadataBase::METADATA_OPUS, metadata->GetKind());
RefPtr<OpusMetadata> opusMeta = static_cast<OpusMetadata*>(metadata.get());
EXPECT_EQ(channels, opusMeta->mChannels);
EXPECT_EQ(sampleRate, opusMeta->mSamplingFrequency);
}
TEST(OpusAudioTrackEncoder, FrameEncode)
{
const int32_t channels = 1;
const TrackRate sampleRate = 44100;
TestOpusTrackEncoder encoder(sampleRate);
EXPECT_TRUE(encoder.TestOpusRawCreation(channels));
// Generate five seconds of raw audio data.
AudioGenerator<AudioDataValue> generator(channels, sampleRate);
AudioSegment segment;
const int32_t samples = sampleRate * 5;
generator.Generate(segment, samples);
encoder.AppendAudioSegment(std::move(segment));
encoder.NotifyEndOfStream();
nsTArray<RefPtr<EncodedFrame>> frames;
EXPECT_TRUE(NS_SUCCEEDED(encoder.GetEncodedTrack(frames)));
// Verify that encoded data is 5 seconds long.
uint64_t totalDuration = 0;
for (auto& frame : frames) {
totalDuration += frame->mDuration;
}
// 44100 as used above gets resampled to 48000 for opus.
const uint64_t five = 48000 * 5;
EXPECT_EQ(five + encoder.GetLookahead(), totalDuration);
}
TEST(OpusAudioTrackEncoder, DefaultInitDuration)
{
const TrackRate rate = 44100;
OpusTrackEncoder encoder(rate);
AudioGenerator<AudioDataValue> generator(2, rate);
AudioSegment segment;
// 15 seconds should trigger the default-init rate.
// The default-init timeout is evaluated once per chunk, so keep chunks
// reasonably short.
for (int i = 0; i < 150; ++i) {
generator.Generate(segment, rate / 10);
}
encoder.AppendAudioSegment(std::move(segment));
encoder.NotifyEndOfStream();
nsTArray<RefPtr<EncodedFrame>> frames;
EXPECT_TRUE(NS_SUCCEEDED(encoder.GetEncodedTrack(frames)));
// Verify that encoded data is 15 seconds long.
uint64_t totalDuration = 0;
for (auto& frame : frames) {
totalDuration += frame->mDuration;
}
// 44100 as used above gets resampled to 48000 for opus.
const uint64_t fifteen = 48000 * 15;
EXPECT_EQ(totalDuration, fifteen + encoder.GetLookahead());
}
uint64_t TestSampleRate(TrackRate aSampleRate, uint64_t aInputFrames) {
OpusTrackEncoder encoder(aSampleRate);
AudioGenerator<AudioDataValue> generator(2, aSampleRate);
AudioSegment segment;
const uint64_t chunkSize = aSampleRate / 10;
const uint64_t chunks = aInputFrames / chunkSize;
// 15 seconds should trigger the default-init rate.
// The default-init timeout is evaluated once per chunk, so keep chunks
// reasonably short.
for (size_t i = 0; i < chunks; ++i) {
generator.Generate(segment, chunkSize);
}
generator.Generate(segment, aInputFrames % chunks);
encoder.AppendAudioSegment(std::move(segment));
encoder.NotifyEndOfStream();
nsTArray<RefPtr<EncodedFrame>> frames;
EXPECT_TRUE(NS_SUCCEEDED(encoder.GetEncodedTrack(frames)));
// Verify that encoded data is 15 seconds long.
uint64_t totalDuration = 0;
for (auto& frame : frames) {
totalDuration += frame->mDuration;
}
return totalDuration - encoder.GetLookahead();
}
TEST(OpusAudioTrackEncoder, DurationSampleRates)
{
// Factors of 48k
EXPECT_EQ(TestSampleRate(48000, 48000 * 3 / 2), 48000U * 3 / 2);
EXPECT_EQ(TestSampleRate(24000, 24000 * 3 / 2), 48000U * 3 / 2);
EXPECT_EQ(TestSampleRate(16000, 16000 * 3 / 2), 48000U * 3 / 2);
EXPECT_EQ(TestSampleRate(12000, 12000 * 3 / 2), 48000U * 3 / 2);
EXPECT_EQ(TestSampleRate(8000, 8000 * 3 / 2), 48000U * 3 / 2);
// Non-factors of 48k, resampled
EXPECT_EQ(TestSampleRate(44100, 44100 * 3 / 2), 48000U * 3 / 2);
EXPECT_EQ(TestSampleRate(32000, 32000 * 3 / 2), 48000U * 3 / 2);
EXPECT_EQ(TestSampleRate(96000, 96000 * 3 / 2), 48000U * 3 / 2);
EXPECT_EQ(TestSampleRate(33330, 33330 * 3 / 2), 48000U * 3 / 2);
}