/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */ /* vim:set ts=2 sw=2 sts=2 et cindent: */ /* 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 #include #include "prlog.h" #include "prdtoa.h" #include "AudioStream.h" #include "nsAlgorithm.h" #include "VideoUtils.h" #include "mozilla/Monitor.h" #include "mozilla/Mutex.h" extern "C" { #include "sydneyaudio/sydney_audio.h" } #include "mozilla/Preferences.h" #if defined(MOZ_CUBEB) #include "nsAutoRef.h" #include "cubeb/cubeb.h" template <> class nsAutoRefTraits : public nsPointerRefTraits { public: static void Release(cubeb_stream* aStream) { cubeb_stream_destroy(aStream); } }; #endif namespace mozilla { #if defined(XP_MACOSX) #define SA_PER_STREAM_VOLUME 1 #endif #ifdef PR_LOGGING PRLogModuleInfo* gAudioStreamLog = nullptr; #endif static const uint32_t FAKE_BUFFER_SIZE = 176400; // Number of milliseconds per second. static const int64_t MS_PER_S = 1000; class NativeAudioStream : public AudioStream { public: ~NativeAudioStream(); NativeAudioStream(); nsresult Init(int32_t aNumChannels, int32_t aRate, const dom::AudioChannelType aAudioChannelType); void Shutdown(); nsresult Write(const AudioDataValue* aBuf, uint32_t aFrames); uint32_t Available(); void SetVolume(double aVolume); void Drain(); void Pause(); void Resume(); int64_t GetPosition(); int64_t GetPositionInFrames(); int64_t GetPositionInFramesInternal(); bool IsPaused(); int32_t GetMinWriteSize(); private: int32_t WriteToBackend(const float* aBuffer, uint32_t aFrames); int32_t WriteToBackend(const short* aBuffer, uint32_t aFrames); double mVolume; void* mAudioHandle; // True if this audio stream is paused. bool mPaused; // True if this stream has encountered an error. bool mInError; }; #define PREF_VOLUME_SCALE "media.volume_scale" #define PREF_USE_CUBEB "media.use_cubeb" #define PREF_CUBEB_LATENCY "media.cubeb_latency_ms" static Mutex* gAudioPrefsLock = nullptr; static double gVolumeScale; static bool gUseCubeb; static uint32_t gCubebLatency; static int PrefChanged(const char* aPref, void* aClosure) { if (strcmp(aPref, PREF_VOLUME_SCALE) == 0) { nsAdoptingString value = Preferences::GetString(aPref); MutexAutoLock lock(*gAudioPrefsLock); if (value.IsEmpty()) { gVolumeScale = 1.0; } else { NS_ConvertUTF16toUTF8 utf8(value); gVolumeScale = NS_MAX(0, PR_strtod(utf8.get(), nullptr)); } } else if (strcmp(aPref, PREF_USE_CUBEB) == 0) { #ifdef MOZ_WIDGET_GONK bool value = Preferences::GetBool(aPref, false); #else bool value = Preferences::GetBool(aPref, true); #endif MutexAutoLock lock(*gAudioPrefsLock); gUseCubeb = value; } else if (strcmp(aPref, PREF_CUBEB_LATENCY) == 0) { // Arbitrary default stream latency of 100ms. The higher this // value, the longer stream volume changes will take to become // audible. uint32_t value = Preferences::GetUint(aPref, 100); MutexAutoLock lock(*gAudioPrefsLock); gCubebLatency = NS_MIN(NS_MAX(value, 20), 1000); } return 0; } static double GetVolumeScale() { MutexAutoLock lock(*gAudioPrefsLock); return gVolumeScale; } #if defined(MOZ_CUBEB) static bool GetUseCubeb() { MutexAutoLock lock(*gAudioPrefsLock); return gUseCubeb; } static cubeb* gCubebContext; static cubeb* GetCubebContext() { MutexAutoLock lock(*gAudioPrefsLock); if (gCubebContext || cubeb_init(&gCubebContext, "AudioStream") == CUBEB_OK) { return gCubebContext; } NS_WARNING("cubeb_init failed"); return nullptr; } static uint32_t GetCubebLatency() { MutexAutoLock lock(*gAudioPrefsLock); return gCubebLatency; } #endif static sa_stream_type_t ConvertChannelToSAType(dom::AudioChannelType aType) { switch(aType) { case dom::AUDIO_CHANNEL_NORMAL: return SA_STREAM_TYPE_SYSTEM; case dom::AUDIO_CHANNEL_CONTENT: return SA_STREAM_TYPE_MUSIC; case dom::AUDIO_CHANNEL_NOTIFICATION: return SA_STREAM_TYPE_NOTIFICATION; case dom::AUDIO_CHANNEL_ALARM: return SA_STREAM_TYPE_ALARM; case dom::AUDIO_CHANNEL_TELEPHONY: return SA_STREAM_TYPE_VOICE_CALL; case dom::AUDIO_CHANNEL_RINGER: return SA_STREAM_TYPE_RING; case dom::AUDIO_CHANNEL_PUBLICNOTIFICATION: return SA_STREAM_TYPE_ENFORCED_AUDIBLE; default: NS_ERROR("The value of AudioChannelType is invalid"); return SA_STREAM_TYPE_MAX; } } AudioStream::AudioStream() : mInRate(0), mOutRate(0), mChannels(0), mAudioClock(this) {} void AudioStream::InitLibrary() { #ifdef PR_LOGGING gAudioStreamLog = PR_NewLogModule("AudioStream"); #endif gAudioPrefsLock = new Mutex("AudioStream::gAudioPrefsLock"); PrefChanged(PREF_VOLUME_SCALE, nullptr); Preferences::RegisterCallback(PrefChanged, PREF_VOLUME_SCALE); #if defined(MOZ_CUBEB) PrefChanged(PREF_USE_CUBEB, nullptr); Preferences::RegisterCallback(PrefChanged, PREF_USE_CUBEB); PrefChanged(PREF_CUBEB_LATENCY, nullptr); Preferences::RegisterCallback(PrefChanged, PREF_CUBEB_LATENCY); #endif } void AudioStream::ShutdownLibrary() { Preferences::UnregisterCallback(PrefChanged, PREF_VOLUME_SCALE); #if defined(MOZ_CUBEB) Preferences::UnregisterCallback(PrefChanged, PREF_USE_CUBEB); #endif delete gAudioPrefsLock; gAudioPrefsLock = nullptr; #if defined(MOZ_CUBEB) if (gCubebContext) { cubeb_destroy(gCubebContext); gCubebContext = nullptr; } #endif } AudioStream::~AudioStream() { } void AudioStream::EnsureTimeStretcherInitialized() { if (!mTimeStretcher) { mTimeStretcher = new soundtouch::SoundTouch(); mTimeStretcher->setSampleRate(mInRate); mTimeStretcher->setChannels(mChannels); mTimeStretcher->setPitch(1.0); } } nsresult AudioStream::SetPlaybackRate(double aPlaybackRate) { NS_ASSERTION(aPlaybackRate > 0.0, "Can't handle negative or null playbackrate in the AudioStream."); // Avoid instantiating the resampler if we are not changing the playback rate. if (aPlaybackRate == mAudioClock.GetPlaybackRate()) { return NS_OK; } mAudioClock.SetPlaybackRate(aPlaybackRate); mOutRate = mInRate / aPlaybackRate; EnsureTimeStretcherInitialized(); if (mAudioClock.GetPreservesPitch()) { mTimeStretcher->setTempo(aPlaybackRate); mTimeStretcher->setRate(1.0f); } else { mTimeStretcher->setTempo(1.0f); mTimeStretcher->setRate(aPlaybackRate); } return NS_OK; } nsresult AudioStream::SetPreservesPitch(bool aPreservesPitch) { // Avoid instantiating the timestretcher instance if not needed. if (aPreservesPitch == mAudioClock.GetPreservesPitch()) { return NS_OK; } EnsureTimeStretcherInitialized(); if (aPreservesPitch == true) { mTimeStretcher->setTempo(mAudioClock.GetPlaybackRate()); mTimeStretcher->setRate(1.0f); } else { mTimeStretcher->setTempo(1.0f); mTimeStretcher->setRate(mAudioClock.GetPlaybackRate()); } mAudioClock.SetPreservesPitch(aPreservesPitch); return NS_OK; } NativeAudioStream::NativeAudioStream() : mVolume(1.0), mAudioHandle(0), mPaused(false), mInError(false) { } NativeAudioStream::~NativeAudioStream() { Shutdown(); } nsresult NativeAudioStream::Init(int32_t aNumChannels, int32_t aRate, const dom::AudioChannelType aAudioChannelType) { mInRate = mOutRate = aRate; mChannels = aNumChannels; if (sa_stream_create_pcm(reinterpret_cast(&mAudioHandle), NULL, SA_MODE_WRONLY, SA_PCM_FORMAT_S16_NE, aRate, aNumChannels) != SA_SUCCESS) { mAudioHandle = nullptr; mInError = true; PR_LOG(gAudioStreamLog, PR_LOG_ERROR, ("NativeAudioStream: sa_stream_create_pcm error")); return NS_ERROR_FAILURE; } int saError = sa_stream_set_stream_type(static_cast(mAudioHandle), ConvertChannelToSAType(aAudioChannelType)); if (saError != SA_SUCCESS && saError != SA_ERROR_NOT_SUPPORTED) { mAudioHandle = nullptr; mInError = true; PR_LOG(gAudioStreamLog, PR_LOG_ERROR, ("NativeAudioStream: sa_stream_set_stream_type error")); return NS_ERROR_FAILURE; } if (sa_stream_open(static_cast(mAudioHandle)) != SA_SUCCESS) { sa_stream_destroy(static_cast(mAudioHandle)); mAudioHandle = nullptr; mInError = true; PR_LOG(gAudioStreamLog, PR_LOG_ERROR, ("NativeAudioStream: sa_stream_open error")); return NS_ERROR_FAILURE; } mInError = false; mAudioClock.Init(); return NS_OK; } void NativeAudioStream::Shutdown() { if (!mAudioHandle) return; sa_stream_destroy(static_cast(mAudioHandle)); mAudioHandle = nullptr; mInError = true; } int32_t NativeAudioStream::WriteToBackend(const AudioDataValue* aBuffer, uint32_t aSamples) { double scaledVolume = GetVolumeScale() * mVolume; nsAutoArrayPtr outputBuffer(new short[aSamples]); ConvertAudioSamplesWithScale(aBuffer, outputBuffer.get(), aSamples, scaledVolume); if (sa_stream_write(static_cast(mAudioHandle), outputBuffer, aSamples * sizeof(short)) != SA_SUCCESS) { return -1; } mAudioClock.UpdateWritePosition(aSamples / mChannels); return aSamples; } nsresult NativeAudioStream::Write(const AudioDataValue* aBuf, uint32_t aFrames) { NS_ASSERTION(!mPaused, "Don't write audio when paused, you'll block"); if (mInError) return NS_ERROR_FAILURE; uint32_t samples = aFrames * mChannels; int32_t written = -1; if (mInRate != mOutRate) { EnsureTimeStretcherInitialized(); mTimeStretcher->putSamples(aBuf, aFrames); uint32_t numFrames = mTimeStretcher->numSamples(); uint32_t arraySize = numFrames * mChannels * sizeof(AudioDataValue); nsAutoArrayPtr data(new AudioDataValue[arraySize]); uint32_t framesAvailable = mTimeStretcher->receiveSamples(data, numFrames); NS_ASSERTION(mTimeStretcher->numSamples() == 0, "We did not get all the data from the SoundTouch pipeline."); // It is possible to have nothing to write: the data are in the processing // pipeline, and will be written to the backend next time. if (framesAvailable) { written = WriteToBackend(data, framesAvailable * mChannels); } else { written = 0; } } else { written = WriteToBackend(aBuf, samples); } if (written == -1) { PR_LOG(gAudioStreamLog, PR_LOG_ERROR, ("NativeAudioStream: sa_stream_write error")); mInError = true; return NS_ERROR_FAILURE; } return NS_OK; } uint32_t NativeAudioStream::Available() { // If the audio backend failed to open, lie and say we'll accept some // data. if (mInError) return FAKE_BUFFER_SIZE; size_t s = 0; if (sa_stream_get_write_size(static_cast(mAudioHandle), &s) != SA_SUCCESS) return 0; return s / mChannels / sizeof(short); } void NativeAudioStream::SetVolume(double aVolume) { NS_ASSERTION(aVolume >= 0.0 && aVolume <= 1.0, "Invalid volume"); #if defined(SA_PER_STREAM_VOLUME) if (sa_stream_set_volume_abs(static_cast(mAudioHandle), aVolume) != SA_SUCCESS) { PR_LOG(gAudioStreamLog, PR_LOG_ERROR, ("NativeAudioStream: sa_stream_set_volume_abs error")); mInError = true; } #else mVolume = aVolume; #endif } void NativeAudioStream::Drain() { NS_ASSERTION(!mPaused, "Don't drain audio when paused, it won't finish!"); // Write all the frames still in the time stretcher pipeline. if (mTimeStretcher) { uint32_t numFrames = mTimeStretcher->numSamples(); uint32_t arraySize = numFrames * mChannels * sizeof(AudioDataValue); nsAutoArrayPtr data(new AudioDataValue[arraySize]); uint32_t framesAvailable = mTimeStretcher->receiveSamples(data, numFrames); int32_t written = 0; if (framesAvailable) { written = WriteToBackend(data, framesAvailable * mChannels); } if (written == -1) { PR_LOG(gAudioStreamLog, PR_LOG_ERROR, ("NativeAudioStream: sa_stream_write error")); mInError = true; } NS_ASSERTION(mTimeStretcher->numSamples() == 0, "We did not get all the data from the SoundTouch pipeline."); } if (mInError) return; int r = sa_stream_drain(static_cast(mAudioHandle)); if (r != SA_SUCCESS && r != SA_ERROR_INVALID) { PR_LOG(gAudioStreamLog, PR_LOG_ERROR, ("NativeAudioStream: sa_stream_drain error")); mInError = true; } } void NativeAudioStream::Pause() { if (mInError) return; mPaused = true; sa_stream_pause(static_cast(mAudioHandle)); } void NativeAudioStream::Resume() { if (mInError) return; mPaused = false; sa_stream_resume(static_cast(mAudioHandle)); } int64_t NativeAudioStream::GetPosition() { return mAudioClock.GetPosition(); } int64_t NativeAudioStream::GetPositionInFrames() { return mAudioClock.GetPositionInFrames(); } int64_t NativeAudioStream::GetPositionInFramesInternal() { if (mInError) { return -1; } sa_position_t positionType = SA_POSITION_WRITE_SOFTWARE; #if defined(XP_WIN) positionType = SA_POSITION_WRITE_HARDWARE; #endif int64_t position = 0; if (sa_stream_get_position(static_cast(mAudioHandle), positionType, &position) == SA_SUCCESS) { return position / mChannels / sizeof(short); } return -1; } bool NativeAudioStream::IsPaused() { return mPaused; } int32_t NativeAudioStream::GetMinWriteSize() { size_t size; int r = sa_stream_get_min_write(static_cast(mAudioHandle), &size); if (r == SA_ERROR_NOT_SUPPORTED) return 1; else if (r != SA_SUCCESS || size > INT32_MAX) return -1; return static_cast(size / mChannels / sizeof(short)); } #if defined(MOZ_CUBEB) class nsCircularByteBuffer { public: nsCircularByteBuffer() : mBuffer(nullptr), mCapacity(0), mStart(0), mCount(0) {} // Set the capacity of the buffer in bytes. Must be called before any // call to append or pop elements. void SetCapacity(uint32_t aCapacity) { NS_ABORT_IF_FALSE(!mBuffer, "Buffer allocated."); mCapacity = aCapacity; mBuffer = new uint8_t[mCapacity]; } uint32_t Length() { return mCount; } uint32_t Capacity() { return mCapacity; } uint32_t Available() { return Capacity() - Length(); } // Append aLength bytes from aSrc to the buffer. Caller must check that // sufficient space is available. void AppendElements(const uint8_t* aSrc, uint32_t aLength) { NS_ABORT_IF_FALSE(mBuffer && mCapacity, "Buffer not initialized."); NS_ABORT_IF_FALSE(aLength <= Available(), "Buffer full."); uint32_t end = (mStart + mCount) % mCapacity; uint32_t toCopy = NS_MIN(mCapacity - end, aLength); memcpy(&mBuffer[end], aSrc, toCopy); memcpy(&mBuffer[0], aSrc + toCopy, aLength - toCopy); mCount += aLength; } // Remove aSize bytes from the buffer. Caller must check returned size in // aSize{1,2} before using the pointer returned in aData{1,2}. Caller // must not specify an aSize larger than Length(). void PopElements(uint32_t aSize, void** aData1, uint32_t* aSize1, void** aData2, uint32_t* aSize2) { NS_ABORT_IF_FALSE(mBuffer && mCapacity, "Buffer not initialized."); NS_ABORT_IF_FALSE(aSize <= Length(), "Request too large."); *aData1 = &mBuffer[mStart]; *aSize1 = NS_MIN(mCapacity - mStart, aSize); *aData2 = &mBuffer[0]; *aSize2 = aSize - *aSize1; mCount -= *aSize1 + *aSize2; mStart += *aSize1 + *aSize2; mStart %= mCapacity; } private: nsAutoArrayPtr mBuffer; uint32_t mCapacity; uint32_t mStart; uint32_t mCount; }; class BufferedAudioStream : public AudioStream { public: BufferedAudioStream(); ~BufferedAudioStream(); nsresult Init(int32_t aNumChannels, int32_t aRate, const dom::AudioChannelType aAudioChannelType); void Shutdown(); nsresult Write(const AudioDataValue* aBuf, uint32_t aFrames); uint32_t Available(); void SetVolume(double aVolume); void Drain(); void Pause(); void Resume(); int64_t GetPosition(); int64_t GetPositionInFrames(); int64_t GetPositionInFramesInternal(); bool IsPaused(); int32_t GetMinWriteSize(); private: static long DataCallback_S(cubeb_stream*, void* aThis, void* aBuffer, long aFrames) { return static_cast(aThis)->DataCallback(aBuffer, aFrames); } static void StateCallback_S(cubeb_stream*, void* aThis, cubeb_state aState) { static_cast(aThis)->StateCallback(aState); } long DataCallback(void* aBuffer, long aFrames); void StateCallback(cubeb_state aState); long GetUnprocessed(void* aBuffer, long aFrames); long GetTimeStretched(void* aBuffer, long aFrames); // Shared implementation of underflow adjusted position calculation. // Caller must own the monitor. int64_t GetPositionInFramesUnlocked(); // The monitor is held to protect all access to member variables. Write() // waits while mBuffer is full; DataCallback() notifies as it consumes // data from mBuffer. Drain() waits while mState is DRAINING; // StateCallback() notifies when mState is DRAINED. Monitor mMonitor; // Sum of silent frames written when DataCallback requests more frames // than are available in mBuffer. uint64_t mLostFrames; // Temporary audio buffer. Filled by Write() and consumed by // DataCallback(). Once mBuffer is full, Write() blocks until sufficient // space becomes available in mBuffer. mBuffer is sized in bytes, not // frames. nsCircularByteBuffer mBuffer; // Software volume level. Applied during the servicing of DataCallback(). double mVolume; // Owning reference to a cubeb_stream. cubeb_stream_destroy is called by // nsAutoRef's destructor. nsAutoRef mCubebStream; uint32_t mBytesPerFrame; uint32_t BytesToFrames(uint32_t aBytes) { NS_ASSERTION(aBytes % mBytesPerFrame == 0, "Byte count not aligned on frames size."); return aBytes / mBytesPerFrame; } uint32_t FramesToBytes(uint32_t aFrames) { return aFrames * mBytesPerFrame; } enum StreamState { INITIALIZED, // Initialized, playback has not begun. STARTED, // Started by a call to Write() (iff INITIALIZED) or Resume(). STOPPED, // Stopped by a call to Pause(). DRAINING, // Drain requested. DataCallback will indicate end of stream // once the remaining contents of mBuffer are requested by // cubeb, after which StateCallback will indicate drain // completion. DRAINED, // StateCallback has indicated that the drain is complete. ERRORED // Stream disabled due to an internal error. }; StreamState mState; }; #endif AudioStream* AudioStream::AllocateStream() { #if defined(MOZ_CUBEB) if (GetUseCubeb()) { return new BufferedAudioStream(); } #endif return new NativeAudioStream(); } #if defined(MOZ_CUBEB) BufferedAudioStream::BufferedAudioStream() : mMonitor("BufferedAudioStream"), mLostFrames(0), mVolume(1.0), mBytesPerFrame(0), mState(INITIALIZED) { } BufferedAudioStream::~BufferedAudioStream() { Shutdown(); } nsresult BufferedAudioStream::Init(int32_t aNumChannels, int32_t aRate, const dom::AudioChannelType aAudioChannelType) { cubeb* cubebContext = GetCubebContext(); if (!cubebContext || aNumChannels < 0 || aRate < 0) { return NS_ERROR_FAILURE; } mInRate = mOutRate = aRate; mChannels = aNumChannels; cubeb_stream_params params; params.rate = aRate; params.channels = aNumChannels; if (AUDIO_OUTPUT_FORMAT == AUDIO_FORMAT_S16) { params.format = CUBEB_SAMPLE_S16NE; } else { params.format = CUBEB_SAMPLE_FLOAT32NE; } mBytesPerFrame = sizeof(AudioDataValue) * aNumChannels; mAudioClock.Init(); { cubeb_stream* stream; if (cubeb_stream_init(cubebContext, &stream, "BufferedAudioStream", params, GetCubebLatency(), DataCallback_S, StateCallback_S, this) == CUBEB_OK) { mCubebStream.own(stream); } } if (!mCubebStream) { return NS_ERROR_FAILURE; } // Size mBuffer for one second of audio. This value is arbitrary, and was // selected based on the observed behaviour of the existing AudioStream // implementations. uint32_t bufferLimit = FramesToBytes(aRate); NS_ABORT_IF_FALSE(bufferLimit % mBytesPerFrame == 0, "Must buffer complete frames"); mBuffer.SetCapacity(bufferLimit); return NS_OK; } void BufferedAudioStream::Shutdown() { if (mState == STARTED) { Pause(); } if (mCubebStream) { mCubebStream.reset(); } } nsresult BufferedAudioStream::Write(const AudioDataValue* aBuf, uint32_t aFrames) { MonitorAutoLock mon(mMonitor); if (!mCubebStream || mState == ERRORED) { return NS_ERROR_FAILURE; } NS_ASSERTION(mState == INITIALIZED || mState == STARTED, "Stream write in unexpected state."); const uint8_t* src = reinterpret_cast(aBuf); uint32_t bytesToCopy = FramesToBytes(aFrames); while (bytesToCopy > 0) { uint32_t available = NS_MIN(bytesToCopy, mBuffer.Available()); NS_ABORT_IF_FALSE(available % mBytesPerFrame == 0, "Must copy complete frames."); mBuffer.AppendElements(src, available); src += available; bytesToCopy -= available; if (mState != STARTED) { int r; { MonitorAutoUnlock mon(mMonitor); r = cubeb_stream_start(mCubebStream); } mState = r == CUBEB_OK ? STARTED : ERRORED; } if (mState != STARTED) { return NS_ERROR_FAILURE; } if (bytesToCopy > 0) { mon.Wait(); } } return NS_OK; } uint32_t BufferedAudioStream::Available() { MonitorAutoLock mon(mMonitor); NS_ABORT_IF_FALSE(mBuffer.Length() % mBytesPerFrame == 0, "Buffer invariant violated."); return BytesToFrames(mBuffer.Available()); } int32_t BufferedAudioStream::GetMinWriteSize() { return 1; } void BufferedAudioStream::SetVolume(double aVolume) { MonitorAutoLock mon(mMonitor); NS_ABORT_IF_FALSE(aVolume >= 0.0 && aVolume <= 1.0, "Invalid volume"); mVolume = aVolume; } void BufferedAudioStream::Drain() { MonitorAutoLock mon(mMonitor); if (mState != STARTED) { return; } mState = DRAINING; while (mState == DRAINING) { mon.Wait(); } } void BufferedAudioStream::Pause() { MonitorAutoLock mon(mMonitor); if (!mCubebStream || mState != STARTED) { return; } int r; { MonitorAutoUnlock mon(mMonitor); r = cubeb_stream_stop(mCubebStream); } if (mState != ERRORED && r == CUBEB_OK) { mState = STOPPED; } } void BufferedAudioStream::Resume() { MonitorAutoLock mon(mMonitor); if (!mCubebStream || mState != STOPPED) { return; } int r; { MonitorAutoUnlock mon(mMonitor); r = cubeb_stream_start(mCubebStream); } if (mState != ERRORED && r == CUBEB_OK) { mState = STARTED; } } int64_t BufferedAudioStream::GetPosition() { return mAudioClock.GetPosition(); } // This function is miscompiled by PGO with MSVC 2010. See bug 768333. #ifdef _MSC_VER #pragma optimize("", off) #endif int64_t BufferedAudioStream::GetPositionInFrames() { return mAudioClock.GetPositionInFrames(); } #ifdef _MSC_VER #pragma optimize("", on) #endif int64_t BufferedAudioStream::GetPositionInFramesInternal() { MonitorAutoLock mon(mMonitor); return GetPositionInFramesUnlocked(); } int64_t BufferedAudioStream::GetPositionInFramesUnlocked() { mMonitor.AssertCurrentThreadOwns(); if (!mCubebStream || mState == ERRORED) { return -1; } uint64_t position = 0; { MonitorAutoUnlock mon(mMonitor); if (cubeb_stream_get_position(mCubebStream, &position) != CUBEB_OK) { return -1; } } // Adjust the reported position by the number of silent frames written // during stream underruns. uint64_t adjustedPosition = 0; if (position >= mLostFrames) { adjustedPosition = position - mLostFrames; } return NS_MIN(adjustedPosition, INT64_MAX); } bool BufferedAudioStream::IsPaused() { MonitorAutoLock mon(mMonitor); return mState == STOPPED; } long BufferedAudioStream::GetUnprocessed(void* aBuffer, long aFrames) { uint8_t* wpos = reinterpret_cast(aBuffer); // Flush the timestretcher pipeline, if we were playing using a playback rate // other than 1.0. uint32_t flushedFrames = 0; if (mTimeStretcher && mTimeStretcher->numSamples()) { flushedFrames = mTimeStretcher->receiveSamples(reinterpret_cast(wpos), aFrames); wpos += FramesToBytes(flushedFrames); } uint32_t toPopBytes = FramesToBytes(aFrames - flushedFrames); uint32_t available = NS_MIN(toPopBytes, mBuffer.Length()); void* input[2]; uint32_t input_size[2]; mBuffer.PopElements(available, &input[0], &input_size[0], &input[1], &input_size[1]); memcpy(wpos, input[0], input_size[0]); wpos += input_size[0]; memcpy(wpos, input[1], input_size[1]); return BytesToFrames(available) + flushedFrames; } long BufferedAudioStream::GetTimeStretched(void* aBuffer, long aFrames) { long processedFrames = 0; EnsureTimeStretcherInitialized(); uint8_t* wpos = reinterpret_cast(aBuffer); double playbackRate = static_cast(mInRate) / mOutRate; uint32_t toPopBytes = FramesToBytes(ceil(aFrames / playbackRate)); uint32_t available = 0; bool lowOnBufferedData = false; do { // Check if we already have enough data in the time stretcher pipeline. if (mTimeStretcher->numSamples() <= static_cast(aFrames)) { void* input[2]; uint32_t input_size[2]; available = NS_MIN(mBuffer.Length(), toPopBytes); if (available != toPopBytes) { lowOnBufferedData = true; } mBuffer.PopElements(available, &input[0], &input_size[0], &input[1], &input_size[1]); for(uint32_t i = 0; i < 2; i++) { mTimeStretcher->putSamples(reinterpret_cast(input[i]), BytesToFrames(input_size[i])); } } uint32_t receivedFrames = mTimeStretcher->receiveSamples(reinterpret_cast(wpos), aFrames - processedFrames); wpos += FramesToBytes(receivedFrames); processedFrames += receivedFrames; } while (processedFrames < aFrames && !lowOnBufferedData); return processedFrames; } long BufferedAudioStream::DataCallback(void* aBuffer, long aFrames) { MonitorAutoLock mon(mMonitor); uint32_t available = NS_MIN(static_cast(FramesToBytes(aFrames)), mBuffer.Length()); NS_ABORT_IF_FALSE(available % mBytesPerFrame == 0, "Must copy complete frames"); uint32_t underrunFrames = 0; uint32_t servicedFrames = 0; if (available) { AudioDataValue* output = reinterpret_cast(aBuffer); if (mInRate == mOutRate) { servicedFrames = GetUnprocessed(output, aFrames); } else { servicedFrames = GetTimeStretched(output, aFrames); } float scaled_volume = float(GetVolumeScale() * mVolume); ScaleAudioSamples(output, aFrames * mChannels, scaled_volume); NS_ABORT_IF_FALSE(mBuffer.Length() % mBytesPerFrame == 0, "Must copy complete frames"); // Notify any blocked Write() call that more space is available in mBuffer. mon.NotifyAll(); } underrunFrames = aFrames - servicedFrames; if (mState != DRAINING) { uint8_t* rpos = static_cast(aBuffer) + FramesToBytes(aFrames - underrunFrames); memset(rpos, 0, FramesToBytes(underrunFrames)); mLostFrames += underrunFrames; servicedFrames += underrunFrames; } mAudioClock.UpdateWritePosition(servicedFrames); return servicedFrames; } void BufferedAudioStream::StateCallback(cubeb_state aState) { MonitorAutoLock mon(mMonitor); if (aState == CUBEB_STATE_DRAINED) { mState = DRAINED; } else if (aState == CUBEB_STATE_ERROR) { mState = ERRORED; } mon.NotifyAll(); } #endif AudioClock::AudioClock(AudioStream* aStream) :mAudioStream(aStream), mOldOutRate(0), mBasePosition(0), mBaseOffset(0), mOldBaseOffset(0), mOldBasePosition(0), mPlaybackRateChangeOffset(0), mPreviousPosition(0), mWritten(0), mOutRate(0), mInRate(0), mPreservesPitch(true), mPlaybackRate(1.0), mCompensatingLatency(false) {} void AudioClock::Init() { mOutRate = mAudioStream->GetRate(); mInRate = mAudioStream->GetRate(); mPlaybackRate = 1.0; mOldOutRate = mOutRate; } void AudioClock::UpdateWritePosition(uint32_t aCount) { mWritten += aCount; } uint64_t AudioClock::GetPosition() { NS_ASSERTION(mInRate != 0 && mOutRate != 0, "AudioClock not initialized."); int64_t position = mAudioStream->GetPositionInFramesInternal(); int64_t diffOffset; if (position >= 0) { if (position < mPlaybackRateChangeOffset) { // See if we are still playing frames pushed with the old playback rate in // the backend. If we are, use the old output rate to compute the // position. mCompensatingLatency = true; diffOffset = position - mOldBaseOffset; position = static_cast(mOldBasePosition + static_cast(USECS_PER_S * diffOffset) / mOldOutRate); mPreviousPosition = position; return position; } if (mCompensatingLatency) { diffOffset = position - mPlaybackRateChangeOffset; mCompensatingLatency = false; mBasePosition = mPreviousPosition; } else { diffOffset = position - mPlaybackRateChangeOffset; } position = static_cast(mBasePosition + (static_cast(USECS_PER_S * diffOffset) / mOutRate)); return position; } return -1; } uint64_t AudioClock::GetPositionInFrames() { return (GetPosition() * mOutRate) / USECS_PER_S; } void AudioClock::SetPlaybackRate(double aPlaybackRate) { int64_t position = mAudioStream->GetPositionInFramesInternal(); if (position > mPlaybackRateChangeOffset) { mOldBasePosition = mBasePosition; mBasePosition = GetPosition(); mOldBaseOffset = mPlaybackRateChangeOffset; mBaseOffset = position; mPlaybackRateChangeOffset = mWritten; mOldOutRate = mOutRate; mOutRate = static_cast(mInRate / aPlaybackRate); } else { // The playbackRate has been changed before the end of the latency // compensation phase. We don't update the mOld* variable. That way, the // last playbackRate set is taken into account. mBasePosition = GetPosition(); mBaseOffset = position; mPlaybackRateChangeOffset = mWritten; mOutRate = static_cast(mInRate / aPlaybackRate); } } double AudioClock::GetPlaybackRate() { return mPlaybackRate; } void AudioClock::SetPreservesPitch(bool aPreservesPitch) { mPreservesPitch = aPreservesPitch; } bool AudioClock::GetPreservesPitch() { return mPreservesPitch; } } // namespace mozilla