virtual const char* open( void * hwnd, unsigned sample_rate, unsigned nch, unsigned max_samples_per_frame, unsigned num_frames )
{
this->hwnd = hwnd;
this->sample_rate = sample_rate;
this->nch = nch;
this->max_samples_per_frame = max_samples_per_frame;
this->num_frames = num_frames;
#ifdef HAVE_KS_HEADERS
WAVEFORMATEXTENSIBLE wfx;
wfx.Format.wFormatTag = WAVE_FORMAT_EXTENSIBLE;
wfx.Format.nChannels = nch; //1;
wfx.Format.nSamplesPerSec = sample_rate;
wfx.Format.nBlockAlign = 2 * nch; //2;
wfx.Format.nAvgBytesPerSec = wfx.Format.nSamplesPerSec * wfx.Format.nBlockAlign;
wfx.Format.wBitsPerSample = 16;
wfx.Format.cbSize = sizeof(WAVEFORMATEXTENSIBLE)-sizeof(WAVEFORMATEX);
wfx.Samples.wValidBitsPerSample = 16;
wfx.SubFormat = KSDATAFORMAT_SUBTYPE_PCM;
wfx.dwChannelMask = nch == 2 ? KSAUDIO_SPEAKER_STEREO : KSAUDIO_SPEAKER_MONO;
#else
WAVEFORMATEX wfx;
wfx.wFormatTag = WAVE_FORMAT_PCM;
wfx.nChannels = nch; //1;
wfx.nSamplesPerSec = sample_rate;
wfx.nBlockAlign = 2 * nch; //2;
wfx.nAvgBytesPerSec = wfx.nSamplesPerSec * wfx.nBlockAlign;
wfx.wBitsPerSample = 16;
wfx.cbSize = 0;
#endif
HRESULT hr = XAudio2Create( &xaud, 0 );
if (FAILED(hr)) return "Creating XAudio2 interface";
hr = xaud->CreateMasteringVoice(
&mVoice,
nch,
sample_rate,
0,
NULL,
NULL );
if (FAILED(hr)) return "Creating XAudio2 mastering voice";
hr = xaud->CreateSourceVoice( &sVoice, &wfx, 0, 4.0f, ¬ify );
if (FAILED(hr)) return "Creating XAudio2 source voice";
hr = sVoice->Start( 0 );
if (FAILED(hr)) return "Starting XAudio2 voice";
hr = sVoice->SetFrequencyRatio((float)1.0f);
if (FAILED(hr)) return "Setting XAudio2 voice frequency ratio";
buffered_count = 0;
buffer_read_cursor = 0;
buffer_write_cursor = 0;
samples_played = 0;
sample_buffer = new int16_t[ max_samples_per_frame * num_frames ];
samples_in_buffer = new UINT64[ num_frames ];
memset( samples_in_buffer, 0, sizeof( UINT64 ) * num_frames );
return NULL;
}
void OutputDeviceNodeXAudio::initSourceVoice()
{
CI_ASSERT( ! mSourceVoice );
auto context = dynamic_pointer_cast<ContextXAudio>( getContext() );
auto wfx = msw::interleavedFloatWaveFormat( getSampleRate(), getNumChannels() );
IXAudio2 *xaudio = context->getXAudio();
UINT32 flags = ( mFilterEnabled ? XAUDIO2_VOICE_USEFILTER : 0 );
HRESULT hr = xaudio->CreateSourceVoice( &mSourceVoice, wfx.get(), flags, XAUDIO2_DEFAULT_FREQ_RATIO, mVoiceCallback.get() );
CI_ASSERT( hr == S_OK );
}
bool XAudio2_Output::init(long sampleRate)
{
if( failed || initialized ) return false;
HRESULT hr;
// Initialize XAudio2
UINT32 flags = 0;
//#ifdef _DEBUG
// flags = XAUDIO2_DEBUG_ENGINE;
//#endif
hr = XAudio2Create( &xaud, flags );
if( hr != S_OK ) {
systemMessage( IDS_XAUDIO2_FAILURE, NULL );
failed = true;
return false;
}
freq = sampleRate;
// calculate the number of samples per frame first
// then multiply it with the size of a sample frame (16 bit * stereo)
soundBufferLen = ( freq / 60 ) * 4;
// create own buffers to store sound data because it must not be
// manipulated while the voice plays from it
buffers = (BYTE *)malloc( ( bufferCount + 1 ) * soundBufferLen );
// + 1 because we need one temporary buffer when all others are in use
WAVEFORMATEX wfx;
ZeroMemory( &wfx, sizeof( wfx ) );
wfx.wFormatTag = WAVE_FORMAT_PCM;
wfx.nChannels = 2;
wfx.nSamplesPerSec = freq;
wfx.wBitsPerSample = 16;
wfx.nBlockAlign = wfx.nChannels * ( wfx.wBitsPerSample / 8 );
wfx.nAvgBytesPerSec = wfx.nSamplesPerSec * wfx.nBlockAlign;
// create sound receiver
hr = xaud->CreateMasteringVoice(
&mVoice,
XAUDIO2_DEFAULT_CHANNELS,
XAUDIO2_DEFAULT_SAMPLERATE,
0,
theApp.xa2Device,
NULL );
if( hr != S_OK ) {
systemMessage( IDS_XAUDIO2_CANNOT_CREATE_MASTERINGVOICE, NULL );
failed = true;
return false;
}
// create sound emitter
hr = xaud->CreateSourceVoice( &sVoice, &wfx, 0, 4.0f, ¬ify );
if( hr != S_OK ) {
systemMessage( IDS_XAUDIO2_CANNOT_CREATE_SOURCEVOICE, NULL );
failed = true;
return false;
}
if( theApp.xa2Upmixing ) {
// set up stereo upmixing
XAUDIO2_DEVICE_DETAILS dd;
ZeroMemory( &dd, sizeof( dd ) );
hr = xaud->GetDeviceDetails( 0, &dd );
ASSERT( hr == S_OK );
float *matrix = NULL;
matrix = (float*)malloc( sizeof( float ) * 2 * dd.OutputFormat.Format.nChannels );
if( matrix == NULL ) return false;
bool matrixAvailable = true;
switch( dd.OutputFormat.Format.nChannels ) {
case 4: // 4.0
//Speaker \ Left Source Right Source
/*Front L*/ matrix[0] = 1.0000f; matrix[1] = 0.0000f;
/*Front R*/ matrix[2] = 0.0000f; matrix[3] = 1.0000f;
/*Back L*/ matrix[4] = 1.0000f; matrix[5] = 0.0000f;
/*Back R*/ matrix[6] = 0.0000f; matrix[7] = 1.0000f;
break;
case 5: // 5.0
//Speaker \ Left Source Right Source
/*Front L*/ matrix[0] = 1.0000f; matrix[1] = 0.0000f;
/*Front R*/ matrix[2] = 0.0000f; matrix[3] = 1.0000f;
/*Front C*/ matrix[4] = 0.7071f; matrix[5] = 0.7071f;
/*Side L*/ matrix[6] = 1.0000f; matrix[7] = 0.0000f;
/*Side R*/ matrix[8] = 0.0000f; matrix[9] = 1.0000f;
break;
case 6: // 5.1
//Speaker \ Left Source Right Source
/*Front L*/ matrix[0] = 1.0000f; matrix[1] = 0.0000f;
/*Front R*/ matrix[2] = 0.0000f; matrix[3] = 1.0000f;
/*Front C*/ matrix[4] = 0.7071f; matrix[5] = 0.7071f;
/*LFE */ matrix[6] = 0.0000f; matrix[7] = 0.0000f;
/*Side L*/ matrix[8] = 1.0000f; matrix[9] = 0.0000f;
/*Side R*/ matrix[10] = 0.0000f; matrix[11] = 1.0000f;
break;
case 7: // 6.1
//Speaker \ Left Source Right Source
/*Front L*/ matrix[0] = 1.0000f; matrix[1] = 0.0000f;
/*Front R*/ matrix[2] = 0.0000f; matrix[3] = 1.0000f;
/*Front C*/ matrix[4] = 0.7071f; matrix[5] = 0.7071f;
/*LFE */ matrix[6] = 0.0000f; matrix[7] = 0.0000f;
/*Side L*/ matrix[8] = 1.0000f; matrix[9] = 0.0000f;
/*Side R*/ matrix[10] = 0.0000f; matrix[11] = 1.0000f;
/*Back C*/ matrix[12] = 0.7071f; matrix[13] = 0.7071f;
break;
case 8: // 7.1
//Speaker \ Left Source Right Source
/*Front L*/ matrix[0] = 1.0000f; matrix[1] = 0.0000f;
/*Front R*/ matrix[2] = 0.0000f; matrix[3] = 1.0000f;
/*Front C*/ matrix[4] = 0.7071f; matrix[5] = 0.7071f;
/*LFE */ matrix[6] = 0.0000f; matrix[7] = 0.0000f;
/*Back L*/ matrix[8] = 1.0000f; matrix[9] = 0.0000f;
/*Back R*/ matrix[10] = 0.0000f; matrix[11] = 1.0000f;
/*Side L*/ matrix[12] = 1.0000f; matrix[13] = 0.0000f;
/*Side R*/ matrix[14] = 0.0000f; matrix[15] = 1.0000f;
break;
default:
matrixAvailable = false;
break;
}
if( matrixAvailable ) {
hr = sVoice->SetOutputMatrix( NULL, 2, dd.OutputFormat.Format.nChannels, matrix );
ASSERT( hr == S_OK );
}
free( matrix );
matrix = NULL;
}
hr = sVoice->Start( 0 );
ASSERT( hr == S_OK );
playing = true;
currentBuffer = 0;
device_changed = false;
initialized = true;
return true;
}
DWORD CWASAPICapture::DoCaptureThread()
{
HANDLE mmcssHandle = NULL;
IXAudio2* xaudio = 0;
IXAudio2MasteringVoice* mastering_voice = 0;
IXAudio2SourceVoice* source_voice = 0;
try {
bool stillPlaying = true;
DWORD mmcssTaskIndex = 0;
HRESULT hr = CoInitializeEx(NULL, COINIT_MULTITHREADED);
if (FAILED(hr))
{
printf_s("Unable to initialize COM in render thread: %x\n", hr);
return hr;
}
mmcssHandle = AvSetMmThreadCharacteristics(L"Audio", &mmcssTaskIndex);
if (mmcssHandle == NULL)
{
printf_s("Unable to enable MMCSS on capture thread: %d\n", GetLastError());
}
//
// XAudioの初期化
//
{
UINT32 flags = 0;
#ifdef _DEBUG
flags |= XAUDIO2_DEBUG_ENGINE;
#endif
if( FAILED( hr = XAudio2Create( &xaudio, flags ) ) )
throw "XAudio2Create";
// Create a mastering voice
if( FAILED( hr = xaudio->CreateMasteringVoice( &mastering_voice ) ) )
throw "CreateMasteringVoice";
// WAVファイルのWAVEFORMATEXを使ってSourceVoiceを作成
if( FAILED( xaudio->CreateSourceVoice( &source_voice, MixFormat() ) ) )
throw "CreateSourceVoice";
// 再生
source_voice->Start();
}
while (stillPlaying)
{
HRESULT hr;
//
// In Timer Driven mode, we want to wait for half the desired latency in milliseconds.
//
// That way we'll wake up half way through the processing period to pull the
// next set of samples from the engine.
//
DWORD waitResult = WaitForSingleObject(_ShutdownEvent, _EngineLatencyInMS / 2);
switch (waitResult)
{
case WAIT_OBJECT_0 + 0: // _ShutdownEvent
stillPlaying = false; // We're done, exit the loop.
break;
case WAIT_TIMEOUT: // Timeout
//
// We need to retrieve the next buffer of samples from the audio capturer.
//
BYTE *pData;
UINT32 framesAvailable;
DWORD flags;
//
// Find out how much capture data is available. We need to make sure we don't run over the length
// of our capture buffer. We'll discard any samples that don't fit in the buffer.
//
hr = _CaptureClient->GetBuffer(&pData, &framesAvailable, &flags, NULL, NULL);
if (SUCCEEDED(hr))
{
UINT32 framesToCopy = min(framesAvailable, static_cast<UINT32>((_CaptureBufferSize - _CurrentCaptureIndex) / _FrameSize));
if (framesToCopy != 0)
{
//
// The flags on capture tell us information about the data.
//
// We only really care about the silent flag since we want to put frames of silence into the buffer
// when we receive silence. We rely on the fact that a logical bit 0 is silence for both float and int formats.
//
if (flags & AUDCLNT_BUFFERFLAGS_SILENT)
{
//
// Fill 0s from the capture buffer to the output buffer.
//
ZeroMemory(&_CaptureBuffer[_CurrentCaptureIndex], framesToCopy*_FrameSize);
}
else
{
//
// Copy data from the audio engine buffer to the output buffer.
//
CopyMemory(&_CaptureBuffer[_CurrentCaptureIndex], pData, framesToCopy*_FrameSize);
// SourceVoiceにデータを送信
XAUDIO2_BUFFER buffer = { 0 };
buffer.AudioBytes = framesToCopy * _FrameSize; //バッファのバイト数
buffer.pAudioData = &pData[ 0 ]; //バッファの先頭アドレス
source_voice->SubmitSourceBuffer( &buffer );
}
//
// Bump the capture buffer pointer.
//
_CurrentCaptureIndex += framesToCopy*_FrameSize;
}
hr = _CaptureClient->ReleaseBuffer(framesAvailable);
if (FAILED(hr))
{
printf_s("Unable to release capture buffer: %x!\n", hr);
}
}
break;
}
}
}
catch( const char* e )
{
std::cout << e << std::endl;
}
// Cleanup XAudio2
if( mastering_voice != 0 ) {
// ここで落ちる
//mastering_voice->DestroyVoice();
mastering_voice = 0;
}
if( xaudio != 0 ) {
// ここでも落ちる
//xaudio->Release();
xaudio = 0;
}
AvRevertMmThreadCharacteristics(mmcssHandle);
CoUninitialize();
return 0;
}