/*****************************************************************************************
* CSimpleDict::RecoEvent()
* Called whenever the dialog process is notified of a recognition event.
* Inserts whatever is recognized into the edit box.
******************************************************************************************/
void CSimpleDict::RecoEvent()
{
USES_CONVERSION;
CSpEvent event;
// Process all of the recognition events
while (event.GetFrom(m_cpRecoCtxt) == S_OK)
{
switch (event.eEventId)
{
case SPEI_SOUND_START:
m_bInSound = TRUE;
break;
case SPEI_SOUND_END:
if (m_bInSound)
{
m_bInSound = FALSE;
if (!m_bGotReco)
{
// The sound has started and ended,
// but the engine has not succeeded in recognizing anything
const TCHAR szNoise[] = _T("<noise>");
::SendDlgItemMessage( m_hDlg, IDC_EDIT_DICT,
EM_REPLACESEL, TRUE, (LPARAM) szNoise );
}
m_bGotReco = FALSE;
}
break;
case SPEI_RECOGNITION:
// There may be multiple recognition results, so get all of them
{
m_bGotReco = TRUE;
static const WCHAR wszUnrecognized[] = L"<Unrecognized>";
CSpDynamicString dstrText;
if (FAILED(event.RecoResult()->GetText(SP_GETWHOLEPHRASE, SP_GETWHOLEPHRASE, TRUE,
&dstrText, NULL)))
{
dstrText = wszUnrecognized;
}
// Concatenate a space onto the end of the recognized word
dstrText.Append(L" ");
::SendDlgItemMessage( m_hDlg, IDC_EDIT_DICT, EM_REPLACESEL, TRUE, (LPARAM) W2T(dstrText) );
}
break;
}
}
}
void CMyDlg::OnRecoEvent()
{
//CMyPackManApp *pApp = (CMyPackManApp *)AfxGetApp(); //View -> App
CMainFrame *pMain = (CMainFrame *)AfxGetMainWnd(); //View -> MainFrm
CMyPackManView *pView = (CMyPackManView *)pMain->GetActiveView();
USES_CONVERSION;
CSpEvent event;
//MessageBox(L"A");
while (event.GetFrom(m_cpRecoCtxt) == S_OK)
{
switch (event.eEventId)
{
case SPEI_RECOGNITION:
{
m_bReco = TRUE;
static const WCHAR wszUnrecognized[] = L"fail";
CSpDynamicString dstrText;
if (FAILED(event.RecoResult()->GetText(SP_GETWHOLEPHRASE, SP_GETWHOLEPHRASE, TRUE,
&dstrText, NULL)))
{
dstrText = wszUnrecognized;
}
dstrText.Append(L" ");
m_Edit = dstrText;
//if (m_Edit == "up " || m_Edit == "Up " || m_Edit == "Down " || m_Edit == "down " || m_Edit == "left " || m_Edit == "Left " || m_Edit == "Right " || m_Edit == "right "
// || m_Edit == "let " || m_Edit == "light " || m_Edit == "night ")
// pView->m_EditV = m_Edit;
////::SendDlgItemMessage(m_hWnd, IDC_EDIT_DICT, EM_REPLACESEL, TRUE, (LPARAM)W2T(dstrText));
//UpdateData(TRUE);
//m_pView->Invalidate();
if (m_Edit == "down " || m_Edit == "Down ")
m_pView->mKey = DOWN;
else if (m_Edit == "up " || m_Edit == "Up ")
m_pView->mKey = UP;
else if (m_Edit == "one " || m_Edit == "One ")
m_pView->mKey = LEFT;
else if (m_Edit == "two " || m_Edit == "Two ")
m_pView->mKey = RITE;
UpdateData(TRUE);
//UpdateData(FALSE);
}
break;
}
}
}
//-----------------------------------------------------------------------------
// Purpose: Debugging, prints alternate list if one is created
// Input : cpResult -
// (*pfnPrint -
//-----------------------------------------------------------------------------
void PrintAlternates( ISpRecoResult* cpResult, void (*pfnPrint)( const char *fmt, ... ) )
{
ISpPhraseAlt *rgPhraseAlt[ 32 ];
memset( rgPhraseAlt, 0, sizeof( rgPhraseAlt ) );
ULONG ulCount;
ISpPhrase *phrase = ( ISpPhrase * )cpResult;
if ( phrase )
{
SPPHRASE *pElements;
if ( SUCCEEDED( phrase->GetPhrase( &pElements ) ) )
{
if ( pElements->Rule.ulCountOfElements > 0 )
{
HRESULT hr = cpResult->GetAlternates(
pElements->Rule.ulFirstElement,
pElements->Rule.ulCountOfElements,
32,
rgPhraseAlt,
&ulCount);
Assert( !FAILED( hr ) );
for ( ULONG r = 0 ; r < ulCount; r++ )
{
CSpDynamicString dstrText;
hr = rgPhraseAlt[ r ]->GetText( (ULONG)SP_GETWHOLEPHRASE, (ULONG)SP_GETWHOLEPHRASE, TRUE, &dstrText, NULL);
Assert( !FAILED( hr ) );
pfnPrint( "[ ALT ]" );
pfnPrint( dstrText.CopyToChar() );
pfnPrint( "\r\n" );
}
}
}
}
for ( int i = 0; i < 32; i++ )
{
if ( rgPhraseAlt[ i ] )
{
rgPhraseAlt[ i ]->Release();
rgPhraseAlt[ i ] = NULL;
}
}
}
int voiceRecognition(string Qtext)
{
HRESULT hr = E_FAIL;
int word=0;
CSpDynamicString Qtextout;
if (SUCCEEDED(hr = ::CoInitialize(NULL)))
{
{
CComPtr<ISpRecoContext> cpRecoCtxt;
CComPtr<ISpRecoGrammar> cpGrammar;
CComPtr<ISpVoice> cpVoice;
hr = cpRecoCtxt.CoCreateInstance(CLSID_SpSharedRecoContext);
if (SUCCEEDED(hr))
{
hr = cpRecoCtxt->GetVoice(&cpVoice);
}
if (cpRecoCtxt && cpVoice &&
SUCCEEDED(hr = cpRecoCtxt->SetNotifyWin32Event()) &&
SUCCEEDED(hr = cpRecoCtxt->SetInterest(SPFEI(SPEI_RECOGNITION), SPFEI(SPEI_RECOGNITION))) &&
SUCCEEDED(hr = cpRecoCtxt->SetAudioOptions(SPAO_RETAIN_AUDIO, NULL, NULL)) &&
SUCCEEDED(hr = cpRecoCtxt->CreateGrammar(0, &cpGrammar)) &&
SUCCEEDED(hr = cpGrammar->LoadDictation(NULL, SPLO_STATIC)) &&
SUCCEEDED(hr = cpGrammar->SetDictationState(SPRS_ACTIVE)))
{
USES_CONVERSION;
CComPtr<ISpRecoResult> cpResult;
Qtextout.operator=(Qtext.c_str());
cpVoice->Speak(Qtextout, SPF_ASYNC, NULL);
if (SUCCEEDED(hr = BlockForResult(cpRecoCtxt, &cpResult)))
{
cpGrammar->SetDictationState(SPRS_INACTIVE);
CSpDynamicString dstrText;
if (SUCCEEDED(cpResult->GetText(SP_GETWHOLEPHRASE, SP_GETWHOLEPHRASE,
TRUE, &dstrText, NULL)))
{
cpResult.Release();
}
if (_wcsicmp(dstrText, L"True") == 0)
{
word = 1;
//break;
}
if (_wcsicmp(dstrText, L"Two") == 0)
{
word = 1;
//break;
}
if (_wcsicmp(dstrText, L"False") == 0)
{
word = 2;
//break;
}
if (_wcsicmp(dstrText, L"Falls") == 0)
{
word = 2;
//break;
}
if (_wcsicmp(dstrText, L"Follows") == 0)
{
word = 2;
//break;
}
if (_wcsicmp(dstrText, L"A") == 0)
{
word = 3;
//break;
}
if (_wcsicmp(dstrText, L"Eight") == 0)
{
word = 3;
//break;
}
if (_wcsicmp(dstrText, L"B") == 0)
{
word = 4;
//break;
}
if (_wcsicmp(dstrText, L"Bee") == 0)
{
word = 4;
//break;
}
if (_wcsicmp(dstrText, L"C") == 0)
{
word = 5;
//break;
}
if (_wcsicmp(dstrText, L"See") == 0)
{
word = 5;
//break;
}
if (_wcsicmp(dstrText, L"Fire") == 0)
{
word = 6;
//break;
}
if (_wcsicmp(dstrText, L"Leave") == 0)
{
word = 7;
//break;
}
if (_wcsicmp(dstrText, L"Leave it") == 0)
{
word = 7;
//break;
}
if (_wcsicmp(dstrText, L"We've") == 0)
{
word = 7;
//break;
}
if (_wcsicmp(dstrText, L"Quit") == 0)
{
word = 7;
//break;
}
if (_wcsicmp(dstrText, L"Quits") == 0)
{
word = 7;
//break;
}
if (_wcsicmp(dstrText, L"Switch") == 0)
{
word = 8;
//break;
}
if (_wcsicmp(dstrText, L"Change") == 0)
{
word = 8;
//break;
}
cout << dstrText.CopyToChar();
cpGrammar->SetDictationState(SPRS_ACTIVE);
}
}
}
::CoUninitialize();
}
return word;
}
LRESULT CALLBACK WndProc(HWND hwnd, UINT message, WPARAM wParam, LPARAM lParam)
{
HDC hdc;
PAINTSTRUCT ps;
switch (message)
{
case WM_CREATE:
{
//初始化COM端口
::CoInitializeEx(NULL, COINIT_APARTMENTTHREADED);
//创建识别引擎COM实例为共享型
HRESULT hr = m_cpRecoEngine.CoCreateInstance(CLSID_SpSharedRecognizer);
//创建识别上下文接口
if (SUCCEEDED(hr))
{
hr = m_cpRecoEngine->CreateRecoContext(&m_cpRecoCtxt);
}
else MessageBox(hwnd, TEXT("error1"), TEXT("error"), S_OK);
//设置识别消息,使计算机时刻监听语音消息
if (SUCCEEDED(hr))
{
hr = m_cpRecoCtxt->SetNotifyWindowMessage(hwnd, WM_RECOEVENT, 0, 0);
}
else MessageBox(hwnd, TEXT("error2"), TEXT("error"), S_OK);
//设置我们感兴趣的事件
if (SUCCEEDED(hr))
{
ULONGLONG ullMyEvents = SPFEI(SPEI_SOUND_START) | SPFEI(SPEI_RECOGNITION) | SPFEI(SPEI_SOUND_END);
hr = m_cpRecoCtxt->SetInterest(ullMyEvents, ullMyEvents);
}
else MessageBox(hwnd, TEXT("error3"), TEXT("error"), S_OK);
//创建语法规则
b_Cmd_Grammar = TRUE;
if (FAILED(hr))
{
MessageBox(hwnd, TEXT("error4"), TEXT("error"), S_OK);
}
hr = m_cpRecoCtxt->CreateGrammar(GID_CMD_GR, &m_cpCmdGramma);
WCHAR wszXMLFile[20] = L"er.xml";
MultiByteToWideChar(CP_ACP, 0, (LPCSTR)"er.xml", -1, wszXMLFile, 256);
hr = m_cpCmdGramma->LoadCmdFromFile(wszXMLFile, SPLO_DYNAMIC);
if (FAILED(hr))
{
MessageBox(hwnd, TEXT("error5"), TEXT("error"), S_OK);
}
b_initSR = TRUE;
//在开始识别时，激活语法进行识别
hr = m_cpCmdGramma->SetRuleState(NULL, NULL, SPRS_ACTIVE);
return 0;
}
case WM_RECOEVENT:
{
RECT rect;
GetClientRect(hwnd, &rect);
hdc = GetDC(hwnd);
USES_CONVERSION;
CSpEvent event;
while (event.GetFrom(m_cpRecoCtxt) == S_OK)
{
switch (event.eEventId)
{
case SPEI_RECOGNITION:
{
static const WCHAR wszUnrecognized[] = L"<Unrecognized>";
CSpDynamicString dstrText;
//取得识别结果
if (FAILED(event.RecoResult()->GetText(SP_GETWHOLEPHRASE, SP_GETWHOLEPHRASE, TRUE, &dstrText, NULL)))
{
dstrText = wszUnrecognized;
}
BSTR SRout;
dstrText.CopyToBSTR(&SRout);
char* lpszText2 = _com_util::ConvertBSTRToString(SRout);
if (b_Cmd_Grammar)
{
if (strstr("打开企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("打开企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
openQQ();
}
if (strstr("关闭企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("关闭企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
closeQQ();
}
if (strstr("隐藏企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("隐藏企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
yincangQQ();
}
if (strstr("显示企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("显示企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
showQQ();
}
if (strstr("上移企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("上移企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
top();
}
if (strstr("下移企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("下移企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
bottom();
}
if (strstr("左移企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("左移企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
left();
}
if (strstr("右移企鹅", lpszText2) != NULL)
{
DrawText(hdc, TEXT("右移企鹅"), -1, &rect, DT_SINGLELINE | DT_CENTER | DT_VCENTER);
right();
}
}
}
}
}
return TRUE;
}
case WM_PAINT:
hdc = BeginPaint(hwnd, &ps);
EndPaint(hwnd, &ps);
return 0;
case WM_DESTROY:
PostQuitMessage(0);
return 0;
}
return DefWindowProc(hwnd, message, wParam, lParam);
}
void CASRwrapper::GetText(std::wstring& speechRes, float* pconfidence, int requestedAlternates, std::wstring alternates[], float alternatesConfidence[])
{
//HRESULT hr = CoCreateInstance(CLSID_SpVoice, NULL, CLSCTX_ALL, IID_ISpVoice, (void **)&m_pVoice);
//hr = m_pVoice->Speak(speechRes, 0, NULL);
//m_pVoice->Release();
//m_pVoice = NULL;
const ULONG maxEvents = 10;
SPEVENT events[maxEvents];
ULONG eventCount;
HRESULT hr;
hr = m_cpRecoCtxt->GetEvents(maxEvents, events, &eventCount);
// Warning hr equal S_FALSE if everything is OK
// but eventCount < requestedEventCount
if (!(hr == S_OK || hr == S_FALSE)) {
return;
}
if (eventCount > 1)
{
speechRes.assign(L"More than one event!");
return;
}
ISpRecoResult* recoResult;
recoResult = reinterpret_cast<ISpRecoResult*>(events[0].lParam);
wchar_t* text;
hr = recoResult->GetText(SP_GETWHOLEPHRASE, SP_GETWHOLEPHRASE, FALSE, &text, NULL);
speechRes.assign(text);
//if (confidence != NULL)
//*confidence = recoResult->->pElements->SREngineConfidence;;
CoTaskMemFree(text);
if (requestedAlternates == 0 && pconfidence == NULL)
return;
const USHORT MAX_ALTERNATES = 100;
if (requestedAlternates > MAX_ALTERNATES)
requestedAlternates = MAX_ALTERNATES;
if (requestedAlternates == 0) //in case asked for confidence. i.e., pconfidence!=NULL
requestedAlternates = 1;
CComPtr<ISpPhraseAlt> pcpPhraseAlt[MAX_ALTERNATES];
SPPHRASE* pPhrase;
std::string betterResult;
float ConfidenceMax = 0.0;
ULONG ulCount;
//std::list<std::string> lWordsRec;
// Retrieve information about the recognized phrase
hr = recoResult->GetPhrase(&pPhrase);
if (SUCCEEDED(hr))
{
// Retrieve a list of alternative phrases related to the recognized phrase
hr = recoResult->GetAlternates(pPhrase->Rule.ulFirstElement,
pPhrase->Rule.ulCountOfElements,
requestedAlternates,
(ISpPhraseAlt**)pcpPhraseAlt,
&ulCount);
}
if (SUCCEEDED(hr))
{
// Browse the list of alternative phrases in order of highest likelyhood with the original phrase
for (unsigned int i = 0; i < ulCount; i++)
{
SPPHRASE* pPhraseAlt;
CSpDynamicString pwszAlternate;
// Retrieve information about the current alternative phrase
pcpPhraseAlt[i]->GetPhrase(&pPhraseAlt);
// Get the phrase's entire text string
hr = pcpPhraseAlt[i]->GetText(SP_GETWHOLEPHRASE, SP_GETWHOLEPHRASE, TRUE, &pwszAlternate, NULL);
if (SUCCEEDED(hr))
{
if (i == 1 && pconfidence != NULL)
*pconfidence = pPhraseAlt->pElements->SREngineConfidence;
if (alternatesConfidence != NULL)
alternatesConfidence[i] = pPhraseAlt->pElements->SREngineConfidence;
if (alternates != NULL)
alternates[i] = pwszAlternate.Copy(); // .CopyToChar();
}
}
}
}
//-----------------------------------------------------------------------------
// Purpose: Given a wave file and a string of words "text", creates a CFG from the
// sentence and stores the resulting words/phonemes in CSentence
// Input : *wavname -
// text -
// sentence -
// (*pfnPrint -
// Output : SR_RESULT
//-----------------------------------------------------------------------------
SR_RESULT ExtractPhonemes( const char *wavname, CSpDynamicString& text, CSentence& sentence, void (*pfnPrint)( const char *fmt, ...) )
{
// Assume failure
SR_RESULT result = SR_RESULT_ERROR;
if ( text.Length() <= 0 )
{
pfnPrint( "Error: no rule / text specified\n" );
return result;
}
USES_CONVERSION;
HRESULT hr;
CUtlVector < WORDRULETYPE > wordRules;
CComPtr<ISpStream> cpInputStream;
CComPtr<ISpRecognizer> cpRecognizer;
CComPtr<ISpRecoContext> cpRecoContext;
CComPtr<ISpRecoGrammar> cpRecoGrammar;
CComPtr<ISpPhoneConverter> cpPhoneConv;
// Create basic SAPI stream object
// NOTE: The helper SpBindToFile can be used to perform the following operations
hr = cpInputStream.CoCreateInstance(CLSID_SpStream);
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Stream object not installed?\n" );
return result;
}
CSpStreamFormat sInputFormat;
// setup stream object with wav file MY_WAVE_AUDIO_FILENAME
// for read-only access, since it will only be access by the SR engine
hr = cpInputStream->BindToFile(
T2W(wavname),
SPFM_OPEN_READONLY,
NULL,
sInputFormat.WaveFormatExPtr(),
SPFEI_ALL_EVENTS );
if ( FAILED( hr ) )
{
pfnPrint( "Error: couldn't open wav file %s\n", wavname );
return result;
}
// Create in-process speech recognition engine
hr = cpRecognizer.CoCreateInstance(CLSID_SpInprocRecognizer);
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 In process recognizer object not installed?\n" );
return result;
}
// Create recognition context to receive events
hr = cpRecognizer->CreateRecoContext(&cpRecoContext);
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to create recognizer context\n" );
return result;
}
// Create a grammar
hr = cpRecoContext->CreateGrammar( EP_GRAM_ID, &cpRecoGrammar );
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to create recognizer grammar\n" );
return result;
}
LANGID englishID = 0x409; // 1033 decimal
bool userSpecified = false;
LANGID langID = SpGetUserDefaultUILanguage();
// Allow commandline override
if ( CommandLine()->FindParm( "-languageid" ) != 0 )
{
userSpecified = true;
langID = CommandLine()->ParmValue( "-languageid", langID );
}
// Create a phoneme converter ( so we can convert to IPA codes )
hr = SpCreatePhoneConverter( langID, NULL, NULL, &cpPhoneConv );
if ( FAILED( hr ) )
{
if ( langID != englishID )
{
if ( userSpecified )
{
pfnPrint( "Warning: SAPI 5.1 Unable to create phoneme converter for command line override -languageid %i\n", langID );
}
else
{
pfnPrint( "Warning: SAPI 5.1 Unable to create phoneme converter for default UI language %i\n",langID );
}
// Try english!!!
langID = englishID;
hr = SpCreatePhoneConverter( langID, NULL, NULL, &cpPhoneConv );
}
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to create phoneme converter for English language id %i\n", langID );
return result;
}
else
{
pfnPrint( "Note: SAPI 5.1 Falling back to use english -languageid %i\n", langID );
}
}
else if ( userSpecified )
{
pfnPrint( "Note: SAPI 5.1 Using user specified -languageid %i\n",langID );
}
SPSTATEHANDLE hStateRoot;
// create/re-create Root level rule of grammar
hr = cpRecoGrammar->GetRule(L"Root", 0, SPRAF_TopLevel | SPRAF_Active, TRUE, &hStateRoot);
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to create root rule\n" );
return result;
}
// Inactivate it so we can alter it
hr = cpRecoGrammar->SetRuleState( NULL, NULL, SPRS_INACTIVE );
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to deactivate grammar rules\n" );
return result;
}
// Create the rule set from the words in text
{
CSpDynamicString currentWord;
WCHAR *pos = ( WCHAR * )text;
WCHAR str[ 2 ];
str[1]= 0;
while ( *pos )
{
if ( *pos == L' ' /*|| *pos == L'.' || *pos == L'-'*/ )
{
// Add word to rule set
if ( currentWord.Length() > 0 )
{
AddWordRule( cpRecoGrammar, &hStateRoot, &wordRules, currentWord );
currentWord.Clear();
}
pos++;
continue;
}
// Skip anything that's inside a [ xxx ] pair.
if ( *pos == L'[' )
{
while ( *pos && *pos != L']' )
{
pos++;
}
if ( *pos )
{
pos++;
}
continue;
}
str[ 0 ] = *pos;
currentWord.Append( str );
pos++;
}
if ( currentWord.Length() > 0 )
{
AddWordRule( cpRecoGrammar, &hStateRoot, &wordRules, currentWord );
}
if ( wordRules.Size() <= 0 )
{
pfnPrint( "Error: Text %s contained no usable words\n", text );
return result;
}
// Build all word to word transitions in the grammar
if ( !BuildRules( cpRecoGrammar, &hStateRoot, &wordRules ) )
{
pfnPrint( "Error: Rule set for %s could not be generated\n", text );
return result;
}
}
// check for recognitions and end of stream event
const ULONGLONG ullInterest =
SPFEI(SPEI_RECOGNITION) | SPFEI(SPEI_END_SR_STREAM) | SPFEI(SPEI_FALSE_RECOGNITION) |
SPFEI(SPEI_PHRASE_START ) | SPFEI(SPEI_HYPOTHESIS ) | SPFEI(SPEI_INTERFERENCE) ;
hr = cpRecoContext->SetInterest( ullInterest, ullInterest );
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to set interest level\n" );
return result;
}
// use Win32 events for command-line style application
hr = cpRecoContext->SetNotifyWin32Event();
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to set win32 notify event\n" );
return result;
}
// connect wav input to recognizer
// SAPI will negotiate mismatched engine/input audio formats using system audio codecs, so second parameter is not important - use default of TRUE
hr = cpRecognizer->SetInput(cpInputStream, TRUE);
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to associate input stream\n" );
return result;
}
// Activate the CFG ( rather than using dictation )
hr = cpRecoGrammar->SetRuleState( NULL, NULL, SPRS_ACTIVE );
if ( FAILED( hr ) )
{
switch ( hr )
{
case E_INVALIDARG:
pfnPrint( "pszName is invalid or bad. Alternatively, pReserved is non-NULL\n" );
break;
case SP_STREAM_UNINITIALIZED:
pfnPrint( "ISpRecognizer::SetInput has not been called with the InProc recognizer\n" );
break;
case SPERR_UNINITIALIZED:
pfnPrint( "The object has not been properly initialized.\n");
break;
case SPERR_UNSUPPORTED_FORMAT:
pfnPrint( "Audio format is bad or is not recognized. Alternatively, the device driver may be busy by another application and cannot be accessed.\n" );
break;
case SPERR_NOT_TOPLEVEL_RULE:
pfnPrint( "The rule pszName exists, but is not a top-level rule.\n" );
break;
default:
pfnPrint( "Unknown error\n" );
break;
}
pfnPrint( "Error: SAPI 5.1 Unable to activate rule set\n" );
return result;
}
// while events occur, continue processing
// timeout should be greater than the audio stream length, or a reasonable amount of time expected to pass before no more recognitions are expected in an audio stream
BOOL fEndStreamReached = FALSE;
while (!fEndStreamReached && S_OK == cpRecoContext->WaitForNotifyEvent( SR_WAVTIMEOUT ))
{
CSpEvent spEvent;
// pull all queued events from the reco context's event queue
while (!fEndStreamReached && S_OK == spEvent.GetFrom(cpRecoContext))
{
// Check event type
switch (spEvent.eEventId)
{
case SPEI_INTERFERENCE:
{
SPINTERFERENCE interference = spEvent.Interference();
switch ( interference )
{
case SPINTERFERENCE_NONE:
pfnPrint( "[ I None ]\r\n" );
break;
case SPINTERFERENCE_NOISE:
pfnPrint( "[ I Noise ]\r\n" );
break;
case SPINTERFERENCE_NOSIGNAL:
pfnPrint( "[ I No Signal ]\r\n" );
break;
case SPINTERFERENCE_TOOLOUD:
pfnPrint( "[ I Too Loud ]\r\n" );
break;
case SPINTERFERENCE_TOOQUIET:
pfnPrint( "[ I Too Quiet ]\r\n" );
break;
case SPINTERFERENCE_TOOFAST:
pfnPrint( "[ I Too Fast ]\r\n" );
break;
case SPINTERFERENCE_TOOSLOW:
pfnPrint( "[ I Too Slow ]\r\n" );
break;
default:
break;
}
}
break;
case SPEI_PHRASE_START:
pfnPrint( "Phrase Start\r\n" );
sentence.MarkNewPhraseBase();
break;
case SPEI_HYPOTHESIS:
case SPEI_RECOGNITION:
case SPEI_FALSE_RECOGNITION:
{
CComPtr<ISpRecoResult> cpResult;
cpResult = spEvent.RecoResult();
CSpDynamicString dstrText;
if (spEvent.eEventId == SPEI_FALSE_RECOGNITION)
{
dstrText = L"(Unrecognized)";
result = SR_RESULT_FAILED;
// It's possible that the failed recog might have more words, so see if that's the case
EnumeratePhonemes( cpPhoneConv, cpResult, sentence );
}
else
{
// Hypothesis or recognition success
cpResult->GetText( (ULONG)SP_GETWHOLEPHRASE, (ULONG)SP_GETWHOLEPHRASE, TRUE, &dstrText, NULL);
EnumeratePhonemes( cpPhoneConv, cpResult, sentence );
if ( spEvent.eEventId == SPEI_RECOGNITION )
{
result = SR_RESULT_SUCCESS;
}
pfnPrint( va( "%s%s\r\n", spEvent.eEventId == SPEI_HYPOTHESIS ? "[ Hypothesis ] " : "", dstrText.CopyToChar() ) );
}
cpResult.Release();
}
break;
// end of the wav file was reached by the speech recognition engine
case SPEI_END_SR_STREAM:
fEndStreamReached = TRUE;
break;
}
// clear any event data/object references
spEvent.Clear();
}// END event pulling loop - break on empty event queue OR end stream
}// END event polling loop - break on event timeout OR end stream
// Deactivate rule
hr = cpRecoGrammar->SetRuleState( NULL, NULL, SPRS_INACTIVE );
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to deactivate rule set\n" );
return result;
}
// close the input stream, since we're done with it
// NOTE: smart pointer will call SpStream's destructor, and consequently ::Close, but code may want to check for errors on ::Close operation
hr = cpInputStream->Close();
if ( FAILED( hr ) )
{
pfnPrint( "Error: SAPI 5.1 Unable to close input stream\n" );
return result;
}
return result;
}
//-----------------------------------------------------------------------------
// Purpose: Given a wavfile and a list of inwords, determines the word/phonene
// sample counts for the sentce
// Input : *wavfile -
// *inwords -
// *outphonemes{ text.Clear( -
// Output : SR_RESULT
//-----------------------------------------------------------------------------
static SR_RESULT SAPI_ExtractPhonemes(
const char *wavfile,
int numsamples,
void (*pfnPrint)( const char *fmt, ... ),
CSentence& inwords,
CSentence& outwords )
{
LogReset();
USES_CONVERSION;
CSpDynamicString text;
text.Clear();
HKEY hkwipe;
LONG lResult = RegOpenKeyEx( HKEY_CURRENT_USER, "Software\\Microsoft\\Speech\\RecoProfiles", 0, KEY_ALL_ACCESS, &hkwipe );
if ( lResult == ERROR_SUCCESS )
{
RecursiveRegDelKey( hkwipe );
RegCloseKey( hkwipe );
}
if ( strlen( inwords.GetText() ) <= 0 )
{
inwords.SetTextFromWords();
}
// Construct a string from the inwords array
text.Append( T2W( inwords.GetText() ) );
// Assume failure
SR_RESULT result = SR_RESULT_ERROR;
if ( text.Length() > 0 )
{
CSentence sentence;
pfnPrint( "Processing...\r\n" );
// Give it a try
result = ExtractPhonemes( wavfile, text, sentence, pfnPrint );
pfnPrint( "Finished.\r\n" );
// PrintWordsAndPhonemes( sentence, pfnPrint );
// Copy results to outputs
outwords.Reset();
outwords.SetText( inwords.GetText() );
Log( "Starting\n" );
LogWords( inwords );
if ( SR_RESULT_ERROR != result )
{
int i;
Log( "Hypothesized\n" );
LogWords( sentence );
for( i = 0 ; i < sentence.m_Words.Size(); i++ )
{
CWordTag *tag = sentence.m_Words[ i ];
if ( tag )
{
// Skip '...' tag
if ( stricmp( tag->GetWord(), "..." ) )
{
CWordTag *newTag = new CWordTag( *tag );
outwords.m_Words.AddToTail( newTag );
}
}
}
// Now insert unrecognized/skipped words from original list
//
int frompos = 0, topos = 0;
while( 1 )
{
// End of source list
if ( frompos >= inwords.m_Words.Size() )
break;
const CWordTag *fromTag = inwords.m_Words[ frompos ];
// Reached end of destination list, just copy words over from from source list until
// we run out of source words
if ( topos >= outwords.m_Words.Size() )
{
// Just copy words over
CWordTag *newWord = new CWordTag( *fromTag );
// Remove phonemes
while ( newWord->m_Phonemes.Size() > 0 )
{
CPhonemeTag *kill = newWord->m_Phonemes[ 0 ];
newWord->m_Phonemes.Remove( 0 );
delete kill;
}
outwords.m_Words.AddToTail( newWord );
frompos++;
topos++;
continue;
}
// Destination word
const CWordTag *toTag = outwords.m_Words[ topos ];
// Words match, just skip ahead
if ( !stricmp( fromTag->GetWord(), toTag->GetWord() ) )
{
frompos++;
topos++;
continue;
}
// The only case we handle is that something in the source wasn't in the destination
// Find the next source word that appears in the destination
int skipAhead = frompos + 1;
bool found = false;
while ( skipAhead < inwords.m_Words.Size() )
{
const CWordTag *sourceWord = inwords.m_Words[ skipAhead ];
if ( !stricmp( sourceWord->GetWord(), toTag->GetWord() ) )
{
found = true;
break;
}
skipAhead++;
}
// Uh oh destination has words that are not in source, just skip to next destination word?
if ( !found )
{
topos++;
}
else
{
// Copy words from from source list into destination
//
int skipCount = skipAhead - frompos;
while ( --skipCount>= 0 )
{
const CWordTag *sourceWord = inwords.m_Words[ frompos++ ];
CWordTag *newWord = new CWordTag( *sourceWord );
// Remove phonemes
while ( newWord->m_Phonemes.Size() > 0 )
{
CPhonemeTag *kill = newWord->m_Phonemes[ 0 ];
newWord->m_Phonemes.Remove( 0 );
delete kill;
}
outwords.m_Words.InsertBefore( topos, newWord );
topos++;
}
frompos++;
topos++;
}
}
Log( "\nDone simple check\n" );
LogWords( outwords );
LogPhonemes( outwords );
ComputeMissingByteSpans( numsamples, outwords );
Log( "\nFinal check\n" );
LogWords( outwords );
LogPhonemes( outwords );
}
}
else
{
pfnPrint( "Input sentence is empty!\n" );
}
// Return results
return result;
}
