Specifying parameters for an inproc Recognizer

Article
03/01/2010

I’ve seen a couple of questions come by the listen alias about using an in-process speech recognizer. When you use an in-process recognizer, you have to specify a lot more than you need to specify with the shared recognizer.

In particular, you need to specify an audio source and a recognition engine. The shared recognizer uses the defaults for both of these.

Most people figure out that they need to specify an audio source; more commonly (and I’ve tripped over this myself) they forget to specify a recognition engine.

And while there’s a nifty helper for the audio input in sphelper.h (namely, SpGetDefaultTokenFromCategoryId(SPCAT_AUDIOIN, &cpToken)), there isn’t a helper for the recognition engine.

So, here’s a helper function that returns the object token for the default shared recognizer:

// Get the object token for the default shared recognizer engine.
HRESULT SpGetDefaultSharedRecognizerToken(ISpObjectToken **ppSharedRecognizerToken)
{
HRESULT hr = S_OK;
const WCHAR pszWindowsCompatAtt[] = L"windowsV6compatible";
LANGID UIlangID = GetUserDefaultUILanguage();
// Find recognizers that match windowsV6compatible and the primary language id of the UI.
// Use primary language because there may be several recognizers available for different sub-languages,
// and there is generally only one UI language per primary language.
// e.g. US and UK English recognizers on a US English system.
// If so, to distinguish these the locale is used below.
CComPtr<IEnumSpObjectTokens> cpEnum;
hr = SpEnumTokensMatchingPrimaryLangID(SPCAT_RECOGNIZERS, PRIMARYLANGID(UIlangID), pszWindowsCompatAtt, &cpEnum);
ULONG ulRecognizers(0);
if (SUCCEEDED(hr))
{
hr = cpEnum->GetCount(&ulRecognizers);
}
if (SUCCEEDED(hr))
{
if (ulRecognizers == 0)
{
// If zero - error.
hr = SPERR_RECOGNIZER_NOT_FOUND;
}
else if (ulRecognizers == 1)
{
// If one - we are done.
hr = cpEnum->Item(0, ppSharedRecognizerToken);
}
else
{
// More than one recognizer matches - we must pick the best default:
BOOL fFoundDefault = FALSE;
// If there's no language-specific default then use the locale.
LANGID localeID = GetUserDefaultLangID();
if (!fFoundDefault && SUCCEEDED(hr) &&
PRIMARYLANGID(UIlangID) == PRIMARYLANGID(localeID))
{
// First see if there's an engine whose language exactly matches the locale.
// For example if the langid is U.S English, but the locale is U.K. English
// and there's a U.K. recognizer then use that.
hr = S_OK;
WCHAR pszLocaleString[] = L"Language=XXXXXXXX";
SpHexFromUlong(pszLocaleString + wcslen(pszLocaleString) - 8, localeID);
for (ULONG ul = 0; SUCCEEDED(hr) && ul < ulRecognizers; ul++)
{
CComPtr<ISpObjectToken> cpToken;
hr = cpEnum->Item(ul, &cpToken);
if (SUCCEEDED(hr))
{
hr = cpToken->MatchesAttributes(pszLocaleString, &fFoundDefault);
}
if (SUCCEEDED(hr) && fFoundDefault)
{
*ppSharedRecognizerToken = cpToken.Detach();
break;
}
}
}
// If there's no engine that directly matches the locale see if the backup list of SupportedLocales is matched.
if (!fFoundDefault && (SUCCEEDED(hr) || hr == SPERR_NOT_FOUND) &&
PRIMARYLANGID(UIlangID) == PRIMARYLANGID(localeID))
{
// See if there's an engine which has a supported Locale matches the current locale.
// For example if the langid is U.S English, but the locale is Australian English
// there may be no recognizer that directly supports Australian English,
// but a U.K. recognizer might specify it can be used in the Australian English locale with this attribute.
hr = S_OK;
WCHAR pszSupportedLocalesString[] = L"SupportedLocales=XXXXXXXX";
SpHexFromUlong(pszSupportedLocalesString + wcslen(pszSupportedLocalesString) - 8, localeID);
for (ULONG ul = 0; SUCCEEDED(hr) && ul < ulRecognizers; ul++)
{
CComPtr<ISpObjectToken> cpToken;
hr = cpEnum->Item(ul, &cpToken);
if (SUCCEEDED(hr))
{
hr = cpToken->MatchesAttributes(pszSupportedLocalesString, &fFoundDefault);
}
if (SUCCEEDED(hr) && fFoundDefault)
{
*ppSharedRecognizerToken = cpToken.Detach();
break;
}
}
}
// We still haven't found a match - just pick the first recognizer.
if (!fFoundDefault && (SUCCEEDED(hr) || hr == SPERR_NOT_FOUND))
{
CComPtr<ISpObjectToken> cpToken;
hr = cpEnum->Item(0, ppSharedRecognizerToken);
}
}
}
return hr;
}
// Return a token enumerator containing all tokens that match the primary language
// of a particular language id. pszRequiredAttributes can be used to specify additional attributes all tokens must have.
HRESULT SpEnumTokensMatchingPrimaryLangID(const LPCWSTR pszCategoryId, LANGID priLangID, LPCWSTR pszRequiredAtts,
IEnumSpObjectTokens **ppEnum)
{
HRESULT hr = S_OK;
// First enumerate the tokens using pszRequiredAtts.
CComPtr<ISpObjectTokenCategory> cpCategory;
hr = SpGetCategoryFromId(pszCategoryId, &cpCategory);
CComPtr<IEnumSpObjectTokens> cpEnum;
if (SUCCEEDED(hr))
{
hr = cpCategory->EnumTokens(pszRequiredAtts, NULL, &cpEnum);
}
ULONG ulTokens(0);
if (SUCCEEDED(hr))
{
hr = cpEnum->GetCount(&ulTokens);
}
// Create enumerator to store new tokens.
CComPtr<ISpObjectTokenEnumBuilder> cpBuilder;
if (SUCCEEDED(hr))
{
hr = cpBuilder.CoCreateInstance(CLSID_SpObjectTokenEnum);
}
if (SUCCEEDED(hr))
{
hr = cpBuilder->SetAttribs(NULL, NULL);
}
// Now, for each token, check language string to see if it matches.
for (ULONG ul = 0; SUCCEEDED(hr) && ul < ulTokens; ul++)
{
LANGID tokenLangID(0);
CComPtr<ISpObjectToken> cpToken;
hr = cpEnum->Item(ul, &cpToken);
if (SUCCEEDED(hr))
{
// Just look at the first language id
hr = SpGetLanguageFromToken(cpToken, &tokenLangID);
}
if (SUCCEEDED(hr) && PRIMARYLANGID(tokenLangID) == PRIMARYLANGID(priLangID))
{
// Add to builder
hr = cpBuilder->AddTokens(1, &(cpToken.p));
}
}
if (SUCCEEDED(hr))
{
hr = cpBuilder->Reset();
}
if (SUCCEEDED(hr))
{
*ppEnum = cpBuilder.Detach();
}
return hr;
}

Specifying parameters for an inproc Recognizer

Additional resources