Share via

ISpEnginePronunciation::GetPronunciations (SAPI 5.4)

  • Article

Microsoft Speech API 5.4

ISpEnginePronunciation::GetPronunciations

ISpEnginePronunciation::GetPronunciations returns all of the pronunciations the engine knows about for a given word.

  
    HRESULT GetPronunciations(
   LPCWSTR                    *pszWord,
   LPCWSTR                     pszLeftContext,
   LPCWSTR                     pszRightContext,
   LANGID                      LangID,
   SPWORDPRONUNCIATIONLIST    *pEnginePronunciationList
);

Parameters

  • pszWord
    [in] The word or term for which a pronunciation is desired (for example, "record"). This is case sensitive.
  • pszLeftContext
    [in] The left portion of a sentence or phrase in which pszWord is used. See the Remarks section below for more information.
  • pszRightContext
    [in] The right portion of a sentence or phrase in which pszWord is used. See the Remarks section below for more information.
  • LangID
    [in] The ID of the language. This parameter is useful if an engine supports multiple languages. To use the engine's default language, set the ID to 0.
  • pEnginePronunciationList
    [out] Returns a list of pronunciations. The application must allocate the SPWORDPRONUNCIATIONLIST structure that pEnginePronunciationList points to. The engine will allocate the space for the pronunciation list itself and provide a pointer to it in pEnginePronunciationList->pvBuffer. The application is responsible for calling CoTaskMemFree() on pEnginePronunciationList->pvBuffer to release the allocation.

Return values

Value
S_OK
E_FAIL
E_INVALIDARG
E_OUTOFMEMORY
SPERR_WORD_NEEDS_NORMALIZATION
SPERR_NO_WORD_PRONUNCIATION

Remarks

In the Microsoft Speech Recognition engine implementation of this interface and method, GetPronunciations returns UPS phone labels instead of phone IDs.

The pszLeftContext and pszRightContextparameters are always case sensitive and either of them can be set to null, in which case the engine ignores the parameter. In addition, the engine may ignore any context string.

Some engines (generally TTS) select pronunciations based on the context in which a word is used. The following table lists two examples of selected pronounciations based on context:

pszWord
"record"
"record"