ISpRecoResult::GetAudio
ISpRecoResult::GetAudio creates an audio stream of the requested words from the audio data in the result object.
HRESULT GetAudio(
ULONG ulStartElement,
ULONG cElements,
ISpStreamFormat **ppStream
);
Parameters
- ulStartElement
- [in] Value specifying from which element in the result data to start the audio stream.
- cElements
- [in] Value specifying the total number of words.
- ppStream
- [out] Address that will receive a pointer to an ISpStreamFormat object containing the audio data requested.
Return values
| Value | Description |
|---|---|
| S_OK | Function completed successfully. |
| E_INVALIDARG | cElements is zero or the expected number of elements to count exceeds the number available. |
| E_POINTER | ppStream is an invalid pointer. |
| SPERR_NO_AUDIO_DATA | This result object does not have any audio data. |
| FAILED(hr) | Appropriate error message. |
Remarks
Even if there are no elements, that is, ulStartElement = 0 and cElements = 0, the audio will still be played. There are "unrecognized" results that have no elements but have audio.
An application can find the time offsets for each element by examining the SPPHRASE object retrieved using ISpRecoResult::GetPhrase.
Example
The following code snippet illustrates the use ISpRecoResult::GetAudio to retrieve the retained audio.
// Declare local identifiers:
HRESULT hr = S_OK;
CComPtr<ISpRecoResult> cpRecoResult;
CComPtr<ISpStreamFormat> cpStreamFormat;
GUID formatId;
WAVEFORMATEX *pWaveFormatEx;
// ... Obtain a recognition result object from the recognizer ...
hr = cpRecoResult->GetAudio(0, 0, &cpStreamFormat;);
if (SUCCEEDED(hr))
{
// Check the format of the stream.
hr = cpStreamFormat->GetFormat(&formatId;, &pWaveFormatEx;);
}
if (SUCCEEDED(hr))
{
// Do stuff here.
}