ISpeechRecoResult
The ISpeechRecoResult automation interface returns information about a recognition attempt.
A recognition result is returned by a recognition context in the three following cases:
- A successful recognition
- An intermediate recognition (also called a hypothesis)
- An unsuccessful recognition (or a false recognition)
A successful recognition is a word or phrase that surpasses a predetermined confidence rating. It is considered to be accurate enough to be passed back to the user as the text that was actually spoken. A hypothesis is an intermediate step toward recognition. The text has been parsed and examined and is available to the user for closer examination. Any number of hypotheses may be produced during a recognition attempt. A hypothesis may not reflect the final recognition and should not be used to predict it. A false recognition is a recognized word or phrase that does not meet or exceed a predetermined confidence rating. The false recognition will still contain a valid recognition result including text representing the speech. However, the text was not able to meet confidence criteria. Any of the following can contribute to a false recognition:
- Background noise
- Inexact pronunciation
- Uncommon words
- Unusual sequence of words
Any one of the three recognition types above is treated the same for a recognition result. A valid recognition result is returned by SAPI and its content may be examined. Information includes the phrase itself, the owning recognition context, the audio format (if the audio was retained) and other properties in this class.
Automation Interface Elements
The ISpeechRecoResult automation interface contains the following elements:
| Properties | Description |
|---|---|
| AudioFormat Property | Gets or sets the audio stream format. |
| PhraseInfo Property | Returns an ISpeechPhraseInfo structure containing detailed information about the last recognized phrase. |
| RecoContext Property | Retrieves the current ISpeechRecoContext for the recognizer. |
| Times Property | Retrieves the time information associated with the result. |
| Methods | Description |
|---|---|
| Alternates Method | Returns a list of alternative words. |
| Audio Method | Creates an audio stream from the audio data in the result object. |
| DiscardResultInfo Method | Discards the requested data from a phrase object. |
| SaveToMemory Method | Saves the phrase portion of the recognition result to memory. |
| SpeakAudio Method | Plays the audio sequence containing the recognized phrase. |