This interface allows an application to enable the speaker-dependent and speaker-independent templates, and to add new words to the speaker-dependent template.
Method |
Description |
| IVCmdUserWord::AddSIFile | Installs speaker-independent template extension files. |
| IVCmdUserWord::GetPhraseList | Gets the words in the installed vocabulary. |
| IVCmdUserWord::ModifyTraining | Specifies which templates are enabled for a particular phrase. |
| IVCmdUserWord::QueryPhrase | Determines what kind of templates a phrase has and whether or not they are enabled. |
| IVCmdUserWord::RemoveSIFile | Uninstalls speaker-independent template extension files. |
| IVCmdUserWord::Train | Called by the application to train a list of user-defined phrases. |
This interface is an extension of the Speech API (SAPI), added to meet the needs of the Auto PC. It is designed specifically for an isolated-word recognizer. Any method call on this interface affects the current speaker only.
Templates hold information that the engine uses to recognize a phrase. There are two types of templates for the Auto PC: speaker independent and speaker dependent. There is one speaker-independent template for each phrase. Speaker independent recognition can only be enabled or disabled. It cannot be modified by the user. Each speaker can have one speaker-dependent template for each phrase. To create a speaker-dependent template, a user must train the object to recognize their particular speech pattern.
The two templates operate independently of each other. An application can enable a speaker-dependent template whether or not the speaker-independent template is available. Enabling both templates may achieve better recognition accuracy.