Create Speech Server Grammars with Grammar Editor

This content is no longer actively maintained. It is provided as is, for anyone who may still be using these technologies, with no warranties or claims of accuracy with regard to the most recent product version or service release.

Use Speech Grammar Editor to develop speech recognition grammars in Speech Server. Speech Grammar Editor is most appropriate for developing grammars to recognize answers where the user provides answers to questions that the system has not asked yet (called mixed-initiative dialog) or for authoring complex rules that resemble regular expressions.

You can also create speech recognition grammars in Speech Server with Conversational Grammar Builder.

More Complex Grammars

Use Speech Grammar Editor to author the grammar if the user's answer is more complex than a single phrase. An example of this is the postal code used in Great Britain, where the postal code itself is the keyword and is best represented as a regular expression.

Speech Grammar Editor can also be used if there is more than one keyword, for example with mixed-initiative or multi-token input.

Mixed-Initiative Dialogs

Mixed-initiative dialogs allow a user to answer multiple questions with a single utterance. For example, an application might ask for a travel destination and the user responds with destination, time, and date. This style enables non-sequential dialog. Mixed-initiative dialogs are typically more difficult to design than system-initiative dialogs, but they provide users with greater flexibility when answering questions. Mixed-initiative dialogs simulate human interaction more closely than system-initiative dialogs.

Large Grammar Editor Files

Large grammars take more time to load in Speech Grammar Editor than medium-sized grammars. To avoid this problem, divide large Speech Grammar Editor grammars according to logical functions. It is more efficient to work with multiple smaller grammars than one large grammar, just as it is easier to break code into functions and subroutines for easier reuse and management.

Authors can create grammar files programmatically, for example using databases with thousands of rows, and then try to edit the resulting files in Speech Grammar Editor. Speech Grammar Editor is intended for manually creating and editing grammars and is not an appropriate tool for editing very large files.

String Comparisons with Recognition Text

The Text property of the RecognitionResult class always returns text as lowercase, regardless of the case entered in the grammar. As a result, when using the value of Recognition.Text in application code, authors should expect that value to be returned as all lowercase.

See Also

Other Resources

Use Speech Grammar Editor to Create Grammars
Use Conversational Grammar Builder to Create Grammars