How to specify the speech recognizer language

Learn how to select an installed language to use for speech recognition.

Here, we enumerate the languages installed on a system, identify which is the default language, and select a different language for recognition.

What you need to know

Technologies

• Windows.Media.SpeechRecognition
• Windows.Globalization

Prerequisites

This topic builds on Quickstart: Speech recognition. You should have a basic understanding of speech recognition and recognition constraints.

To complete this tutorial, have a look through these topics to get familiar with the technologies discussed here:

• Install Microsoft Visual Studio.
• Get a developer license. For instructions, see Develop using Visual Studio 2013.
• Create your first Windows Store app using C# or Visual Basic.
• Roadmap for Windows Store apps using C# or Visual Basic
• Learn about events with Events and routed events overview
• See Speech design guidelines for helpful tips on designing a useful and engaging speech-enabled app.

Instructions

Step 1: Identify the default language

A speech recognizer uses the system speech language as its default recognition language. This language is set by the user on the device Settings > System > Speech > Speech Language screen.

We identify the default language by checking the SystemSpeechLanguage static property.

var language = SpeechRecognizer.SystemSpeechLanguage; 

Step 2: Confirm an installed language

Installed languages can vary between devices. You should verify the existence of a language if you depend on it for a particular constraint.

Note  A reboot is required after a new language pack is installed. An exception with error code SPERR_NOT_FOUND (0x8004503a) is raised if the specified language is not supported or has not finished installing.

Determine the supported languages on a device by checking one of two static properties of the SpeechRecognizer class:

• SupportedTopicLanguages—The collection of Language objects used with predefined dictation and web search grammars.

• SupportedGrammarLanguages—The collection of Language objects used with a list constraint or a Speech Recognition Grammar Specification (SRGS) file.

Step 3: Specify a language

To specify a language, pass a Language object in the SpeechRecognizer constructor.

Here, we specify "en-US" as the recognition language.

var language = new Windows.Globalization.Language(“en-US”); 
var recognizer = new SpeechRecognizer(language); 

Remarks

A topic constraint can be configured by adding a SpeechRecognitionTopicConstraint to the Constraints collection of the SpeechRecognizer and then calling CompileConstraintsAsync. A SpeechRecognitionResultStatus of TopicLanguageNotSupported is returned if the recognizer is not initialized with a supported topic language.

A list constraint is configured by adding a SpeechRecognitionListConstraint to the Constraints collection of the SpeechRecognizer and then calling CompileConstraintsAsync. You cannot specify the language of a custom list directly. Instead the list will be processed using the language of the recognizer.

An SRGS grammar is an open-standard XML format represented by the SpeechRecognitionGrammarFileConstraint class. Unlike custom lists, you can specify the language of the grammar in the SRGS markup. CompileConstraintsAsync fails with a SpeechRecognitionResultStatus of TopicLanguageNotSupported if the recognizer is not initialized to the same language as the SRGS markup.

Related topics

Responding to speech interactions

Designers

Speech design guidelines