Internationalization of Voice Applications 

Purpose

Many companies require their voice applications to be available in more than one language. The Voice Kit satisfies these requirements.

Implementation Considerations

When you are addressing an international audience, and you want to provide the same options in each language required by your customers, the Voice Kit provides the means to do this easily. You can use recorded audio files in the languages you require, store these files in the same directory, and use one application to satisfy the language needs. You can use a single, default Voice gateway with several language options, which enables you to manage several language versions inexpensively.

Integration

You can adapt an existing voice-enabled application into a multiple language application, using the default voice gateway and recorded prompts, as well as developing new applications for this purpose. A few concepts need to be considered for multiple language development and implementation. First, you want to keep the program consistent regardless of the language your caller chooses.

Features

Internationalization provides these capabilities.

●      Multilingual alternative texts in Speak items

You can add tabs for each language key you have defined and create a separate set of speak items for each. That set of speak items is used only when that language key is activated at runtime.

More information: Adding Another Language to Voice Applications

●      Multiple input phrases for list-of-choices grammars

If you use a grammar that accepts user-defined words or phrases (such as a menu), you must define versions of those phrases or utterances in different languages. You can define these utterances in the dialog for creating these grammars in a field for each configured locale. Jumpwords are implemented in a similar fashion.

●      Languages and componentization

Since each voice application and each voice component have their own set of configured language keys, it is possible for an application to use a voice component that does not have the same language keys. For example, the application uses a French male voice actor named Jacques, and the component uses a female French Canadian actress named Gina. If a mismatch occurs when a component is called from a parent, the runtime is designed to use the voice defined in the component that most closely matches the voice set in the parent. If there is no voice that matches the language, then the runtime uses the component's default language.

Constraints

If you intend to run a multilingual application on a monolingual gateway, then you have the following constraints.

●      All prompts must be coded as audio files. Text to speech (TTS) is not possible.

●      Only DTMF (touch-tones) can be used in grammars. Spoken phrases are not recognized.