To begin defining a character, provide the character's name by entering it in the Name text box (32 characters maximum). Because Microsoft Agent uses the name to allow users to access the character, specify a user-friendly name. Supply a name that can be pronounced using conventional spelling, or you may disable speech input for the character. You can also specify a short optional description (256 characters) for your character in the Description text box. The server exposes what you enter in the Description text box to client applications.
For your character's spoken output, Microsoft Agent provides the choice of a synthesized, text-to-speech (TTS) voice or a voice that uses recorded sound files. If you want to use a synthesized voice, check the Use Synthesized Speech For Voice Output option. This will add a Voice page for selecting the characteristics of the voice. Choose the Voice page and use the controls on it page to select a voice, speed, and pitch of any compatible TTS engines you have installed. The range of the voice parameters you can select depends on TTS engines. If you have not yet installed a TTS engine, the Voice ID list will be empty. You must have a TTS engine installed before you define your character's voice settings in the Agent Character Editor.
If you plan to use a TTS engine for your character's output, you must also install that engine on the user's system. For more information on how to install the Lernout & HauspieŽ TruVoice engine on a user's system from a Web script, see Programming The Microsoft Agent Control. If you are using another TTS engine, check with the vendor for appropriate installation information. If you select a voice based on a particular TTS engine, but the user has a different TTS engine installed, the server attempts to fit the voice based on the characteristics you defined in the Agent Character Editor.
If you plan to use recorded sound files (.WAV files) for your character's spoken output, do not check the Use Synthesized Speech For Voice Output option. Instead, you will need to record the spoken output audio files separately and load them from your application code.
The Use Word Balloon on the character's Properties page enables you to determine whether you want to support this feature for your character. Once defined, it cannot be changed through the Microsoft Agent programming interface.
When the Use Word Balloon option is checked, you can access the Word Balloon page. The options on the Word Balloon page enable you to change the default characteristics of your word balloon. The Characters Per Line setting enables you to define the width of the balloon based on the average number of characters per line. You can set the default height based on either a fixed number of lines you want to display at once or automatically sized to the text you supply in the Speak method. You can also set whether the balloon automatically hides after a Speak method is completed and whether the balloon automatically displays or "paces" words to the character's speech output speed setting.
The Word Balloon page also enables you to set the default font for the character's word balloon and the balloon's display colors. However, be aware that users can override your word balloon font settings using the Microsoft Agent property sheet.
Each character requires a unique identifier (GUID). The server uses the identifier to differentiate characters. When you create a new character, the Editor automatically creates a new identifier for your character. You need to change a character's identifier only if you copied the character definition file of another character or if you intentionally want to differentiate a character from a former version. To change a character's identifier, click the New GUID button and the Editor will generate a new identifier.