In this section there is a diagram showing German, French and US English voice databases. In the US English database, the structure is shown. This comprises a voice table, a voice directory for the language containing voice segments, and a system voice directory also containing voice segments.
There is a voice segment database for each language. By creating an application profile for each language, you can make the same application work with different languages.
Each voice segment has:
- A segment ID
- A description
- Digitized audio data in either compressed or uncompressed format, or both
Each voice segment is stored in a voice directory, and can also be referenced by one or more voice tables.
A voice directory has:
- A name
- A voice directory ID (retained only for compatibility with earlier releases)
The voice directory must exist before you create the voice segment.
Voice tables are optional, and you can create them at any time. Voice tables are particularly useful if you have applications that use voice segments which can be referenced using an index value, such as the spoken words for the months of the year.
A voice table has:
- A name
- A description