|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||
TEXT TO SPEECH Applications that have information not known in advance (such as current news events), or have a very large vocabulary (such as a dealer locater with 10,000+ dealer names and addresses) require text to speech. This is due to the fact that the audio either cannot be recorded in advance (unless you can predict the future!), or is impractical to record since it is so large. DAC Systems applications use text to speech only where necessary (as described above). Human recorded audio is used for the remaining portion of the application. For example, in an auto battery dealer locator application, regular professionally recorded audio is used throughout the application except where the dealer names and addresses are presented. This section uses text to speech. Without text to speech it would be unwieldy to record 55,000 dealer names and addresses. The sound of Text to Speech has greatly improved. Prior to 1999, it was hard to understand, and had a highly robotic sound. In our early systems we implemented text to speech only for those applications where callers would use it on a daily or weekly basis, such as for field service where technicians would receive work orders over the phone and text to speech would present organization names, part names, etc. Today, Text to Speech is natural sounding and barely sounds robotic. It is available in male and female voices and in many languages.