| |
Phrase Splicing
Phrase Splicing is the functionality of the SVOX engine to mix prerecorded phrases with the speech portions generated with text-to-speech. This capability is recognized by our customers as one of the most powerful and useful features of the SVOX technology.
For any part of an input sentence, an audio file can be specified for use instead of text-to-speech. The engine will make sure that the prosody of the text-to-speech parts of the input sentence are processed in a way that smoothens the transition between recorded sentence frames and text-to-speech inclusions.
While other competing systems require extensive computing power to accomplish this process (oftentimes disabling this feature in an embedded system), SVOX phrase splicing requires the same or less processing power than full text-to-speech.
In addition, the SVOX engine can automatically map parts of the input text to a database of pre-recorded splices, making the mixing of text-to-speech with voice prompts transparent to the application.
This is particularly important if the phrase splicing technique is to be used in an environment where the source of the text input is unknown. For example in an automotive system, when input texts can be sent in an XML format over the data bus by arbitrary electronic subsystems or by mobile devices.
|
|