A form of speech synthesis
in which sentences are formed by concatenating
pre-recorded words from a database
. Unlike a TTS
system, which uses speech synthesis to form spontaneous sentences and/or phrases based on human phonetics, a voice response system operates with a limited vocabulary in situations where the sentences and/or phrases that are formed follow a strict predetermined pattern. For example, a train station may use a voice response system to notify passengers of schedule information or a train s status. The synthesized speech is created from a pool of words that are strung together based on the input of a human operator, and the pool only contains a limited number of words as there are a limited number of combinations of words that are necessary for the train station’s purpose; financial institutions also use voice response systems to aid customers in getting account information over the telephone. The same principle applies here — since the financial institution only needs to provide the caller with a limited amount of information, it does not need to be able to generate spontaneous sentences in response to customer inquiries.
Contrast with TTS.