Speech Synthesis by Rule: An Acoustic Domain Approach.

摘要

A novel approach to speech synthesis by rule is presented and evaluated. A discrete set of input symbols is converted to continuous control signals for driving a serial terminal analog speech synthesizer. The synthesizer converts the control signals to a continuous acoustic output. The synthesizer was simulated on a digital computer. The input set is a linguistic description of the utterance. It includes information about which phonemes are present, which vowels are stressed, and where sentence and word boundaries and pauses occur.