ABSTRACT
This final chapter provided the third selected area of speech technology, speech synthesis, which forms a core module in larger text-to-speech systems. After a brief introduction to speech synthesis, where major steps in a typical system were outlined, basic appoaches and methods for speech synthesis were presented. We covered three main methods — articulatory method, spectral method, and waveform method. A recent trend in speech synthesis has centered on the use of data-driven, optimization principles to automatically select speech units in waveform-based concatenation methods. We provided a case study to elaborate on such a new trend. We also covered intonation, text pre-processing, and evaluation aspects of speech synthesis and text-to-speech.
