Speech Synthesis

Three steps to synthesis speech from text.  First break down the text into words. It involve reducing ambiguity by normalizing elements like numbers, dates, times, abbreviations and special characters etc into words. Try w second step is to break down words into phonemes (distinct unit of sound). The third step is to synthesize sound. 

There are 3 types of technics to generate sound. Concatenative is to string up pre-recorded sound. Formant is to add a combination of synthesized frequencies and shapes using model. Articulatory synthesis models human vocal tract to generate sound.