Text-to-Speech

Historical notes

1. What three early paradigms of waveform synthesis emerged in the 1950s?
   Formant synthesis (generating artificial spectrograms, e.g. Haskins Pattern Playback, Klatt’s synthesizer and DECtalk), concatenative synthesis (splicing recorded units), and articulatory synthesis (modeling vocal tract physics).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 17, Historical notes, p. 395
2. How did concatenative synthesis develop?
   From splicing magnetic tape for phones (Harris, 1953), to diphone databases with prosodic variants and join costs (implemented decades later), to unit selection synthesis in the 1980s–90s with larger units of varying length and target costs.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 17, Historical notes, p. 396
3. What was probably the first true text-to-speech system?
   Umeda and Teranishi’s system (1968), which included a parser assigning prosodic boundaries, accent and stress; earlier systems took phonemes as input.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 17, Historical notes, p. 396