Spoken Language Generation and Understanding

Volume 59 of the series NATO Advanced Study Institutes Series pp 397-412

An Over-View of Speech Synthesis

  • Jean-Sylvain LienardAffiliated withLaboratoire d’Informatique pour la Mécanique et les Sciences de l’Ingénieur (L.I.M.S.I.-C.N.R.S.)

* Final gross prices may vary according to local VAT.

Get Access


This paper gives a survey of the main present day approaches to speech synthesis. Several processing levels are envisioned, according to the nature of the input: acoustical synthesis, phonetic and prosodic command, phonetic and prosodic transcription of written text, as well as phonetic message generation from a conceptual formula, and synthesis by words. Intelligibility and quality of synthetic voices cannot be estimated without referring to the specific constraints and problems encountered at each processing level.