A simplified engineering super-set of phonetic notation for European multi-language text to speech development

In this paper, an approach of designing a simplified engineering super-set of phonetic notation (abbreviated SESSet) for European multi-language Text-to-Speech (TTS) is presented. The super-set's representation is engineering encoding of IPA. It covers 8 languages that belong to 3 language groups of Indo-European family. For engineering purpose, the proposed SESSet is designed as machine readable, easy typing-in and convenient for programming. Every phonemic notation takes two 7-bit ASCII letters. The whole SESSet contains 91 phonemes that are able to describe 342 phonemes of all 8 languages. The sharing factor of phonemic notation for the 8 languages is 3.7, which means significant resource saving as the SESSet is adopted for engineering realization of multi-language TTS.