This paper describes several aspects of the development of grapheme-to-phoneme systems for 6 different languages. We report on the lexical data bases we are using for these developments. Our method for the automatic alignment of orthographic and phonetic transcriptions is reviewed. We describe how the method can be used to facilitate the development of pronunciation rules. Performances are given for word data and name data. In the last part of the paper we report on some orthographic/phonetic properties of the different languages using the concepts of entropy and mutual information.

Keywords: text-to-speech, lexical data bases, grapheme-to-phoneme conversion