DIACEX will identify foreign accent (FA) cues which most contribute to adverse communicative effects. In the first stage of the project, this will be achieved by implementing a new speech manipulation technique based on speech synthesis, which will be evaluated with respect to two existing techniques, splicing and language switching by bilinguals. Blingual synthesis promises to permit controlled manipulation of the degree of FA while minimising irrelevant variation, thus enabling more robust FA evaluations. Synthetic foreign-accented English will be used in subsequent stages of the project which use perceptual experiments to evaluate the isolated features.
Apart from providing a better scientific understanding of FA, we anticipate that the findings of DIACEX will be applicable in areas such as speech technology (e.g. adaptation to interlocutors), language learning (e.g. automatic pronunciation evaluation) and experimental methods for speech perception research (e.g. controlled voice manipulation).