Allophonic variation in spoken and sung speech

Abstract

Research suggests that sung and spoken speech differ systematically in terms of laryngeal, intonational, rhythmic, and articulatory performance. In the present study we examined control of pronunciation rules [Ladefoged, ACourseinPhonetics (HBJ, 1993)] in singers. Three doctoral students of vocal performance sang and then spoke a set of sentences. Simple melodies were composed to fit the rhythmic figure of each sentence. The sentences contained five examples of each of 24 pronunciation rules of English. Recordings were transcribed, rhythm and stress patterns noted, and acoustic analysis was performed (GW Soundscope) to ascertain the degree to which each pronunciation rule was/was not implemented in sung versus spoken production. Results indicate that singing induces both systematic and nonsystematic changes in pronunciation across singers which can be related to the more stereotyped production strategies associated with singing.