Melodic Analysis of Speech (MAS): Phonetics of Intonation

Abstract

The acoustic-perception-based Melodic Analysis of Speech method (MAS) is a formal, objective and complete method that encompasses the most relevant aspects to be considered when analysing intonation. It provides researchers who wish to study intonation with criteria for establishing a corpus, the identification of melodic units, the extraction and relevance of acoustic data, its standardisation and the representation and interpretation of graphs, as well as the execution of perception tests and validation of the results obtained. It also offers a way of interpreting the melodic data based on three levels of intonation: one which provides an explanation of the structure of speech, the prelinguistic level, the linguistic level and the paralinguistic level, giving it significance with regard to the speaker's intentions or the communicative context. In addition, a phonological description based on the Autosegmental-Metrical model (AM) and its labelling convention known as Tones and Break Indices (ToBI), which is complementary and totally compatible with the phonetic description of MAS, can be added. The method is constantly advancing, being reformulated and added to as new requirements arise resulting from its use by researchers in the description of different languages, interlanguages and dialects. It is these developments, which have occurred over more than a decade and particularly in the two years, together with the basic pillars of the method, which are presented in this article.