Methods and Tools for Prosodic Analysis of a Spoken Italian Corpus

In the last few years, a number of actions has been carried out in Italy with the goal of collecting, annotating and making available a considerable amount of data of spoken Italian varieties. After a first phase, in which the AVIP corpus has been collected and transcribed at both segmental and suprasegmental levels, now research efforts have been concentrating on corpus analysis, starting from two preliminary yet crucial aspect, namely: a) developing strategies and software tools for controlling the semantic coherence of the AVIP database; and b) designing a DBMS scheme for allowing easy access to the data and for rendering the results of the online queries in a user-friendly manner, also by means of special graphical interfaces. In this paper both aspects are presented and discussed, focussing on the prosodic analysis of the database, in terms of the methodologies followed in the intonation labelling phase as well as the consequent strategies adopted in the implemetation of software tools for prosodic analysis.