Comparison between Simulated and in-situ Measured Speech Intelligibility in the Multilingual Context of the Free University of Bozen-Bolzano

Classrooms acoustics can affect studentś speech intelligibility and learning performance depending on its background noise level and/or reverberation. Speech intelligibility is usually assessed in real classrooms through a subjective approach, by performing speech intelligibility tests, or through an objective approach, by evaluating speech transmission index (STI) from impulse response, speech and noise level measurements. An acoustic simulation technique makes it possible to assess acoustical conditions for speech reception in virtual environments, thus allowing for predicting intelligibility before a classroom is built or renovated. However, in order to obtain reliable results, the simulation model needs to be calibrated and validated with in-situ measurements. The aim of this work is to compare tests performed in-situ on a group of people, with tests performed on the same people by reproducing the auralized test signal through headphones, in terms of intelligibility scores (IS), response times (RT), listening efficiency values (DE) and related STI values. Simulations have been carried out using the room acoustic software Odeon version 14.01. The investigation focused on a university classroom, which is part of the Classroom Spaces Living Lab of the Free University of Bozen-Bolzano, currently equipped with devices for monitoring energy and indoor comfort conditions, as well as detailed external weather conditions. By exploiting the bilingual context in South Tyrol, Diagnostic Rhyme Tests (DRT) in the Italian language were administered to both Italian and German native speaker students, the latter with an Italian level at least equal to B2, according to the common European framework of reference for languages. In this way, speech reception performance of the two groups has been investigated and compared.