On the Enhanced Acoustic Design of the Indoor Environments: Correspondence of Perceptual Quantities between Real and Simulated Sound Fields

In the design of indoor spaces where speech communication takes on a central role (e.g., classrooms, conference rooms…), the influence of the sound environment on the occupants’ performance needs to be addressed. In order to guarantee a comfortable communication experience, the acoustic design of such spaces have to ensure, beside a high percentage of correctly heard words (i.e., intelligibility), also a minimal effort in the speech reception process. An effortful listening, as produced for instance by the presence of background noise and/or reverberation, requires the involvement of an increased amount of cognitive resources, and, if sustained over a prolonged period may compromise occupants’ learning and cognitive achievements. The present study specifically addresses the issue of an improved acoustical design of the rooms for speech, based on both intelligibility and listening effort results. For the scope, the correspondence between the results of speech-in-noise tests presented within an existing university classroom and via headphones, using auralized signals obtained from acoustic simulations of the same environment is investigated. In fact, whereas the reliability of acoustical simulations has been widely confirmed as regards the predicted objective acoustic parameters, and the speech intelligibility data, to date an ecological validation of the listening effort metrics is still lacking, and is needed to understand how well results obtained by virtual acoustics predict the everyday realistic communication situations. Simulations of a university classroom with a volume of 198 m3 and acoustical treatment on a lateral wall were carried out with an acoustical CAD software. In order to obtain realistic simulations, the model was calibrated with octave-band field measurements of reverberation and clarity parameters. Binaural impulse responses (BRIRs) were calculated in two listening positions within the classroom, and convolved with anechoic speech and stationary noise, to obtain the auralized stimuli for the speech-in-noise tests. Speech and noise sound pressure levels were calibrated with reference to the values measured during the in situ tests. Consonant confusion tests (Diagnostic Rhyme Tests) in the Italian language were proposed to normal-hearing young adults. The tests were firstly presented in the real classroom, and then in laboratory conditions, via headphones. During the experiments, data on the number of words correctly recognized, auditory response times (RT, behavioral measure of listening effort) and subjective ratings of listening effort (LE) were collected. The statistical analyses showed that both IS and RT data in auralized conditions matched the corresponding results obtained with in situ testing; the RT metric showed a greater sensitivity than IS, being able to discriminate between the listening position within the classroom. As concern LE, the results were found to depend on the mode of presentation, suggesting that beside the auditory stimulus other factors (e.g., attention, experimental setup…) affect the subjective response. Based on the comparison of the results in auralized and in situ conditions, it can be said that the auralization techniques allow to recreate perceptually equivalent environments as regards the IS and RT measures, and that the integration of the two metrics would be of benefit the acoustical design process.