On the use of the Italian matrixed word tests in room acoustics for evaluating speech reception

This work aims to investigate the use of the newly developed Italian matrixed word test for room acoustics applications and assess the speech reception performance in complex acoustical environments. The detrimental effects of background noise and reverberation were taken into account and the listening experience was evaluated with both intelligibility scores and response times; the combined metric “listening efficiency” was also calculated, as the ratio of accuracy and latency measures. Twenty-one native Italian speakers were involved in the experiment, all of them reporting normal hearing. The listening tests were proposed in the closed-set format. SNR and reverberation were combined to create four different listening conditions, presented in a silent room via a three-dimensional audio rendering system. A stationary speech shaped noise and a non-stationary but continuous (ICRA) noise were used as maskers, aiming to cover different aspects of speech reception in noise (energetic masking and listening into the gaps). The results show that when reverberation is added to non-stationary noise, the fluctuating masking benefit is substantially reduced; the differences in accuracy and cognitive load that are present between the noises in anechoic conditions, are absent in a reverberant condition with T=0.94 s. A too long reverberation acts differently upon speech reception depending on the background noise, whereas a noise level increase of 3 dB similarly impairs the accuracy performance for both background noises.