Real-Time Auralization of Large Spaces and its Effects on Singing Performance

Many studies have examined how singers adapt their vocal delivery to the acoustic properties of different performance spaces. However, most virtual acoustic simulations rely on binaural rendering rather than multi-speaker systems. This thesis explores whether a similar adaptive behavior occurs when reverberant environments are reproduced through real-time auralization. A real-time convolution system was developed within the Audio Space Lab (ASL) at the Department of Energy, Politecnico di Torino, using 3rd Order Ambisonics impulse responses from four contemporary churches in Milan, Italy. The laboratory features a 16-loudspeaker spherical array, allowing singers to experience immersive acoustic auralizations. Six professional singers originally performed in these churches, where the acoustic properties were measured following ISO 3382-1, and their vocal delivery was analyzed. A questionnaire has also been used to collect feedback on vocal comfort while performing in each on those environments. In the ASL, amateur singers were invited to perform freely in the auralized environments and rate their preferences through a subjective questionnaire similar to the one provided in the actual churches. Also, their vocal performances were recorded while singing a well-known, simple melody. Finally, vocal parameters were calculated and compared to the ones from the professional singers, along with the preference charts obtained from both groups. This study aimed to evaluate the feasibility of real-time convolution within the ASL and investigate how amateur singers perceive and adapt to auralized spaces. The findings provide insight into the effectiveness of multi-speaker auralization in vocal performance research.