Parametric Evaluation of Ensemble Vocal Performance Using an Immersive Network Music Performance Audio System

Abstract

This paper describes an immersive audio Network Music Performance (NMP) system designed for group singing. A prototype of this design (audio only) was deployed to ten singers across Europe, who participated in a duet vocal performance study, operating the system from their home networks. Parametric evaluation of these vocal performances was conducted in order to provide characterization of musical interactivity between performers and explore the challenges and opportunities presented for immersive audio NMP systems in practical usecase settings. Results demonstrate that it is possible to achieve performance that conforms to expectations of live interactivity and estimate the conditions under which this may be achieved. Significant effect of latency, and in one case virtual room "type," is observed across performances. Informal questionnaire responses present discussion of the potential for virtual acoustics and latency to impact the perceptual experience of networked performers.

Evaluation of Metaverse Music Performance With BBC Maida Vale Recording Studios

“…Using immersive audio rendering on top of NMP audio streaming is another method of providing a shared virtual space for online performance, here through interactive auditory display and interface. This may be implemented with no visual display and interface [8,24,25,26], or alongside telepresence [23,27,28,29] or VR/XR [9,21,23] visual display and interface. This is particularly relevant to VR/XR NMP with respect to the importance of coherence between audio and visual display and interface.…”

“…Ambisonic encoding of direct sound is used to spatialise sources in several cases [24,27,28]. Other Ambisonic systems provide virtual room simulation through convolution with measured Spatial Impulse Response (SIR) [8,23,25], or by modelling of acoustics via analysis of IR capture [29]. In other cases a hybrid approach is used, such as rendering direct sound using Ambisonic encoding methods, and rendering reverberation through SIR convolution [9].…”

“…Evaluation of immersive audio NMP over WANs provided comparison of performance synchrony between duet singers in different virtual rooms, and observed more stable tempi ratio in a medium hall simulation compared to studio booth and cathedral alternatives [8]. With respect to spatialisation factors, one study conducted with percussive instrumentation suggests that source panning to provide seperation may influence better performance synchrony in NMP contexts [34].…”

“…The introduction of new modality presented by VR/XR systems has the potential for interplay with latency perception. This is true with respect to visual performance timing cues (which are expected to influence performance timing in live [5], video [6] and XR [7] cases), and with respect to the potential for virtual acoustics to affect collaborative music-making experiences [8]. As such, further exploration of the NMP experience and effect of latency is prudent in this context.…”