Virtual Acoustics in Joint Performance
Virtual Acoustics in Joint Performance: Perception of Room Acoustic Variations in Ensemble Performance
Project Report
Published: April 14, 2025
Authors
Ying-Ying Zhang (McGill University) and Jithin Thilakan (Hochschüle für Musik Detmold)
Introduction
The acoustic characteristics of a performance space significantly influences musicians' auditory experience, influencing the ability to interact and synchronize within an ensemble. Understanding how room acoustics affect musical performance is crucial for optimizing acoustic attributes of performance spaces, and supporting musicians in various performance settings. This study explores how changes in room acoustic conditions impact musicians’ awareness and perception of their surroundings, with a particular focus on ensemble performance. To achieve this, we employ virtual acoustics, a technique that allows precise manipulation of room acoustic properties through the process of auralization. By systematically adjusting acoustic characteristics of performance space, this pilot study aims to assess musicians' sensitivity to environmental changes, determine preferred acoustic conditions, and investigate the role of acoustic feedback in performance. Through this approach, it seeks to offer insights into the perceptual and cognitive effects of varying acoustic conditions on ensemble performance and musical blending.
Virtual acoustic environments involved in this study are achieved through a process called ‘auralization’: when an audio signal is “processed to contain features of a simulated acoustical space” [1]. The term auralization is introduced, by analogy with visualization, to describe the process of rendering audible sound fields. Mendel Kleiner, who coined this term defines it as follows; Auralization is the process of rendering audible, by physical or mathematical modeling, the sound field of a source in a space, in such a way as to simulate the binaural listening experience at a given position in the modeled space [2]. Auralization involves the use of measured, synthesized, or simulated numerical data [3] to generate virtual acoustic environments. The process consists of several stages, including the capture of the sound source signal, its convolution with the Room Impulse Response (RIR), and subsequent sound field synthesis or reproduction. The RIRs, which represent the transfer function from the sound source to the receiver, detail how the acoustic environment modifies the signal as it travels from the source to the receiver.
Auralization can be used to modify the acoustic properties of an enclosed environment, enabling the tuning of acoustic features of the room to meet specific requirements. This can be achieved through real-time auralization by utilizing a loudspeaker array designed for sound field synthesis. This approach also allows for the recreation of virtual environments over multi-channel speaker systems, enabling researchers to replicate the acoustic characteristics of one space and overlay them onto another. Also referred to as active acoustics systems, the resulting combination allows for “the adjustability of the virtual environment to suit the needs of the music and the performer” [4]. These reproduction systems have the capacity to perceptually augment the sonic environment of a real-time performance, for example by making a relatively small room sound like a concert hall.
Virtual Acoustics and the MMR
Figure 1a: The MMR shown with all banners and curtains deployed to limit its reverberation time.
Figure 1b: MMR with passive absorbers lifted and/or retracted, showing custom diffusion panels installed in the space.
The Multimedia Room (MMR) is a research and performance space located in McGill University’s Schulich School of Music. The space features acoustic treatment designed by the architectural acoustics firm Kirkegaard to minimize the reflections and acoustic responses of the physical room. Specially made diffusers break up sound waves as they encounter surfaces, scattering them so that they no longer have a uniform time response. Further acoustic absorption can be achieved through the deployment of banners and curtains, limiting the space’s early reflections and reverberation time. Early reflections are sound waves that reach the listener shortly after the direct sound, typically within the first 50–80 milliseconds, and represent the initial interaction of sound with the room’s surfaces, most notably the floor, ceiling, and walls, which help perceptually define an enclosed environment. The reverberation time is the time it takes for the sound to decay by 60 dB in an acoustic environment after the sound source has stopped. When combined with the virtual acoustics system, the room’s inherent characteristics can therefore be significantly attenuated in favor of simulated spaces.
Aside from its passive acoustic treatments, the MMR is equipped with a Meyer Constellation system for active acoustics. This is a proprietary virtual acoustic system that is highly calibrated to every space in which it is installed. Because of a prolonged testing and adjustment period, the Constellation is able to create a variety of room responses in real time without the risk of feedback. As the MMR also features passive acoustic treatments, these different auralizations must take into consideration the relationship between feedback suppression and fixed passive acoustic positions in the space. Once chosen, these auralization settings can then be “tweaked” by changing the ratio of early reflections to reverb tail, adjustment of tail length, as well as the application of high and low-pass filters.
Schulich Singers and Virtual Room Perception
This preliminary experiment was conducted as part of a series of experiments on choral blend held at the MMR. Findings from this experiment were used to inform virtual acoustics decisions We invited the Schulich Singers, a McGill University choir, to conduct a typical rehearsal in the MMR. While they rehearsed, and without their knowledge, we changed the acoustic settings of the space. The time intervals of these changes were randomized but followed a pattern of increasing reverberation tail length. We provided four different room acoustic settings, which are listed in the table below. Their corresponding reverberation times, assessed using the T30 parameter (RT60 estimated from a 30 dB sound decay), are also included in the table.
Condition A
Constellation Setting: No Simulated Acoustics
Description of Reverb Characteristics: No active reverb. Passive banners at 50% (more than 0.5 seconds)
Condition B
Constellation Setting: Percussion
Description of Reverb Characteristics: Short RT (Approximately 1.2 seconds)
Condition C
Constellation Setting: Concert Hall
Description of Reverb Characteristics : Second longest RT (more than 1.5 seconds)
Condition D
Constellation Setting: Cathedral
Description of Reverb Characteristics: Longest RT (more than 2.5 seconds)
Figure 2: The Schulich Singers in rehearsal in the MMR. An informal experiment was conducted on their perception of the room response and their preference.
As part of our investigation, we asked singers if they were aware of environmental changes during their performance. Out of 20 participants who responded to the exit survey, 10 of them reported that they were not aware of two or more of the reverb changes, and 7 reported that they were not aware of any acoustic changes at all.
After the initial portion of the experiment, we also returned to each condition and, this time informing them of each change, had them perform a long-held note with a defined cut, and invited them to pay attention to the reverb after the note was cut. We then asked them to rank these conditions in order of preference. Cathedral and No simulated acoustics were both ranked as the second most preferred environment. It is interesting to note that these were our two extremes, as No simulated acoustics had the shortest reverberation time—it being the natural condition of the room— and the Constellation’s “Cathedral” preset boasts its longest reverberation time. These settings seemed to be polarizing in general among our subject pool, as they also were our least preferred (No simulated acoustics, at 10 participants) and second least preferred (Cathedral, at 7 participants) acoustic settings.
Figure 3: Acoustic settings ranked by preference and the number of subjects who gave each ranking.
Around half of our participants most preferred the system’s Concert Hall setting which, with a reverberation time of around 1.5 seconds, most closely resembles a mid-sized performance hall. In fact, an overwhelming majority ranked this setting within their top two, while Percussion had the greatest number of participants calling it second best.
The results of this preliminary study gave us the opportunity to tweak some of the Constellation factory presets before we continued with further experiments. They also showcase some of the issues we experience when conducting perception research with musicians. That half of our participants were unaware of two or more reverb changes shows that, while completing the cognitively complex tasks of singing and rehearsing, singers may not be immediately aware of uncued changes in their acoustic environments. This may have to do with the cognitive load in performance and rehearsal, but they may also experience a very high running reverberance ratio. Running reverberance is the ratio of a sound source’s direct sound against its room reflections, which we could presume would be a significant value for choir singers.
As a singer’s direct sound source is a physiological instrument located inside their body, the way they experience direct sound is quite different from instruments that are separate from and held further away from the body. For example, an electric guitarist’s guitar amplifier may be several feet from the performer, who may hear its direct sound blended with acoustic reflections from the floor, ceiling, walls, and other surfaces. In this way, understanding the relationship between how a singer experiences both the ratio of direct to indirect sound and how it differs from other instrumentalists is critical to understanding how room acoustics affect their performance and ability to blend with other musicians.
Works Cited
[1] Lokki, T., & Savioja, L. (2008). Virtual acoustics. In Handbook of Signal Processing in Acoustics (pp. 761-771). New York, NY: Springer New York.
[2] Kleiner, M., Dalenbäck, B. I., & Svensson, P. (1993). Auralization-an overview. Journal of the Audio Engineering Society, 41(11), 861-875.
[3] Vorländer, M. (2020). Auralization: fundamentals of acoustics, modelling, simulation, algorithms and acoustic virtual reality. Springer International Publishing.
[4] Woszczyk, W., Beghin, T., de Francisco, M., & Ko, D. (2009, October). Recording multichannel sound within virtual acoustics. In Audio Engineering Society Convention 127. Audio Engineering Society.