Spaces in Between – Towards Ambiguity in Immersive Audio Experiences Pedro Rebelo Michael McKnight Sonic Arts Research Centre, Queen’s University Belfast p.rebelo@qub.ac.uk Sonic Arts Research Centre, Queen’s University Belfast mmcknight13@qub.ac.uk ABSTRACT Immersive audio experiences position the listener at the centre of a sound field created by loudspeakers or head- phones often attempting to achieve a sense of belonging and embodiment. The focus on placing the listener amongst multiple sound sources with individual localisa- tion and propagation patterns goes somewhat towards recreating everyday like listening experiences, which typically involve hundreds if not thousands of sources in complex acoustics perceived differently by each listener based on movement and head rotation not to mention individual hearing profiles. This paper focuses on the, as of yet, largely unexplored listening space which sits be- tween headphone and loudspeaker reproduction. This unexplored space affords multiple approaches when it comes to creating ambiguous listening situations in which there is a deliberate attempt to blur, augment or even contradict the relationship between headphone and loudspeaker sound fields. We propose a set of strategies which can be deployed in the development of Auditory Virtual Environments (AVE) aimed at creating spaces in between. The paper combines technical, design and crea- tive approaches based on recent artworks, to illustrate ways in which to explore the notion of ambiguity in im- mersive audio experiences. 1. INTRODUCTION The desire for immersive listening experiences has been driving spatial and 360° audio technologies for over a century. Binaural recordings date back as far as 1881 when the “Théâtrophone” microphone array was placed along the front of the stage at the Opera Garnier in Paris. The performance was distributed through the telephone system and listened to through a special headset. Throughout the twentieth century new technologies and practices have emerged to provide ever more accurate recording and reproduction systems. Current binaural technique uses two channels, one delivered to the left ear and the other to the right. Although an argument can be made that all stereo sound is binaural, ‘of two ears’, the term is reserved for sound that has been filtered by a combination of time, intensity and spectral cues that rep- licate the effect of human ears and body to derive exter- nalisation and localisation of sound objects. It is widely known that binaural audio works best through the use of headphones due to excellent channel separation but play- back on loudspeakers’ is possible. It requires careful consideration to eliminate crosstalk between the speakers and therefore only works if the listener is in a fixed posi- tion. This is called the equalisation zone or often referred to as the sweet spot. Outside of this small zone the spa- tial illusion will be lost, even small movements within this area or head turns as much as 10° will significantly degrade the spatial reproduction as noted by Gardner [1]. There are a number of spatial techniques that relate to loudspeaker arrays such as ambisonics and wave field synthesis, both used in immersive audio experiences. Ambisonics is a technique that emerged in the 1970s for recording or creating a sound field that envelops a listener with sound from all directions. In contrast to stereo or multichannel surround sound arrangements, where the audio is mixed to a predefined number of channels and loudspeaker placements, ambisonics is loudspeaker ag- nostic so that the 360° sound field can be decoded and reproduced by any number of loudspeakers. However, the higher number of channels used in capturing or en- coding sound the greater the number of dimensions is achieved in reproduction, which in turn is perceived as greater spatial accuracy. The most recent resurgence of virtual reality (VR) and augmented reality (AR) has led to the convergence of a number of technologies and a reimagining of their uses to create new immersive experiences. Ubiquitous VR/AR production and reproduction systems now seamlessly include head tracking, positioning and haptic sensors combined with high resolution displays, 360° cameras, new ambisonic microphones with support built into re- corders and supporting audio/visual software. Workflows for mixing and delivering spatial audio for VR experienc- es using ambisonics are beginning to establish themselves amongst industry and practitioners. For the purpose of the work discussed here, we have chosen ambisonics as our methodology for creating spatial audio. This is due not only to its popularity within the VR community but also its scalability; it can be decoded to binaural, to include headtracking, as well as large scale loudspeaker arrays. One single system can be used to derive and manage both audio streams and facilitate cross over of sound objects from headphones to loudspeaker. Copyright: © 2018 Pedro Rebelo et al. This is an open-access article dis- tributed under the terms of the Creative Commons Attribution License 3.0 Unported, which permits unrestricted use, distribu- tion, and reproduction in any medium, provided the original author and source are credited.