AES 2026 AVARIG Conference

This year marks the tenth anniversary of the biennial AES International Conference on Audio for Virtual and Augmented Reality. Starting with the inaugural event co-located with the 2016 AES Convention in Los Angeles, we’ll recap highlights from the previous five AVAR conferences to celebrate where we’ve been, where we are, and where we’re going.

The rapid mainstreaming of spatial audio has created a need for design frameworks that translate research into production-ready practice. This roundtable brings together contributing authors from Immersive Sound Volume II: The Design and Practice of Binaural and Multi-Channel Experiences for an open, cross-disciplinary conversation about where the field stands and where it is headed. Participants will draw on the text's core themes: perceptual grounding, system design, and the creative practice of building immersive sound experiences. The session is structured to encourage dialogue between contributors and attendees, surfacing points of debate, unresolved questions, and divergent perspectives across binaural and multi-channel workflows. Through case studies and cross-disciplinary perspectives, the session offers a practical roadmap for audio engineers, sound designers, and researchers navigating the evolving immersive audio industry. Attendees will leave with concrete frameworks applicable to both studio production and real-time XR deployment.

This paper presents an application-driven objective measurement framework for benchmarking spatial audio reproduction in smart glasses and extended reality (XR) headsets. Wearable XR devices render virtual spatial audio while users simultaneously perceive the physical acoustic environment, creating evaluation cIRCAM:Galleryenges distinct from conventional headphone-based playback. Existing approaches are often inconsistent, focusing on limited device classes or metrics, and do not support unified cross-device benchmarking. The proposed framework derives benchmark attributes from two application dimensions: the acoustic role of the device and the usage context. Measurements are organized into four groups: baseline playback checks, cue fidelity, sound leakage, and robustness to wearing variability. The framework adopts a system-level methodology that characterizes observable device behavior without requiring access to proprietary internal parameters, enabling reproducible cross-device comParison. An illustrative application of the framework is presented in a companion paper.

Eclipsa Audio, based on the Immersive Audio Model and Format (IAMF) specification developed by members of the Alliance for Open Media, represents an open and royalty-free approach to immersive audio creation and delivery. Eclipsa Audio provides a growing ecosystem for producing and distributing spatial audio content, with hardware integration and streaming platform support, including YouTube, actively being rolled out. This panel brings together practitioners, researchers, and engineers directly involved in the development of IAMF and Eclipsa Audio to inform the audio engineering community about the current state of the format and its evolving toolkit for immersive audio production and delivery. The presenters will discuss how the Eclipsa Audio ecosystem can continue growing in the live and interactive realms, including 360 videos, streaming, gaming and the combination of both, eg. in e-sports. Future directions for development will also include developers' perspective on how Eclipsa Audio can be embraced by interactive environments.

Augmented reality (AR) systems require listeners to wear head-worn devices (HWDs) such as headphones and head-mounted displays (HMDs), which can alter spatial hearing by modifying the acoustic cues reaching the listeners’ ears. Although acoustical and perceptual effects have been reported for isolated HWDs, most studies rely on simplified paradigms such as static sound source localisation tasks, providing limited insight into spatial perception in more ecological settings. In everyday listening, spatial perception is an active multisensory process in which listeners coordinate head and eye movements to build a stable representation of the environment, which may be disrupted by altered auditory cues. In this work, the perceptual impact of wearing HWDs was investigated using an auditory-aided visual search task with continuous tracking of head and eye movements. Multiple HWD configurations were compared, including two pairs of headphones with and without an HMD, to assess how scattering introduced by these devices affects spatial hearing in ecologically relevant AR scenarios. Results showed small but statistically significant effects of HWDs on exploration behaviour, primarily reflected in increased eye-movement search time, while head movements were only marginally affected. Across conditions, eye movements preceded head movements, with subtle differences in movement onset timing but limited impact on overall search performance. Overall, the findings indicate that HWDs introduce measurable but moderate changes in eye–head coordination, while largely preserving spatial search performance in ecologically valid listening conditions.

Capturing acoustic scenes with head-worn microphone arrays is cIRCAM:Galleryenging due to a limited number of sensors and constrained placement flexibility. Nevertheless, binaural reproduction based on these arrays has been recently proposed using binaural signal matching (BSM), showing high robustness and computational efficiency, but inferior performance compared to the more computationally complex signal-dependent methods, in particular at low reverberation conditions. To address this gap, this paper investigates the Field-of-View informed Binaural Signal Matching (FoVi-BSM) method for far-field sources. FoVi-BSM incorporates a diagonal spatial weight matrix directly into the error formulation, redefining the filter solution to prioritize spectral and spatial fidelity within a predefined FoV, showing performance comparable to signal dependent methods but with the same computation complexity as BSM. The performance of the method is evaluated through objective single-source anechoic simulations, multi-source reverberant Monte Carlo simulations, and a subjective MUSHRA listening test. Results demonstrate that prioritizing the FoV improves rendering accuracy within the targeted region over the standard baseline BSM method, achieving perceptual quality comparable to signal-dependent parametric methods while maintaining baseline-level performance outside the FoV.

The rapid growth of extended reality (XR) technologies has highlighted the critical role of immersive audio in enabling natural, effective, and socially meaningful interactions. While visual realism has traditionally dominated XR development, auditory perception remains a key driver of presence, communication, and behavioural response. The EU-funded SONICOM project (GA 101017743), ended in June 2026, aimed at addressing this cIRCAM:Galleryenge by advancing the scientific, technological, and perceptual foundations of immersive audio for augmented and virtual reality, with a particular focus on personalisation, interaction, and reproducibility. This workshop presents a comprehensive overview of SONICOM’s outcomes, structured around its core Work Packages (WPs), and delivered directly by WP leaders. It is designed as an interactive session, where each WP will briefly introduce not only scientific findings, but also tangible outputs including datasets, software tools, hardware prototypes, and open research resources. In addition, SONICOM will be represented across the broader conference through multiple contributions, including demonstrations, papers, posters, and tutorials. This workshop will explicitly connect to these activities, guiding participants towards relevant sessions where specific aspects of the project can be explored in greater depth. The first part of the workshop will focus on WP1 (Immersion), which has advanced the modelling of both listeners and environments for high-fidelity spatial audio rendering. Key contributions include novel approaches to Head-Related Transfer Function (HRTF) personalisation, combining physical modelling, machine learning, and perceptual validation. The SONICOM HRTF dataset, comprising over 300 subjects with associated 3D morphological data, will be presented as a major open resource for the community, supporting reproducible research and data-driven approaches to auditory modelling. Additional outputs include parametric pinna models, numerical simulation frameworks, and techniques for real/virtual acoustic blending, addressing one of the central cIRCAM:Galleryenges of audio augmented reality. The workshop will then move to WP2 (Interaction), which investigates how spatial audio influences human behaviour, perception, and social interaction. This includes studies on auditory proxemics, adaptation to non-individual HRTFs, and speech intelligibility in complex environments. WP2 has generated a range of experimental paradigms and datasets linking acoustic features to behavioural and cognitive responses, providing new insights into how sound shapes social dynamics in XR. In WP3 (Integration), research outcomes from WP1 and WP2 are consolidated into a unified technological framework. Central to this effort is the Binaural Rendering Toolkit (BRT), a flexible and extensible software platform enabling real-time, personalised spatial audio rendering. The workshop will briefly present the architecture and capabilities of the BRT, as well as its integration with novel hardware prototypes, including self-personalising headphones equipped with multiple microphones and loudspeakers. WP4 (Experience) extends these technologies into ecologically valid scenarios, evaluating immersive audio in realistic use cases ranging from social communication to professional XR applications. The workshop will present the design of these scenarios and early insights into how personalised audio affects user experience, task performance, and perceptual plausibility. The final technical section will focus on WP5 (Beyond), which ensures long-term impact through open science and community engagement. SONICOM has contributed to and expanded key infrastructures such as the Auditory Modelling Toolbox and SOFA standard, while also developing its own ecosystem linking datasets, models, and tools. A key highlight is the Listener Acoustic Personalisation (LAP) cIRCAM:Galleryenge, which engaged the international community in advancing data-driven approaches to HRTF personalisation. Throughout the workshop, emphasis will be placed on open outputs and reuse: participants will be guided on how to access SONICOM datasets, software, and models, and how these can be integrated into their own research and development pipelines. The session will conclude with a panel discussion involving all WP leaders, focusing on future cIRCAM:Galleryenges in immersive audio, including scalability of personalisation, as well as integration with emerging AI technologies.

Integrating microphone arrays into head-worn devices, such as augmented reality (AR) and virtual reality (VR) headsets, as well as hearing aids, has become increasingly popular for capturing and reproducing acoustic scenes. A common requirement in many such systems is a dense set of array transfer functions~(ATFs). However, dense ATFs are cumbersome to measure, and practical setups commonly yield sparse grids rather than the uniform dense sampling often required. This motivates the use of interpolation to reconstruct dense ATF sets from sparse measurements. This paper evaluates spherical harmonics and natural-neighbor interpolation, each combined with onset-based time-alignment and post-interpolation magnitude correction, for a head-worn array across sampling densities. To examine how interpolation errors propagate to binaural rendering, the interpolated ATFs are substituted into two recent filter design methods: signal-independent binaural signal matching (BSM) and a signal-dependent method combining COMPASS parametric spatial coding with BSM (COM). Results show that BSM remains largely robust to interpolation errors, while COM substantially degrades under sparse sampling conditions with errors comparable to BSM at the lowest density, but achieves considerably lower errors than BSM as the sampling grid density increases. This is because BSM averages errors across all steering directions, while COM relies on individual steering vectors for source-directed beamforming.

Come see the newest developements of our sponsors

We propose a baffleless circular array of radially outward facing cardioid microphones that produces standard ambisonic signals. The array produces an $N$th-order ambisonic signal from $2N+1$ microphones. It can be seen as a baffleless variant of the previously proposed equatorial microphone array, which uses a rigid spherical baffle. The simplicity of the microphone arrangement comes at the price of not being able to extract certain spatial information from the captured sound field. The ambisonic output signal represents a horizontal projection of the captured sound field. We demonstrate that interaural elevation cues are maintained when binaural rendering is performed despite the horizontal projection.

Demonstration of the Binaural Rendering Toolbox (BRT) capabilities through the BeRTA standalone application

A standalone VR experience for Meta Quest 3 showcasing practical applications of the Binaural Rendering Toolbox.

Immersive binaural audio guide around Notre-Dame Cathedral. Bring your phone if you can. Headphones supplied if needed.

A VR demo showcasing the impact of various low level immersive features such as (HRTF individualization, reverberation, source directivity) on the quality of the experience in a remote music lesson

Interactive demo showcasing spatialized audio rendering for multiple participants over a teleconference tool.

This work reports an exploratory perceptual assessment of a binaurally rendered virtual choir, focusing on how spatial, audiovisual, and performative (including conductor-related) dimensions shape perceived plausibility of the depicted musical event. A four-part a cappella passage from Purcell's "Dido and Aeneas" was recorded at six positions, five within the choir and one at the conductor, using first-order Ambisonic recorders in a controlled acoustic environment. These recording positions were later matched to six teleportation points in a virtual performance space. The experience was reproduced over open-back headphones via a dual-engine system coupling Unity (visuals, interaction) with REAPER (audio rendering) via OSC; head-tracked scene rotation and algorithmically added late reverberation were applied in real time, with binaural decoding using an individualized KEMAR HRTF set. 24 participants with choral backgrounds wore a Meta Quest 3 headset and navigated among the teleportation points. Seven items were rated inside the virtual environment on 5-point scales: source--position connection, conductor--music temporal coherence, voice-part localizability, reverberation--room congruence, co-presence, conductor naturalness, and event plausibility. Results suggested an uneven perceptual profile where audio-spatial aspects of the scene were experienced as comparatively coherent and the depicted musical event as plausible, whereas the animated conductor was the clearest limitation. The match between reverberation and visible room size was rated as moderate, and co-presence with the choir and conductor varied widely across participants.

Indoor navigation remains a significant cIRCAM:Galleryenge for blind and visually impaired (BVI) individuals, particularly in complex environments where conventional assistive solutions rely on visual interfaces or dedicated infrastructure. Spatial audio offers a non-visual alternative by encoding navigational information as virtual sound sources that can be perceived and followed in three-dimensional space. This work presents a demonstration of an indoor navigation system developed within the SONICOM project, using spatial audio rendering on standard smartphones with open-ear headphones. The system provides continuous directional guidance through virtual auditory targets, updated in real time based on user motion and orientation, without requiring additional physical infrastructure. A core research objective of SONICOM has been to improve the perceptual realism of spatial audio and to investigate its impact on navigation performance. In this context, controlled experiments were conducted to evaluate how different rendering parameters—such as head-related transfer function individualization and rendering fidelity—affect user performance, confidence, and cognitive load during navigation tasks. The results of these experiments are presented in the accompanying poster. The demonstration allows participants to experience multiple audio rendering conditions, including high-fidelity spatial audio, simplified spatial cues, and non-spatialized audio. By directly comparing these conditions, participants can perceive how spatial audio realism influences orientation and movement. This work highlights the potential of spatial audio as an infrastructure-independent approach to accessible indoor navigation and provides an experiential complement to experimental findings on the role of auditory spatialization in assistive technologies.

The Vaulted Harmonies project explores the co-evolution of architecture, acoustics, and music in the cathedral Notre-Dame de Paris through immersive audiovisual experiences. Following the production of a one-hour animated feature film combining historically informed visual reconstructions and dynamic spatial audio renderings, three additional dissemination formats were developed: a dome version for planetarium diffusion, a standalone virtual reality experience, and a web-based interactive experience. While derived from the same core assets and archaeoacoustic reconstruction workflow, each platform required specific perceptual, technical, and distribution-oriented adaptations. This paper presents the shared adaptation workflow, platform-specific constraints and rendering strategies, and discusses scalable dissemination approaches for immersive heritage experiences.

We summarise recent clinical studies on neurosensory differences between humans across age and gender, in particular related to Auditory Envelopment and reproduced sound. In audio engineering, listening quality is generally explained and tested considering just snapshot (frequency domain) metrics: Frequency response, sound pressure level, distortion and direction of sound. However, the two elusive dimensions, time and change, play the most important roles. Not only in hearing, but in all our five primary senses. Human sensory physiology is summarised, along with recent studies on a particular time-domain dimension, Auditory Envelopment (AEV). Naive and professional subjects of any age and across genders, point to AEV as a coherent and universal inducer of emotion in sound reproduction. With a conference in Paris attendees are furthermore able to experience AEV at one of most conducive places for it to emerge naturally: Inside the newly renovated Notre Dame cathedral. Using the quadrant model from Nordic universities, it is discussed if audio engineering, like medicine and other “objective” sciences, may have been tending primarily to adult male needs and preferences. Our literature is abundant with investigation of frequency-domain attributes such as power, punch and mobilisation; but ignores physiological and mental effects of the time-domain modulation that happens on basically any time-scale, as we listen. It appears stonemasons centuries ago knew important things about humans and hearing that would have been lost by now, had it not been for their magnificent monuments. Because of those efforts, however, we still have guides to what sound might achieve on new platforms, when we remember to listen slowly.

Including diffraction modelling in an acoustic simulation is known to improve the plausibility of rendered room acoustics in Virtual Reality (VR). In VR, acoustic rendering only needs to satisfy the expectations raised by the visual room impression. In Augmented Reality (AR), however, the user’s natural acoustic environment provides an additional reference, which typically increases the perceptual demands. This study assesses a selection of diffraction modelling approaches in an augmented reality (AR) setting in an L-shaped corridor. The participants rated the plausibility and similarity using a paired-comParison paradigm. ComParisons were included between acoustic simulations and between simulations and a real sound source. This is, to the best of the authors’ knowledge, the first experiment investigating diffraction perception in an AR context. The results indicated that room auralisation including diffraction was rated as more plausible than auralisation without, similar to VR experiments. However, the real sound source was rated as more plausible than all of the simulations. These observations suggest that the relative performance of room acoustic modelling is perceived similarly in VR and AR experiments, but needs further improvement to be suitable for occlusion scenarios in AR, where diffraction modelling might not be the main limitation. In general, perceptually accurate acoustic modelling of a complex real environment remains a cIRCAM:Galleryenge in AR.

Auralization enables multisensory evaluation of architectural designs in Virtual Reality (VR), yet physically accurate acoustic simulations remain computationally prohibitive for interactive workflows. This study investigates efficient artificial reverberation methods as lightweight proxies for different stages of VR-based architectural design. After assessing the predictive capabilities of geometrically informed models, a hierarchical 3-Alternative-Forced-Choice listening experiment with a transferring task paradigm was conducted in VR using binaural audio. In this experiment, the measured room impulse responses of a physical space in untreated and acoustically treated conditions were compared with those from five auralization techniques. These techniques ranged from industry-standard simulations to artificial reverberators, all calibrated to the measured Energy Decay Curves. Statistical analysis revealed that the pure Image-Source Method was easily detected, likely because the late reverberation's temporal density was insufficient. Conversely, when incorporating a dense late reverberant tail, computationally efficient methods achieved perceptual comparability with high-fidelity simulations. Participants prioritized the timbral quality of late reverberation over geometric early reflections. This suggests that computationally efficient models can serve as convincing, scalable rendering tools for interactive design and presents this audiovisual VR paradigm as an ecologically valid platform for multisensory architectural assessment.

The Brittany-based consortium HDaudio3D (LabSTICC, Brest; Noise Makers, Rennes; and Feichter Audio, Lannion) has been working for five years on the development of a professional 3D audio recording solution: one offering sufficient spatial accuracy and signal-to-noise ratio, compatibility with headphones and speakers, and the ability to manipulate the audio scene in post-production. The solution is now operational and includes a 7th-order ambisonic sensor (with 480 MEMS on the surface of a 16 cm icosahedral tetrahedron), signal processing and a software suite. Pascal Rueff, a sound engineer specialising in binaural sound and François Salmon, research engineer at Noise Makers, members of the HDaudio3D consortium, will present the system and play excerpts from recordings made for various use cases.

In audio for virtual and augmented reality systems it is often necessary to measure very large data sets. For example, HRTFs from different spatial angles, or spatial impulse responses in a room as function of angle or position, or even training sets for neural networks. This can result in a large number of measurements being required to collect the dataset with adequate fidelity. Even then compromises are often necessary to make the data set manageable. Sparse sampling is a technique that can overcome these limitations. Sparse Sampling allows one to sample the signal at apparently less than the Nyquist/Shannon limit of two times the highest frequency, without losing any signal fidelity. If some loss of fidelity is allowed the signal can be sampled at an even lower average rate. How can this be? The answer is that the effective information rate is actually lower than the highest frequency. The purpose of this tutorial is to give a (mostly) non-mathematical introduction to Sparse Sampling, and its application to audio for virtual and augmented reality systems. We will examine the difference between “Sparse” and “Dense” signals and define what is meant by “rate of innovation” and see how it relates to sample rate. We will then go on to see how we can create sparse signals either via transforms or filters to provide signals that can be sample at much lower rates. We will then show how some of these methods are already used in audio, and suggest other areas of application, such as measurements in audio for virtual and augmented reality systems, and parameter sets for neural networks. The tutorial will be accessible to everyone, you will not have to be an electronic engineer to understand the principles behind this alternative approach to sampling for audio for virtual and augmented reality systems.

This study investigates whether humans can adapt to manipulated auditory distance cues in virtual environments. While adaptation to remapped auditory localization cues is well established, it remains unclear whether similar processes apply to distance perception, particularly when natural acoustic cues are systematically modified. Virtual reality (VR) systems often employ non-ecological distance laws to improve the intelligibility of distant sound sources, which can introduce conflicts between auditory and visual information. To examine perceptual adaptation under such conditions, we modified a binaural real-time room acoustic simulation engine with rendering in six-degrees-of-freedom. The manipulation consisted of holding the direct sound level constant across distance while applying a distance-dependent low-pass filter. Over four consecutive days, participants completed a training protocol combining alternating testing phases with gamified training sessions. Results show that participants successfully adapted to the altered distance cues, with most learning occurring within the first two days. Initial exposure to the manipulation severely disrupted distance perception, rendering participants unable to make reliable judgments. However, following training, perceived distance functions approached veridical performance for far distances, exhibiting slopes close to unity. In contrast, judgments at close distances remained highly variable, suggesting that the available spectral cues were insufficient for accurate estimation in this range. These findings demonstrate that auditory distance perception can be recalibrated through short-term perceptual learning, even when initial perceptual mappings are strongly degraded. Adaptation generalizes across contexts within the virtual environment, although limitations persist for near-field perception.

Acoustic performance remains insufficiently addressed in early-stage architectural design, where visual and spatial considerations typically guide decision-making. The present research examines the integration of real-time spatial auralization into multi-user virtual reality environments to facilitate collaborative evaluation of architectural acoustic performance during design exploration. A multi-user VR framework has been developed that embeds real-time binaural auralization within a collaborative design evaluation context. The framework supports shared inhabitation of virtual models, auditory exploration of spatialized sound sources, and systematic investigation of how spatial configurations and material properties affect perceived acoustic characteristics. Temporal soundscapes and acoustic implications of alternative materials and spatial arrangements can be evaluated through a gesture-driven interface. Spatialized voice communication further supports co-inhabitation of virtual environments and acoustically informed discussion. By embedding real-time auralization within collaborative virtual environments, the framework repositions acoustic performance as an experiential dimension of architectural design, enabling earlier incorporation of auditory considerations into the design process. An exploratory user study indicates that immersive, real-time auralization can integrate acoustic feedback into collaborative architectural design workflows, supporting multisensory evaluation and reducing dependence on late-stage corrective interventions.

Reproducing the acoustic consequences of source or receiver motion from a set of discrete static room impulse responses (RIRs) is a fundamental cIRCAM:Galleryenge in spatial audio processing, with direct relevance to the generation of training and evaluation data for machine learning systems operating on reverberant speech and audio. This paper presents a comparative evaluation of three offline algorithms for acoustic motion simulation, two of which are ports of previously published methods and one of which is an independent implementation conceptually related to prior work. All three methods are evaluated against a common reference using two complementary validation approaches: an objective analysis based on interaural time difference (ITD) estimation from synthetic binaural signals, and a perceptual evaluation conducted under the MUSHRA protocol using stimuli drawn from a controlled moving-receiver database. Results indicate that frequency-domain interpolation between neighbouring impulse responses provides the most accurate binaural cue reproduction and the highest perceptual similarity to the reference under spectrally demanding stimuli, while nearest-neighbour switching produces the most pronounced artefacts under broadband excitation. Time-domain crossfading between fully convolved signals yields intermediate performance, achieving parity with frequency-domain interpolation for speech and noise but falling significantly behind for music. The combination of ITD-based objective analysis and MUSHRA perceptual evaluation proved informative in characterising method differences, and the two measures converged on a consistent performance ordering across methods.

Multi-channel compact linear loudspeaker arrays combined with Crosstalk Cancellation (CTC) can deliver binaural audio to a user by directing sound precisely at the listener’s ears. This enables the reproduction of binaural, and thus all given spatial audio formats, without the need for headphones. The technique traditionally suffers from a small sweet-spot, which can be overcome by combining real-time head-tracking and position-adaptive beamforming. This workshop introduces the concept of CTC and beamforming combined with user head-tracking, covering both its theoretical foundations and the practical considerations of incorporating head tracking. A live demonstration will showcase head-tracked binaural audio delivered through a position-adaptive CTC soundbar.

We have proposed a method called V2MA (VSVerb Virtual Microphone Array). This method virtually generates spatial room impulse responses (SRIRs) captured by a conventional microphone array using only a set of four impulse responses (IRs) measured by an A-format microphone in the target space. V2MA is based on the concept of geometrical acoustics, that involves a virtual sound source, also known as a mirror source. After measuring the four IRs using an A-format microphone, we calculate the instantaneous sound intensities in the x, y, and z directions. The “source intensities,” that contain sound source information, are detected from these sound intensities. Then, we estimate the locations, strengths, and phase characteristics of the sound sources. The spatial properties of the obtained virtual sound sources can be considered the fingerprint of a target space's reverberant characteristics. Using the manner of geometrical acoustics, we can update the spatial properties of the virtual sound sources to match a neighboring receiver position and a desired directivity. Lastly, we can obtain the SRIRs at any receiver (microphone) position with any directivity in the target room by translating the spatial information of the updated virtual sound sources into time responses. For immersive recording, large microphone arrays are often used. Using V2MA, we can virtually make an immersive recording using such a virtual microphone array, provided that we measure four IRs using an A-format microphone in the target space. In our previous study, we developed the framework of V2MA. To verify the plausibility of V2MA, this paper compares the responses of virtual (V2MA) and real (conventional) microphone arrays using measurement results collected in a practice IRCAM:Gallery under the various conditions. The results show similar overall characteristics, but also suggest the difficulty of a detailed evaluation. We also introduce practical examples of immersive recording using V2MA.