Spatial Hearing Augmentation to Improve Hearing Assistive Technology
A 3-year EPSRC New Investigator Award led by Aidan Hogg at QMUL, in collaboration with Imperial College London and Eargym Ltd, expected to begin in September 2026.
This project explores a new approach to hearing assistive technology: not just deciding what sounds to suppress, but restoring and enhancing the spatial cues people rely on to understand and navigate the world.
The research asks whether artificial, enhanced spatial cues can help people with hearing loss better localise sounds, follow conversations in noisy environments, and maintain situational awareness. By combining machine learning, immersive audio, behavioural experiments, and perceptual training, the project aims to lay the foundations for a new generation of hearing assistive technologies that augment perception rather than simply restore lost function.
Overview
Spatial hearing is central to everyday life. It helps us detect danger, orient ourselves, separate competing sound sources, and follow speech in noisy environments. However, for many users of hearing aids and other hearing assistive technologies, these spatial cues are significantly degraded.
Current hearing aids often focus on noise reduction and signal enhancement, but this can come at the expense of spatial awareness. This project investigates a different strategy: using AI and computational auditory modelling to develop enhanced artificial spatial cues that support listening, localisation, and autonomy in complex acoustic scenes.
In particular, the project will develop and evaluate superhuman head-related transfer functions (HRTFs) derived from virtual ear shapes that exaggerate spatial filtering effects beyond those found in natural human anatomy.
Vision
We’re rethinking hearing aids — not just what to suppress, but how to restore and enhance the spatial cues people rely on to understand and navigate the world.
The longer-term vision is to establish a new perceptually driven framework for hearing assistive technology, in which intelligent systems do not simply restore lost function but actively augment perception. Beyond conventional hearing aids, this research may also inform future technologies such as AR audio devices and spatial hearing glasses.
Research questions
The project is structured around four core questions:
- Which spatial cues most effectively support behavioural tasks such as speech-in-noise understanding and localisation?
- How should spatial cues be modified to create optimal enhanced or “superhuman” HRTFs?
- To what extent can users adapt to these enhanced, non-individualised spatial cues?
- Do enhanced cues result in measurable performance improvements relative to current hearing-aid approaches?
Work packages
WP1: Creation
Create a large open dataset of superhuman HRTFs using parametric models of ear geometry.
WP2: Optimisation
Use machine learning and auditory models to identify HRTF modifications that improve specific behavioural tasks such as localisation and speech intelligibility.
WP3: Adaptation
Investigate how well users can adapt to enhanced spatial cues through structured perceptual training.
WP4: Evaluation
Evaluate benefits in realistic listening conditions and compare performance against current hearing-aid baselines.
Create a large open dataset of superhuman HRTFs using parametric models of ear geometry.
WP2: Optimisation
Use machine learning and auditory models to identify HRTF modifications that improve specific behavioural tasks such as localisation and speech intelligibility.
WP3: Adaptation
Investigate how well users can adapt to enhanced spatial cues through structured perceptual training.
WP4: Evaluation
Evaluate benefits in realistic listening conditions and compare performance against current hearing-aid baselines.
Why it matters
Many hearing-aid users still struggle to locate sounds, follow conversations in noisy places, and maintain awareness of their surroundings. These difficulties affect autonomy, communication, and quality of life.
This project targets those problems directly by asking whether better spatial cue design can improve listening in the real world. If successful, the work could help shape more effective and more usable hearing assistive technologies, while also contributing to broader research in auditory perception, immersive audio, and sensory augmentation.
Team and partners
Project Lead
Aidan Hogg
Lecturer in Computer Science, Queen Mary University of London
Aidan Hogg
Lecturer in Computer Science, Queen Mary University of London
Project Partners
Audio Experience Design group at Imperial College London and Eargym Ltd. Imperial contributes specialist spatial audio facilities, equipment, and expertise, while Eargym provides advisory support, platform access, and participant engagement.
Funding
This project is funded by the Engineering and Physical Sciences Research Council (EPSRC) through a New Investigator Award.
Project value: £607,233
Duration: 3 years
Planned start: September 2026
Duration: 3 years
Planned start: September 2026
Opportunities
We expect to recruit a full-time 3-year Postdoctoral Research Associate (PDRA) to work on the project. The post will likely be advertised in the coming months, once the project setup has been finalised.
Publications
Publications arising from the project will be listed here as they appear.
Updates
Updates on recruitment, team members, talks, datasets, and other project milestones will be posted here.
Contact
For collaborations, recruitment, or project enquiries, please contact Aidan Hogg via the contact details on the Contact page.