Improving broadcast accessibility for hard of hearing individuals : using object-based audio personalisation and narrative importance

Ward, L 2020, Improving broadcast accessibility for hard of hearing individuals : using object-based audio personalisation and narrative importance , PhD thesis, University of Salford.

[img]
Preview
PDF
Download (12MB) | Preview

Abstract

Technological advances in broadcasting can be the impetus for advances in accessibility services. For the 11 million individuals in the United Kingdom with some degree of hearing loss, the advent of object-based broadcasting and it’s personalisation features has the potential to facilitate a transition towards more accessible broadcast audio. Part I of this work conducts a systematic review of previous object-based accessibility research, identifying the personalisation of redundant non-speech objects as a potentially high impact yet unexplored area of research. Guided by these findings, and the results of a survey of end-user needs, the specific research questions of this work are then developed as: - What is the relationship between redundant non-speech audio objects and broadcast speech intelligibility, for normal and hard of hearing listeners? - Can a system be designed which allows end-users to control the balance between audio objects for dramatic content which is simple to use and preserves comprehension? Part II of this work shows that the presence of redundant non-speech sounds improve speech recognition in noise in normal hearing listeners, even when the sound partially masks the speech. Subsequent investigations show that this effect exists within hard of hearing cohorts also, and the benefit yielded by non-speech sounds can be predicted by the severity of hearing loss in an individual's better hearing ear. Part III work translates these novel findings into practical broadcast accessibility technology, through the development of a new conceptual framework called: `Narrative Importance’. Based on this framework, production tools and an end-user interface are developed and deployed in a large scale public trial. The results of this trial demonstrate that this new approach to accessible audio can deliver content which is more enjoyable with reduced energetic masking of speech, whilst still maintaining the creative integrity and comprehension of the content.

Item Type: Thesis (PhD)
Contributors: Shirley, BG (Supervisor) and Davies, WJ (Supervisor)
Schools: Schools > School of Computing, Science and Engineering
Depositing User: Lauren Ward
Date Deposited: 15 Jun 2020 11:37
Last Modified: 15 Jul 2020 02:30
URI: http://usir.salford.ac.uk/id/eprint/57144

Actions (login required)

Edit record (repository staff only) Edit record (repository staff only)

Downloads

Downloads per month over past year