posted on 2022-03-29, 03:19authored byAdam Weisser
In everyday life, the ears constantly receive information from the environment that is carried by sound waves. Despite the uncontrolled nature of this information, normal hearing listeners can make sense of the unstructured sounds. In contrast, hearing has mostly been studied using acoustic stimuli that are rigidly controlled. The difference between understanding such artificial stimuli and real-world hearing has produced the distinction of ’complex acoustic environments‘ (CAEs) – a catch-all term for everything that the laboratory-based stimuli are not. However, an exact definition has been elusive, despite growing use in hearing research – especially in research focusing on hearing-impaired individuals, who struggle communicating in cocktail-party like scenarios. In this work, several aspects of CAEs were studied with the intention to understand where the complexity lies. The work consists of four main parts in manuscript form. The first paper presents the Ambisonic Recordings of Typical Environments (ARTE) database, which contains 3D recordings of everyday scenarios. These were the primary stimuli used in all subsequent parts of this work. In the second paper, the speech and noise levels of real conversations were measured at two distances between talkers. The third paper reviews the origins and uses of the CAE concept in literature and presents a framework that summarizes and defines the characteristics that can drive complexity in a given environment. A subset of these characteristics were tested by 65 listeners who listened to the ARTE scenes (also with target speech), while answering a three-stage questionnaire. The analysis revealed that listeners’ perceived complexity is associated with both the loudness and the variability of the scenes. The final paper analyzes the responses of listeners to an open-ended auditory scene analysis task using ARTE, and the observed limitations are discussed. A concluding discussion dissects how the auditory system deals with realistic amounts of acoustic information.
History
Table of Contents
1. Introduction -- 2. The ambisonic recordings of typical environments (ARTE) database -- 3. Conversational speech levels and signal-to-noise ratios in realistic conditions -- 4. Complex acoustic environments: review, framework and subjective model -- 5. Exploring the ability of listeners to identify events in realistic environments -- 6. General discussion and conclusions -- Appendices.
Notes
Bibliography: pages 191-215
Thesis by publication.
Awarding Institution
Macquarie University
Degree Type
Thesis PhD
Degree
PhD, Macquarie University, Faculty of Human Science, Department of Linguistics
Department, Centre or School
Department of Linguistics
Year of Award
2018
Principal Supervisor
Jörg Buchholz
Additional Supervisor 1
Gitte Keidser
Rights
Copyright Adam Weisser 2018
Copyright disclaimer: http://mq.edu.au/library/copyright