Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.617231
Title: The impact of sound on virtual landscape perception : an empirical evaluation of aural-visual interaction for 3D visualization
Author: Lindquist, Mark
ISNI:       0000 0004 5349 2010
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2014
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
An understanding of quantitative and qualitative landscape characteristics is necessary to successfully articulate intervention or change in the landscape. In landscape planning and design 3D visualizations have been used to successfully communicate various aspects of landscape to a diverse population, though they have been shown to lag behind real-world experience in perceptual experiments. There is evidence that engaging other senses can alter the perception of 3D visualizations, which this thesis used as a departure point for the research project. Three research questions guide the investigation. The first research question is: How do fundamental elements in visualizations (i.e. terrain, vegetation and built form) interact with fundamental sound types (i.e. anthropogenic, mechanical and natural) to affect perceived realism of, and preference for, 3D landscape visualization? The research used empirical methods of a controlled experiment and statistical analysis of quantitative survey responses to examine the perceptual responses to the interaction aural and visual stimuli in St. James’s Park, London, UK. The visualizations were sourced from Google Earth, and the sounds recorded in situ, with Google Earth chosen as it is being used more frequently in landscape planning and design processes, though has received very little perceptual research focus. The second research question is: Do different user characteristics interact with combined aural-visual stimuli to alter perceived realism and preferences for 3D visualization? The final research question emerged out of the experiment design concentrating on research methodology: How effective is the Internet for aural-visual data collection compared to the laboratory setting? The results of the quantitative analysis can be summarized as follows: For research question 1 the results show that sound alters 3D visualization perception both positively and negatively, which varies by landscape element. For all visual conditions mechanical sound significantly lowers preference. For visualizations showing terrain only perceived realism and preference are significantly lowered by anthropogenic sound and significantly raised by natural sound for both realism and preference. For visualizations showing a combination of terrain with built form anthropogenic and mechanical sound significantly raises perceived realism. For visualizations showing a combination of terrain, vegetation and some built form a more complicated interaction occurs for realism, which is moderated by the amount of built form in the scene, e.g. with no buildings in the scene traffic and speech significantly lower realism ratings in similar ways while a small amount of built form visible resulted in speech significantly raising realism ratings. Preference was significantly lowered by anthropogenic and mechanical sound the most out of all three visual conditions. For research question 2 the results confirm that perception can vary for realism by gender and first language differences, and preference by age, first language, cultural and professional background and 3D familiarity. Finally for research question 3 and implications for Internet-based multisensory experiments there is strong evidence that audio hardware and experimental condition (laboratory vs. online) do not significantly alter realism and preference ratings, though larger display sizes can have a significant but very small effect on preference ratings (+/- 0.08 on a 5-point scale). The results indicate that sound significantly alters the perception of realism and preference for landscape simulated via 3D visualizations, with the congruence of aural and visual stimuli having a strong impact on both perceptual responses. The results provide important empirical evidence for future research to build upon, and raise important questions relating to authenticity of landscape experience, particularly when relying solely on visual material as visuals alone do not accurately simulate landscape experience. In addition the research confirms the cross-sensory nature of perception in virtual environments. As a result the inclusion of sound for landscape visualization and aesthetic research is concluded to be of critical importance. The research results suggest that when using sound with 3D visualizations the sound content match the visualized material, and to avoid using sounds that contain human speech unless there is a very strong reason to do so (e.g. there are humans in the visualization). The final chapter discusses opportunities for integrating sound with 3D visualizations in order to increase the perception of realism and preference in landscape planning and design processes, and concludes with areas for future research.
Supervisor: Lange, Eckart ; Kang, Jian Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.617231  DOI: Not available
Share: