In-Situ Study of Blind Individuals Listening to Audio-Visual Contents
Claude Chapdelaine · 2010 · Proceedings of the 12th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2010) · doi:10.1145/1878803.1878816
Summary
This paper presents an in-situ observational study examining how ten legally blind individuals (WHO levels 3-5) listen to and comprehend audio-visual content such as films, television dramas, and science reports. The research aims to understand the type, quantity, and frequency of visual information that blind viewers need described, and how these needs differ based on residual vision. Participants were divided into two groups: Group A (five totally blind individuals, WHO levels 4-5) and Group B (five individuals with residual vision, WHO levels 3-4). Both groups watched four different video clips — two short films, a drama, and a science report — varying in complexity, number of actors, dialogue density, contrast quality, and background noise. Using a think-aloud protocol, participants verbally requested information whenever they felt they were missing something, and a sighted control group independently identified the key visual concepts in each clip. The study builds on the cognitive science understanding that vision is the dominant sensory modality, auditory processing is more cognitively demanding than visual processing, and blind individuals must rely more heavily on short-term memory when filling in missing visual information.
Key findings
Group A (totally blind) made 57.4% of all information requests compared to Group B's 42.6%, confirming that residual vision reduces but does not eliminate the need for description. Requests fell into six categories: who is present, where the action occurs, what action is happening, description of people/objects, facial expression, and sound identification. Action was by far the most requested category for both groups (52.4% for Group A, 60% for Group B), indicating that understanding what is happening is more critical than knowing who or where. Facial expression requests came almost exclusively from Group A, since Group B could often perceive expressions with their residual vision. Content complexity had a significant effect: when contrast was poor (as in the drama with dark night scenes), Group B's behavior converged with Group A's, effectively losing their residual vision advantage. The science report produced an unexpected reversal where Group B made more inquiries than confirmations, likely because the narration provided enough context that residual vision became less relevant. Requests were classified as either confirmations (51% for Group A, 52.1% for Group B) or inquiries, suggesting blind viewers frequently seek reassurance about what they partially understand rather than entirely new information.
Relevance
This research challenges the one-size-fits-all approach to audio description by demonstrating that blind individuals with different levels of residual vision have measurably different information needs. The findings have direct implications for audio description production: a basic level should cover action (the dominant need), but flexible, on-demand description should be available for details like facial expressions and scene descriptions that some viewers can perceive independently. The study led to an accessible video player with a toggle between standard and extended description, plus three on-demand buttons for actions, actors, and scene names. For accessibility practitioners, the key takeaway is that the visual impairment spectrum requires adaptive rather than uniform accessibility solutions — a totally blind viewer and one with residual vision watching the same content have fundamentally different gaps in comprehension. The finding that poor contrast eliminates the advantage of residual vision also reinforces the importance of visual quality in broadcast content.
Tags: audio description · video description · visual impairment · blindness · residual vision · multimedia accessibility · video accessibility · user research
Standards referenced: WHO Visual Impairment Classification · ITC Guidance on Standards for Audiodescription