저시력자를 위한 멀티미디어 콘텐츠용 음성 안내의 UX 디자인 가이드라인

Background: Because of the recent popularization of smart devices, it is common for people with low-vision to enjoy multimedia contents on an everyday basis. Although auditory guidance is an essential function in those situations, it is not carefully considered in designing the UX of it for better accessibility. The study especially focuses on finding out how many differences are required in sound levels and what amount of information is possibly acceptable for people with lowvision when enjoying multimedia contents. Methods: The study conducted user tests providing 4 sets of multimedia contents(verbal/nonverbal) and auditory guides (verbal/non-verbal) for one, and a maximum of 9 chunks of words for the other. 30 people with low-vision aged 18~50 were tested and interviewed afterwards. Results: 1) Voice guidance showed that more than 80% accuracy was provided in contrast to +45dB(unrelated contents type), and the recognition of non-verbal content was +15dB or more with verbal contents, while non-verbal contents were highly recognized in the equal volume. 2) When provided from 3 to 8 chunks, the accuracy was more than 80%. In case be to provide 3 to 8 chunks, but users preferred to provide less than 6 chunks. This is related to the analysis of two groups(students/normal), which is similar until 5 chunks, but the difference occurs at 6 chunks. Conclusions: The results showed that different level differences (verbal contents: 45dB(verbal)-15dB(non-verbal)/nonverbal contents: 45dB(verbal)-0dB(non-verbal)) are required, and that more than 5 chunks are optimal, among other findings. The study also summarizes the findings and converts them into UX design guidelines for designing audio guidance for people with low-vision.