MirrorMirror

Abstract

Many people around the world have difficulties in day-to-day conversation due to hearing loss. Hearing aids often fail to offer enough benefits and have low adoption rates. However, people with hearing loss find that speechreading can improve their understanding during conversation, but speechreading is a challenging skill to learn. Speechreading classes can improve acquisition, however there are a limited number of classes available and students can only practice effectively when attending class. To address this, we conducted a postal survey with 59 speechreading students to understand students' perspectives on practicing. Using our findings, we developed an Android application called MirrorMirror -a new Speechreading Acquisition Tool (SAT) that allows students to practice their speechreading by recording and watching videos of people they frequently speak with. We evaluated MirrorMirror through three case studies with speechreading students and found that they could effectively target their speechreading practice on people, words and situations they encounter during daily conversations.

Adaptive Subtitles: Preferences and Trade-Offs in Real-Time Media Adaption

“…Participants reported using a variety of services to watch scripted entertainment: Netfix (82 participants), Amazon Video (45), Virgin Media ( 14), Freeview ( 14), NowTV ( 13), Terrestrial TV (12), Sky (8), Apple TV (6), Sky Go (4), Not Given (2). Additionally, participants reported if they used any of the following: 'On Demand TV' (59 participants), 'Recorded Programmes' (25), 'On the Go Live TV' (7).…”

“…We chose clips using the following criteria (similar to [11]): (a) content was not originally broadcast within a month of our study taking place (to reduce potential familiarity with content); (b) content did not contain ofensive language, or potentially disturbing material; (c) there were extensive talkinghead shots (i.e., as much spoken dialogue as possible); (d) content containing speakers with local regional accents (i.e. local to study location) was excluded; (e) excerpts were around fve minutes in duration; (f) similar levels of activity and engagement across content; (g) subtitles were not superimposed on content before transcoding; (h) content contained challenges that afect people's ability to lipread/speechread such as speakers turning away, and, facial hair [24,25]. iPlayer subtitles are presented using EBU-TT (timed text) format [64].…”