A carregar...
Publicação
Concurrent speech feedback for blind people on touchscreens
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.13 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Smartphone interactions are demanding. Most smartphones come with limited physical buttons, so users can not rely on touch to guide them. Smartphones come with built-in accessibility
mechanisms, for example, screen readers, that make the interaction accessible for blind users.
However, some tasks are still inefficient or cumbersome. Namely, when scanning through a document, users are limited by the single sequential audio channel provided by screen readers. Or
when tasks are interrupted in the presence of other actions.
In this work, we explored alternatives to optimize smartphone interaction by blind people by
leveraging simultaneous audio feedback with different configurations, such as different voices and
spatialization. We researched 5 scenarios: Task interruption, where we use concurrent speech to
reproduce a notification without interrupting the current task; Faster information consumption,
where we leverage concurrent speech to announce up to 4 different contents simultaneously; Text
properties, where the textual formatting is announced; The map scenario, where spatialization
provides feedback on how close or distant a user is from a particular location; And smartphone
interactions scenario, where there is a corresponding sound for each gesture, and instead of reading
the screen elements (e.g., button), a corresponding sound is played. We conducted a study with
10 blind participants whose smartphone usage experience ranges from novice to expert. During
the study, we asked participants’ perceptions and preferences for each scenario, what could be
improved, and in what situations these extra capabilities are valuable to them.
Our results suggest that these extra capabilities we presented are helpful for users, especially if
these can be turned on and off according to the user’s needs and situation. Moreover, we find that
using concurrent speech works best when announcing short messages to the user while listening
to longer content and not so much to have lengthy content announced simultaneously.
Descrição
Tese de Mestrado, Engenharia Informática, 2023, Universidade de Lisboa, Faculdade de Ciências
Palavras-chave
Acessibilidade Saída de voz Sonificação Deficiência Visual Concurrent speech Teses de mestrado - 2023