Método de avaliação da qualidade de audiodescrição – MADE

• Ano de defesa: 2020
• Autor(a) principal: Cordeiro, Rafael Farias
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Informática; Programa de Pós-Graduação em Informática; UFPB
• Palavras-chave em Português: Audiodescrição; Acessibilidade; Experiência do usuário; Audiodescription; Accessibility; User Experience; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/20277

Audio description (AD) is a narrative audio track that assists people with visual impairments in understanding audiovisual content. The professional responsible for generating an AD is the audio describer. It has the task of transforming visual information from videos or audio presentations to viewers. The transformation process takes time and has a high cost. As a support for these professionals, with a view to reducing the time and the cost of executing audio description, automatic audio description generator systems that were used. The aim of this research is to develop and apply a method to assess the quality of audio description, whether generated by humans or machines. Method of Audio Description Evaluation– (MADE), allows to evaluate the quality of the AD according to the feasibility and aspects of interest of the evaluators, considering three evaluation methods: (i) from the analysis of technical aspects by the evaluator (which evaluates how essential visual information is described), without the involvement of the end user (ii) from the evaluation of the visually impaired user experience when using an AD; or (iii) hybrid scenario that includes the technical and user aspects and the complementarity of the results achieved. A MADE application was analyzed for the hybrid scenario from a case study. In this article, three types of AD were used for three short films and were attended by eight visually impaired people. As a result, it was possible to identify how audio descriptions behave to evoke the feelings of users, demonstrate a perception of usefulness and assist in understanding the story, as well as aspects of improvement for ADs. MADE can be used by researchers and developers in the field of AD, to identify improvements in audio description and ensure minimum quality before it is made available to the final public.