Audio Subtitles or Spoken Subtitles/Captions: An ecological media accessibility service

Abstract

Subtitles are the most common and versatile access media service across platforms, languages, and technologies. Subtitles, which traditionally were between languages, have now expanded to same language subtitles, subtitles for the deaf and hard of hearing, and verbatim transcriptions. The production of subtitles has also changed greatly from the manual production with a subtitle editor, to respeaking, stenography, velotype, and the many possibilities offered by machine translation and crowdsourcing. Subtitles can be read on the main or on a secondary screen, in a wearable, as the mobile phone, intelligent glasses or a head mounted device for XR (eXtended Reality) environments. Finally, subtitles can become a hybrid access service when are generated with Easy-to-Read guidelines, or are read aloud. This last case is the objective of study in this article: spoken or audio subtitles (AST). In the first section, an introduction will serve as a state of the art regarding audio subtitle literature in the field of Audiovisual Translation or Media Accessibility. Most research has focused at audio subtitles for movies from a descriptive approach, with a growing interest in multilanguage movies. Little has been written regarding the technology towards the generation of audio subtitles and the many workflows allowing for a versatile audio subtitle service in many presentations. This will be the content of the second part of the article. After looking at some possible use cases, it will look at existing solutions for three very different situations. The third part of the article will summarise existing requirements looking to widen the use cases, taking advantage of existing technology. The final section will look at possible future applications, and avenues for research.