Workshop D: Respeaking: Accessibility through Speech Recognition
Pablo Romero Fresco
Roehampton University


The present workshop aims to provide ideas, information and some of the skills required to produce live subtitles (mainly for the deaf and hard of hearing) using speech-recognition technology. First, respeaking will be introduced within the context of live subtitling, as opposed to other methods that have been and are still being used to produce subtitles in real time. Subsequently, an overview will be given of the respeaking profession as it is carried out across countries and across channels within the same country. Participants will then be trained to recognise some of the skills required to respeak and will be able to perform hands-on respeaking with the professional software package Protitle, designed by Ninsight. The focus will be placed on the skills to be implemented before the process, that is, the use of speech recognition software (how it works, what types of software are available, how it is to be trained) and on the preparation of the respeaker (anticipation of difficulties, creation of glossaries, macros, etc.). Attention will then turn to the skills to be implemented during the process, namely split-attention, the use of punctuation as well as the management of speed (edited vs verbatim respeaking), rhythm and errors. This practice will be carried out with “real-life” clips from different genres. Finally, participants will also be introduced to other uses of respeaking-based accessibility outside the TV environment, such as respeaking in museums or in classrooms.