More and more live subtitling is being produced across Europe for deaf and hard-of-hearing people in response to legislative pressure for equal access. Production methods are changing from the days when most live subtitles were created using fast keyboard systems, to an ever-increasing adoption of speech recognition using a trained re-speaker, sometimes supported by a downstream corrector. The art of speech-based subtitle production has matured from an experimental technique towards an established discipline with its own need for practical training and skills development based on a pragmatic understanding of the problems and challenges of the task and the quality issues that affect the audience.
The use of speech recognition as a rapid and economic means of accurate real-time transcription during live programme broadcasts sounds fine in theory, but in practice it involves significant specialist training and the development of novel techniques to overcome the problems inherent in phonetic transcription. The subtitle re-speaker acts as a channel of communication to the subtitle viewer, and at the same time the quality of the resultant subtitles affects the perceived quality of the broadcast channel itself. It is thus of key importance that re-speakers are adequately trained and skilled, and thereby equipped to deliver the best possible results.
This talk examines in some detail the seven key cognitive skills which underpin the art of successful re-speaking. Understanding and mastering these skills – and the means to attain them – is key to the delivery of quality services. By looking at the tasks in isolation it is possible to establish the building blocks necessary for successful training, and understanding how develop and then integrate the skills without leading to cognitive overload is vital to successful training. A good re-speaker is generally a confident re-speaker, and this paper shows how that confidence can be developed and achieved. As a result, we can look forward to increasing levels of service provision in the drive to provide media access for all.

Andrew LAMBOURNE is one of the pioneers of the use of speech input tools for subtitling, and having trained re-speakers in the UK and Europe for many years, he is well placed to reflect on the core skills which are necessary for success. His own work in the subtitling field started 30 years ago as a researcher at the University of Southampton in the UK, from where he went to ORACLE Teletext and founded the UK’s first live news subtitling operation. Leaving to form his own business in 1990, Andrew led the development of the successful WinCAPS subtitling system. Since then he has worked alongside many broadcasters and subtitling houses to assist them in breaking new ground in service provision. His firm belief is that new technology should be harnessed to provide subtitlers with the assistance to enable higher productivity, but that at the core of success is effective development and deployment of the human skills which are vital to achieving and maintaining quality.