As progressing social mores and new legislation mandate greater levels of accessibility across broadcast television, online video, education, and the workplace, it becomes ever more important to deliver captioned content at lower cost, with higher quality, and more quickly. To reach those goals, attention is shifting toward automation. While many attempts have been made to apply large vocabulary ASR transcription for this task, the results on noisy, post-processed, multi-speaker recordings are insufficient for most real-world applications, to say the least. More viable is the automated synchronization of a human-prepared transcript with the media, and automating the breaking up of the text into well-structured, meaningful captions. SyncWords is a platform that incorporates innovative, originally developed algorithms for both of these tasks, and presents an easy-to-use commercial Web interface for submitting work and receiving captions. In addition, an interactive review tool is available in which user corrections are fed back and cause a re-evaluation of the results in real time.
Bibliographic reference. Dubinsky, Aleksandr (2014): "Syncwords: a platform for semi-automated closed captioning and subtitles", In INTERSPEECH-2014, 1500-1501.