Broadcast Beat Magazine 2018 NAB Show Edition - Page 93

AI and machine learning bring new levels of efficiency and costs savings to closed captioning workflows By Giovanni Galvez, Product Manager for captioning And subtitling, Telestream Thanks to the power of the cloud, Machine Learning (ML) and Artificial Intelligence (AI) are finally moving out of their hype phase and are starting to affect just about every aspect of video production and distribution. As the world of broadcasting continues to under- go fundamental changes, cloud-based solutions are increasingly helping to solve more and more of the obstacles traditionally faced by video pro- duction professionals. From the creation of musi- cal scores, to the automation of QC processes, AI and ML will have a profound impact on how content creators and broadcast professionals do their work over the next decade. But one area that is poised to benefit the most from some of the recent developments in AI and ML is closed captioning and subtitling. Video captioning remains a challenge. New regulations may make it even more so. Multi-platform delivery to broadcast, VoD and web streaming channels requires, by law, the presence of captions and subtitles. Content pro- viders who don’t have them now risk significant fines. Captions must be done and they must be done cost-effectively. There are many variables in the audio of broad- cast TV content that can cause problems for auto transcription software and human operators. When attempting to decipher words, noisy action scenes, crowds, or heavily accented speakers prove to be a major challenge. Misspelling, inau- dible designations, and difficulty in interpreting the context of spoken words mean lots of costly human scrutiny in order to finalize a video asset. The trailer for Amazon Studio’s Oscar-