Guilherme Castelo Branco

Closed captioning has been a requirement on open TV since 2006. The requirement was only for part of the daily content, but in 2017 closed captioning became mandatory for all programming. With the growing demand for VOD streaming and FAST channels, combined with the global reach of these platforms, the requirements to generate closed captions and subtitles in other languages for these contents has increased even more. Closed Caption generation is a time-consuming and costly job that also requires qualified professionals. The presentation will address how the use of ML/AI techniques such as NLP (Natural language Processing) and STT (Speech to Text) in the automatic closed captions generation has helped thousands of broadcasters around the world and the main challenges in using these technologies to the correct segmentation, punctuation, alignment, accuracy and correction of the generated files, since closed captioning is not just the transcription of audio into text.